International Journal of Modelling, Identification and Control (95 papers in press)
Front-end matching optimised algorithm of Cartographer with multi-resolution layered search strategy
by Yiqun Di, Xianghua Ma, DONDongG Guo
Abstract: Cartographer is used in LiDAR SLAM (Simultaneous Localization and Mapping), and it produces a stable mapping results and uses loop closure detection to eliminate accumulated errors. However, considerable calculation is needed to guarantee the accuracy of the mapping results and the decrease of accumulated errors when CSM (Correlative Scan Matcher) is used to carry out the global search of the front-end matching. In order to resolve the above problem, a structure, multi-resolution layered search strategy (MLSS), is proposed in the frond-end of Cartographer. MLSS structure makes use of layered strategy and the full-number locally optimal principle to decrease accumulated errors from sources and time costs of front-end matching. Verification and comparison studies demonstrate the effectiveness of the proposed structure.
Keywords: SLAM; multi-resolution layered search strategy; front-end matching; full-number locally optimal principle.
Outlier detection algorithm based on deviation characteristic
by Yong Wang, Hongbin Wang, Pengcheng Sun, Xinliang Yin
Abstract: Outlier mining focuses on researching rare events through detection and analysis to dig out the valuable knowledge from them. In the static data set environment, the traditional LOF algorithm calculates the local outlier factor through the whole data set and requires a lot of computing time. To solve this problem, the algorithm divides the data space into grids, and calculates the local outlier factor based on the centroids of the grids. Since the grid number is less than data point number, the time complexity is obviously reduced under acceptable error. When the new data points are added, it can rapidly detect outliers. The contrast experiment results show that the new algorithm can reduce the computation time and improve the efficiency, while achieving comparable accuracy.
Keywords: outlier detection; local outlier factor; deviation characteristic; fast LOF detection algorithm.
Multi-stopping criterion multi-feature-based multi-objective cohort intelligence algorithm for thermoacoustic engine optimisation
by Mukundraj Patil, Satish Kumar
Abstract: The aim of this research is to investigate the performance characteristics of a thermoacoustic engine (TAE) using the multi-stopping criterion multi-feature-based Multi-objective Cohort Intelligence (MOCI) algorithm. MOCI and the state-of-the-art algorithms are applied to study performance characteristics of a thermoacoustic engine. Exploratory and statistical analyses revealed better performance of the MOCI algorithm on qualitative and quantitative performance metrics. Post-optimality analysis showed a better region of interest for an analyst and the desirable working ranges for each variable of TAE design. The pressure-frequency relationship showed high correlation and it is useful for future study and detailed investigation of the thermoacoustic phenomenon. MOCI established competitive results that are useful in benchmarking TAE performances in future researches. The effective design of the TAE using the MOCI algorithm aids in the sustainable development of society in terms of affordable and clean energy, clean climate and responsible consumption and production.
Keywords: multi-objective; cohort intelligence; thermal device optimisation; stopping criterion; thermoacoustic; thermoacoustic prime mover; performance benchmarking.
Hybrid fuzzy level set approach for multiple sclerosis lesions assessment in magnetic resonance brain images
by Chaima Dachraoui, Aymen Mouelhi, Cyrine Drissi, Salam Labidi
Abstract: Multiple sclerosis is a neurological autoimmune disease characterised by progressive degeneration due to the myelin attack on the central nervous system. The diagnosis is based essentially on clinical features and additional examinations, mainly magnetic resonance imaging findings. The diagnosis of multiple sclerosis requires all defined criteria that aim to study spatial and temporal dissemination. Thus, in this work, the automatic segmentation of multiple sclerosis plaques is chosen in order to computerise the process and the follow-up. This approach is a hybrid method allowing to combine the fuzzy C-Means method with geodesic models until obtaining an automatic task. This is a retrospective study in which data were collected from the National Institute of Neurology in Tunisia. The proposed method is for medical neuroradiology research. The eventual results are improved after some pre-treatments, therefore, there is interest in pre-processing. High accuracy was achieved for the models discussed in this paper (93% - 84%). Accordingly, the suitability and practical usefulness of the 'simple' pre-treatments to achieve multiple sclerosis classification are demonstrated.
Keywords: multiple sclerosis; brain; MRI; automatic segmentation; hybrid approach; geodesic contour; fuzzy C-means; lesions segmentation; T2 FLAIR.
An optimal control problem associated with Lorentz group SO(3; 1)
by Archana Tiwari, Kishor Chandra Pati
Abstract: The Lorentz group is the group of transformation of spatial and time coordinates associated with the special theory of relativity. It is both a group and admits a topological description as a smooth manifold. Hence, the Lorentz group can act as a configuration manifold of control systems. This opens up the scope to study the controllability and optimal control problems of control systems on the Lorentz group. Here, a left-invariant driftless control system is defined on the group. An optimal control problem is formulated with an objective to minimise the cost function and satisfy the given dynamical constraints. The stability of the system around equilibrium points is studied. Two unconventional numerical integrators, the Kahan and Lie-Trotter integrator and the conventional Runge-Kutta integrator, are implemented to study the system dynamics and their corresponding trajectories are shown.
Keywords: Lorentz group; control system; optimal control; stability.
PFOID-SMC approach to mitigate the effect of disturbance and parametric uncertainty on the quadcopter
by Sanjay Kumar, Lillie Dewan
Abstract: This paper investigates the possible types of disturbance, viz aerodynamic factors, sensor noise, random noise, wind effect, and parametric uncertainty, acting on an unmanned aerial vehicle quadcopter and their adverse effects on the performance. To ensure the desired performance and increase the quadcopter's stability, accuracy, and task reliability, disturbances and uncertainties detection and diagnosis are very important. Proportional-Fractional Order-Integral-Derivative (PFOID) surface-based Sliding Mode Controller (SMC) is proposed to mitigate the effect of disturbances and uncertainties. The system's stability is proved using the Lyapunov criterion, and performance is validated by simulation. Results of the PFOID surface-based SMC are compared with proportional-integral-derivative surface-based SMC.
Keywords: nonlinear systems; quadcopter dynamics; aerodynamic effects; parametric uncertainty; sensor noise; wind noise; proportional integral derivative; proportional fractional order integral derivative; sliding surface; sliding mode control.
Investigation and realisation of PID and LQR control methods in Parrot Mambo minidrone
by Mohamed Okasha, Jordan Kralev, Maidul Islam
Abstract: A quadcopter is a multivariate, unstable, and highly nonlinear dynamic system, which requires a proper controller to ensure the stability and performance of the system. This study aims to investigate different types of control methods for Parrot Mambo minidrone. In this study, different control methods used on quadcopters such as Proportional Integral Derivative (PID) and Linear Quadratic Regulator (LQR) are investigated and implemented. First, the Parrot built-in PID controller is tested in simulation and experimentally validated using MATLAB and Simulink, followed by the design of the LQR controller. For both controllers, the operating point is selected such that the minidrone can hover along the vertical dimension. The design and tuning of the LQR is carried out by giving weight on the inertial coordinates and on the motor signals, which determine the performance of the minidrone with minimisation of quadratic cost function. The LQR controller shows that the system tends to have less overshoot in vertical trajectory. In many testing scenarios, the LQR controller shows better overall performance compared with the PID controller in both simulation and experimental testing.
Keywords: PID; LQR; UAV; Parrot Mambo minidrone; minidrone; control methods.
Research on the emerging mechanism of complex networks community dividing based on cellular automata
by Xiaodong Qian, Yanfu Luo, Peng Zhang
Abstract: This paper takes the complex network as the research object. It firstly proposes a simple network centre node determination method, and based on the connection status between the non-central node and the central node in the network, divides the network nodes into three categories, that are ? (non-central node, and only directly connected to one central node), ? (non-central node, and directly connected to multiple central nodes) and ? (non-central node, and not directly connected to any central node). Then, the cellular automaton is used as the computing framework, and the central node, the node connected to the centre, and the node not connected to the centre are used to construct the three-dimensional cell space, and defines rules for updating the connection among ?, ? and ?. Defining the rules for updating the connection between and the central nodes and the edge update rules between and with central nodes which allows nodes in a network community to connect to community centres with only a certain weight. Based on the above rules to explore the process of community structures emerging. Finally, using a number of data sets to verify the algorithm proposed in this paper, and the results show that the rules defined in this paper can be used to complete the division of the network community while a new structure emerges on the network.
Keywords: cellular automata; complex networks; community dividing; emerging mechanism.
Quality prediction method for plasticising process of single-base gun propellant based on modified multiway JY-PLS transfer model
by Mingyi Yang, Junyi Wang, Di Huang, Zhigang Xu, Tingjiang Yu, Shubo Chen
Abstract: To solve the problem that the data of a new plasticising process of single-base gun propellant is not enough to build an accurate model for predicting the product quality, a modified multiway joint-Y partial least squares (MMJY-PLS) transfer model for final quality prediction in the plasticising process is proposed. Firstly, the modelling efficiency of the new plasticising process is improved by transfer learning using the plasticising process data of a similar source domain. Secondly, the internal structure of the JY-PLS transfer model is improved so that the regression coefficients of the source and the target domain are no longer the same, and the prediction performance of the model is further improved. Finally, the MMJY-PLS process transfer model is applied to predict the final quality index in the plasticising process of two single-base gun propellant products, and the experimental results show its effectiveness.
Keywords: gun propellant; plasticizing process; JY-PLS; quality prediction; model transfer;online update.
An improved deep forest classification algorithm
by Jiaman Ding, Cuihua Liu, Runxin Li, Jinguo You, Lianyin Jia
Abstract: Recent studies suggest that deep forest is both a kind of deep learning while overcoming the problems of excessive parameters and hard-to-adjust parameters in deep neural networks. However, the deep forest does not consider the classification contributions of each forest and both the time and memory consumption caused by high-dimensional features, thus leading to inefficiency. To solve these problems, an improved deep forest (IgcForest) classification algorithm is proposed in this paper. IgcForest first preserves the class distribution vectors of the main features by a pooling strategy to realise the feature reduction and reuse. Next, an adaptive weighting strategy is proposed to calculate the weight of each forest in the cascade structure. Experiments are carried out on multiple UCI datasets to verify the effectiveness of IgcForest, and the results demonstrate that IgcForest achieves better results in different evaluation indexes than other algorithms.
Keywords: deep learning; deep forest; ensemble methods; pooling strategy; self-adaptive differential evolution algorithm.
Recursive identification of an IIR system using binary input/output measurements
by Hicham Oualla, Mathieu Pouliquen, Miloud Frikel, Said Safi
Abstract: The identification of a system using binary measurements of both the input and output is a problem relevant to a number of applications. Although the identification of Finite Impulse Response (FIR) models has been thorough, more complex model structures still need to be investigated. In this paper, the problem of identifying an Infinite Impulse Response (IIR) systems is studied. A three-step real-time algorithm is proposed. The proposed algorithm is based on the use of a Least Mean Square (LMS) algorithm to estimate the correlation function, then estimate the parameters of the system. A convergence analysis of the proposed algorithm based on those of the LMS algorithm is presented. Some simulation results are included to illustrate the performance of the proposed algorithm.
Keywords: real-time identification; IIR systems; binary input/output.
Development of control-oriented models for a building under regular heating, ventilation, and air-conditioning operation: a comparative simulation study and an experimental validation
by Heman Shamachurn, S.Z. Sayed Hassen
Abstract: The development of models is a major barrier to the fast and widespread adoption of model predictive control for building HVAC systems. This paper proposes the subspace identification technique, refined through the prediction error method, to quickly obtain a model for the accurate indoor temperature prediction, even with little identification data, even in the presence of large unmeasured disturbances and noisy identification data, and even using data which was collected during the regular HVAC operation of a building. The identification issues associated with grey-box models were thoroughly investigated. In particular, the development of a grey-box model was found to be a complex, lengthy and computationally intensive process, even for a single-zone building, and the models were not physically meaningful. The proposed method was found to be much easier and faster, with a potential for direct practical application. Analysis on experimental data from an existing building provided promising results.
Keywords: RC model; subspace identification; regular HVAC operation; open-loop data; closed-loop data; experimental validation; DesignBuilder.
Voltage reconfiguration faulttolerant technique for twolevel three-phase inverter integrated in a wind energy system
by Mokhtar Mahmoud Mohammedi, Azeddine Bendiabdellah, Tayeb Allaoui
Abstract: The present paper focuses on the study of a system composed of a Doubly-Fed Induction Generator (DFIG) linked to the network by the back-to-back converter and related to a wind turbine constituting the Wind Energy Conversion System (DFIG-WECS). In order to reach a maximum power point tracking (MPPT), a control strategy has been used. The IGBT open circuit fault presence can lead to a disturbance of the continuity of service. To overcome such a problem, a new voltage reconfiguration fault-tolerant technique for the twolevel three-phase inverter is integrated to our DFIG-WECS system with the aim to preserve its performance and stability under faulty operation. The paper also introduces a performance study of the resulting reconfiguration voltage under MATLAB-Simulink environment, and illustrates well the merits and efficiency of the proposed fault-tolerant method through the faulty condition.
Keywords: compensation voltages identification; voltage reconfiguration; DFIG; fault detection; fault–tolerant; IGBT open-circuit fault; inverter; MPPT; wind turbine.
Modelling, bifurcation analysis and multistability of a new 4-D hyperchaotic system with no balance point, its master-slave synchronisation and MultiSim circuit simulation
by Sundarapandian Vaidyanathan, Khaled Benkouider, Aceng Sambas
Abstract: This research manuscript focuses on the modelling, bifurcation analysis and multistability of a new four-dimensional hyperchaotic system with no balance point. The proposed hyperchaos system has four quadratic nonlinear terms and three parameters. We perform a detailed bifurcation analysis of the hyperchaos system and illustrate that it has coexisting hyperchaotic attractors. The proposed system has no balance point and it is equipped with hidden attractors. Using integral sliding mode control, we derive control results for the global hyperchaos synchronisation of a pair of hyperchaos systems and Matlab plots are shown as illustrations. Finally, circuit simulations of the new 4-D hyperchaotic system with no balance point are carried out using MultiSim.
Keywords: bifurcation; hyperchaos; multistability; rest point; Lyapunov exponents; circuit simulation; sliding mode control; attractor.
Sensorless DC-link voltage regulation strategy for single-phase grid-connected solar photovoltaic systems
by Raja Owais, Sheikh Javed Iqbal
Abstract: A sensorless DC-link voltage control technique is proposed for a two-stage, single-phase grid-connected solar photovoltaic (PV) system. The developed control strategy is predicated on the fact that if PV power is directly fed into the grid, the DC-link voltage would stabilize. However, losses must be properly compensated to maintain a flat DC-link voltage profile. To account for the system losses, teaching-learning-based optimization (TLBO) is used offline. After the offline implementation, a regression model based on maximum power is built to estimate system losses. The regression model is applied to the system to predictively compensate for system losses. As a result, the DC-link control loop is avoided, and sensorless control is achieved. A unit-template-based control is used to synchronise the PV system with the grid. Simulation experiments using MATLAB/Simulink are used to verify the superiority of the proposed scheme over the conventional sensor-based technique.
Keywords: photovoltaic; sensorless control; TLBO; hysteresis-current control; unit-template; Lambert-W function.
Suppression of positive emotions during pandemic era: a deep learning framework for rehabilitation
by Ahona Ghosh, Sriparna Saha
Abstract: The rapidly growing popularity of computer and Android gaming in the pandemic era has led many researchers to focus on the emotional changes and adverse effects of gaming on a player. Sometimes games are played to give relaxation and help to recover from anxiety and stress, but some recent incidents are alarming since vigorous games playing can lead to anxiety and stress. This paper attempts to overcome the shortcomings of existing literature by a novel approach to detect emotions acquired from electroencephalographic signals using a deep learning algorithm with high accuracy of emotion recognition. The feature extraction and classification algorithms individually outperformed the existing ones in the related area, and the combination of them also has shown better performance than the state-of-the-art literature. Outcomes of the proposed framework have shown its wide range of applicability from parental control to cyber security by emotion detection and can help in providing rehabilitation.
Keywords: deep learning; electroencephalogram; support vector machine; convolutional neural network; Chebyshev type I filter; cognitive rehabilitation; emotion management; pandemic era; circumplex model; Android game.
Control points searching algorithm for multiple mobile robots
by Imen Hassani, Imen Maalej, Rekik Chokri
Abstract: In the robotic field, an important issue is to find the adequate path for the mobile robot navigation free from collision. This paper deals with the development of the navigation strategy for multiple mobile robots, and investigates two types of navigation strategy: navigation with one robot and navigation with two robots. The proposed algorithm is based on searching for the value of the visibility condition of the trajectory to detect the collision problem. Then, control points are selected to find the safe path for navigation. On the other hand, a sliding mode controller (SMC) is used for tracking the planned trajectory. Hence, to improve the performance given by the SMC, a fuzzy logic control approach is applied for tracking the desired trajectory. Finally, simulation results are given to highlight the applicability of the proposed approach.
Keywords: mobile robot; autonomous navigation; control points; sliding mode; fuzzy logic.
Optimising processes and generating knowledge by interpreting a new algebraic inequality
by Michael Todinov
Abstract: This paper focuses on optimising processes and generating knowledge based on interpreting a new algebraic inequality. The inequality permits segmentation or aggregation of controlling factors and can be used to generate new knowledge and optimise processes and systems in virtually any domain of science and technology. The proposed method consists of interpreting the left- and the right-hand side of the proposed algebraic inequality as the outputs of two alternative processes delivering the same required function. In this way, the superiority of one of the competing process designs is established. The proposed method opens wide opportunities for enhancing the performance of processes and systems and is very useful for design in general. An interpretation of the new inequality yielded a strategy for minimising the amount of pollutants released from an industrial process. An alternative interpretation of the same inequality established that the deflection of n elastic elements connected in series is at least n^2 times larger than the deflection of the same elements connected in parallel, irrespective of the individual stiffness values of the elements. In addition, an alternative interpretation of the new inequality revealed a counter-intuitive result concerning picking a winning lottery ticket from a randomly selected urn. Finally, the paper also features novel interpretations of algebraic inequalities related to improving reliability by increasing the level of balancing. This is done primarily by assessing the probability of selecting items of the same variety and determining the lower and upper bounds of this probability.
Keywords: algebraic inequalities; interpretation of algebraic inequalities; reducing the amount of pollutants; increasing the level of balancing.
Volterra series based nonlinear system identification methods and modelling capabilities
by Gargi Trivedi, Tarun Rawat
Abstract: The current study discusses Volterra series based non-linear system models such as Taylor series, Time-delay neural network (TDNN) and Non-linear autoregressive model (NARX). The study aims to construct a truncated second-order Volterra model that can be used to identify non-linear systems and compare its performance to that of a TDNN using benchmark cases. The feasibility of feedback and feedforward networks is evaluated using a dataset of cortical responses evoked by wrist joint manipulation. It is observed that TDNN is a mathematical model with more customizable parameters and require less computation time than Volterra system with particle swarm optimization (PSO). Also, open-loop connections with less a-prior system assumptions, such as Volterra can estimate 42% of wrist dynamics and closed-loop connections like NARX model can estimate 93% of complex non-linear dynamics.
Keywords: Volterra; TDNN; NARX; model structure.
Investigation of transverse vibration suppression of hoisting catenaries in mine hoists by virtual prototype and uniform design
by Yu Zhu, Tong Xu, Jiannan Yao
Abstract: In the hoisting process of an ultra-deep mine shaft, the winding movement of ropes on the Lebus drum will cause transverse vibrations of the catenaries, leading to intense swing of the rope and even winding confusion. In this paper, the transverse vibration virtual model of the catenary was firstly established using bushing sleeve force method in ADAMS, then the displacement response of the catenary under the drum excitation was obtained. Secondly, the correctness of the established model was verified by comparing the numerical simulation results to the ADAMS model results. Finally, a spring-damper device was proposed to further suppress the transverse vibration of the catenary based on the boundary control. The reasonable values of the stiffness and damping parameters of the damper were obtained by applying the uniform experimental design. This study provides a theoretical support for the reduction of the transverse vibration of hoisting catenary.
Keywords: hoisting catenary; transverse vibration; virtual prototype; spring-damper device; uniform design; vibration suppression.
Generating synthetic wind speed scenarios using artificial neural networks for probabilistic analysis of hybrid energy systems
by Jun Chen, Junhui Zhao
Abstract: Hybrid energy systems (HES) have been proposed to include and co-optimize multiple energy inputs and multiple energy outputs to enable increasing penetration of clean energy such as wind power. To optimize the system design, extensive data sets of renewable resources for the given location are required, whose availability may be limited. To address this limitation, this paper proposes an innovative methodology to generate synthetic wind speed data. Specifically, artificial neural networks are adopted to characterize historical wind speed data and to generate synthetic scenarios. In addition, Fourier transformation is used to capture the characteristics of the low frequency components in historical data, allowing the synthetic scenarios to preserve seasonal trend. The proposed methodology enables the possibility of Monte Carlo simulation of HES for probabilistic analysis using large volumes of heterogeneous scenarios. Case study of probabilistic analysis is then performed on a particular HES configuration, which includes nuclear power plant, wind farm, battery storage, electric vehicle charging station, and desalination plant. Wind power availability and requirements on component ramping rate are then investigated.
Keywords: artificial neural networks; Fourier transformation; hybrid energy systems; synthetic scenarios; wind energy.
Unknown input observer For linear singular systems with variable delay: the continuous and the discrete time cases
by Fatma Hamzaoui, Malek Khadhraoui, Hassani Messaoud
Abstract: In this paper, we propose a new approach to design a functional observer for linear singular systems with unknown inputs. Variable-Time delay is present in both state and input vectors. The proposed observer is developed in time and frequency domains. The time domain observer design is obtained for both continuous and discrete time systems. The procedure design is based on the unbiasedness of the estimation error dynamic of the observer using Lyapunov functional. The problem is solved by means of Linear Matrix Inequalities to nd the optimal gain implemented in the functional observer design. Frequency domain procedure is derived from time domain results, where we propose a suitable co-prime Matrix Fractions Descriptions. The main interest is that the proposed algorithm estimates both a functional state and the unknown input dependently from the considered delay in time and frequency domains. The proposed observer proves its eectiveness on a given numerical example.
Keywords: continuous and discrete time cases; time and frequency domains; unknown additional inputs; functional observers; variable delay; MFD; LMI.
Adaptive PID computed-torque control of robot manipulators based on DDPG reinforcement learning
by Akram Ghediri, Kheireddine Lamamra, Abdelaziz Ait Kaki, Sundarapandian Vaidyanathan
Abstract: This paper presents a design of an adaptive PID gain tuning based on Deep Deterministic Policy Gradient reinforcement learning agent for PID Computed-torque control of robot manipulators, taking the presence of unmodeled dynamics and external disturbances into consideration. The proposed approach adaptively computes the outer-loop PID controller gains, that minimize trajectory tracking errors and reject disturbances, with the closed-loop dynamics remain stable. Since the control scheme requires the knowledge of the robots dynamics, both kinematic and dynamic equations of n-link serial manipulator are developed. The agent is implemented on UR5e robot manipulator model, using the most valid dynamic and kinematic parameters provided by the manufacturer and related works. Simulation results show that the proposed approach is robust against bounded internal and external disturbances, and achieves a good trajectory tracking performance, due to the adaptability of gain tuning over the conventional PID controller.
Keywords: UR5e robot manipulator; adaptive PID control; computed-torque control; DDPG reinforcement learning; trajectory tracking; disturbance rejection.
Advanced genetic algorithm based PID controller for air levitation system
by D.P. Gaikwad, B. Patil, L. Patil
Abstract: In industrial control systems, PID controllers are being widely used due its simple working principles. Many control and instruments engineers and operators use PID controllers in daily life. PID controllers allows for many variations which can cope with a wide range of systems and conditions. For increasing performances of PID controller, fine tuning of its parameters are required. Many authors have used different optimization algorithms to tune parameters of PID controllers. These optimization algorithms offer less performance. In this paper, the fine-tuned PID controller is proposed for the air levitation system. Advanced genetic algorithm is used for tuning parameters of PID controllers. For demonstration of efficiency and applicability of the proposed PID controller, simulation based experimentations have conducted. The proposed PID design method has linked with other three optimization techniques. Ant colony optimisation, particle swarm optimisation and fuzzy logic are used for performance comparison of advanced genetic algorithm based PID controllers. In experimental results, we have got very small values of IAE, ISE and ITAE using the proposed method. It indicates that the proposed PID design method offers improved performance over the other three optimisation based PID design methods and other existing methods.
Keywords: PID; integrating; process model; tuning; stability.
An enhancement in parallel cascade scheme for non-minimum phase system
by Manish Yadav, Hirenkumar Patel
Abstract: This paper aims to control non-minimum phase (NMP) systems with dead time in the existence of uncertainty and disturbances. The parallel cascade control is utilized to motivate such problems, especially for slow process dynamics and actuator nonlinearities. The novelty lies in this work, combination of a higher-order fractional-filter with an inverse response and dead-time compensator in the Internal Model Control (IMC) framework for designing the outer loop controller. The inner loop controller assumes the standard IMC controller. This modified structure is offered refinement in the gain margin for better robustness. The Riemann sheet principal is used to stability investigation of factional quasi characteristic polynomial arises from the non-minimum phase systems with dead time. Further, a robustness test is also carried out via sensitivity analysis. The efficacy of the suggested method is illustrated via two case studies.
Keywords: non-minimum phase; parallel cascade control; IMC controller; robustness.
Kharitonov polynomial based interval reduced order modelling of Cuk Converter
by V.P. Meena, V.P. Singh
Abstract: This paper proposes method of reduced order modelling for the Cuk converter using a state-space-averaging (SSA) technique, in which combined state-space description is obtained and output to control for the Cuk converter transfer function is determined. However, there may be some variation in parameters of system owing to uncertainties and imperfect modelling that are addressed using interval modelling. Thus, the obtained interval model is reduced further using Kharitonov polynomials. Routh-Pad
Keywords: Cuk converter; interval modelling; interval systems; Kharitonov polynomials; order reduction; parametric uncertainty; state-space averaging.
A TCPN-based model for testing distributed systems with timing constraints
by Salma Azzouzi, My El Hassan Charaf
Abstract: This paper focuses on extending the basis of distributed testing to address the testing process for time-sensitive distributed systems. In this context, we present an alternative approach to manage the problems that occur in this area, commonly referred to as controllability, observability and synchronisation issues. The key point of the proposed study is to define the activities of each tester through a set of timed distributed testing rules. Thus, we present our algorithm for the generation of such rules. Each rule is handled as a data structure comprising the data to be sent or received, the guard to be controlled and the set of clocks to be updated at the end of each transition. Afterwards, we show how we can use a timed coloured Petri nets model to cope with the complex tasks of the monitoring of testers in the distributed test context. The simulation results revealed the effectiveness of our approach in providing correct execution of the system actions and also show how the response time of each tester can be improved by considering the temporal constraints. Indeed, the test becomes non-blocking and stops immediately by returning a false verdict if the temporal constraints have not been met.
Keywords: distributed testing; controllability; observability; synchronisation; timed coloured Petri net.
A fixed wing UAV with VTOL capabilities: design, control and energy management
by Luca Pugi, Alberto Mela, Alberto Reatti, Armando Casazza, Roberto Fiorenzani, Giuseppe Mattei
Abstract: There is an increasing interest for UAVs (Unmanned Aerial Vehicles) with mixed, multi-rotor propulsion layouts able to assure desirable feature of both fixed wing systems (efficiency high cruising speed autonomy) and capabilities of rotating wing ones (hovering, vertical take off and landing capabilities). This work investigated a mixed propulsion layout with five electric propellers fed by an hybrid energy management system able to assure an higher autonomy respect to a pure electric solution. The proposed system is investigated through the development of a model able properly to simulate complex interactions arising between different propulsion, control and energy management subsystems. In this way, it was possible to propose and calibrate an efficient energy management policy and to evaluate how different transition policies between hovering and fixed wing cruising should affect involved energy consumptions. Finally, the proposed model was used to simulate a complex mission profile in order to both verify manoeuvring capabilities of the system and predict energy consumption. At the end it was possible to verify not only the feasibility of the proposed solution with respect to the completion of a complex mission profiles but also the potentialities and utility of the adopted simulation models.
Keywords: mechatronics; UAV; VTOL; hybrid propulsion.
Orientation effect on the stability conditions of fronts in a liquid and porous medium
by Hamza Rouah, Ahmed Taik
Abstract: In this article, we have studied the influence of orientation on the conditions of stability of the reaction fronts in two cases: the first case where the liquid monomer is converted into solid polymer in a liquid medium and the second case where the monomer and the polymer are liquid in a porous medium. Zeldovich and Frank-Kamenetskii method are used to perform the asymptotic analysis taking the inverse of Zeldovich number as a critical parameter. The linear stability analysis is fulfilled to investigate the resulting interface models for both cases. The dispersion relation obtained for both cases is solved numerically and then the conditions for cellular and oscillatory instability are determined. We showed that the angle of inclination affects the conditions of thermal and convective instability in a liquid and porous medium.
Keywords: frontal polymerisation; reaction-diffusion equations; reaction fronts; stability analysis.
Power quality assessment and power quality improvement in a hospital facility
by Sachin R, Nagesh H B
Abstract: The power quality (PQ) assessment is carried out at a KIMS hospital facility by conducting the harmonic study as per IEEE3002.8 guidelines. As per PQ data obtained, the hospital facility is experiencing a significant number of PQ disturbances and violating IEEE519 limits for PQ in special application systems. It is found necessary to use a compensator for PQ improvement. A cost-effective compensator dynamic voltage restorer (DVR) is designed to improve the PQ in the hospital facility. The designed DVR controller model is tested in the RT-LAB hardware-in-loop real-time simulation platform using OPAL-RT with multicore FPGA processors to validate the compatibility of the designed controller to be applied for real system implementation. Thus, PQ assessment helps in the appropriate design of compensating controllers to suit practicality. A comparative study of the choice of designed DVR controllers, such as PI, Fuzzy, ANN, ANFIS-PI optimized and ANFIS-Fuzzy optimised controllers tested in real-time for different cases and for different loading conditions is summarised.
Keywords: adaptive neuro fuzzy inference system; artificial neural network; dynamic voltage restorer; fuzzy logic controller; proportional integral controller; power quality.
Identification of the most conservative stability bounds for a class of multi-rate haptics controllers
by Suhail Ganiny, Majid H. Koul, Babar Ahmad
Abstract: This work identifies the most conservative stability bounds applicable to a class of multi-rate haptic controllers that involve sampling of a single state variable at two distinct rates for rendering a virtual wall. In particular, the uncoupled stability boundaries of a dual-rate haptic controller have been determined. In contrast to the prior research, the current work extends the scope in terms of identification of the impedance parameters that establish the worst-case stability limits that are closest to the experimental results. Our analysis reveals that the transformation sequence ZOH-Tustin-ZOH yields the most conservative, while the half-sample delay approximation approach yields the least conservative estimates of the stability bounds. Specifically, the relative root-mean-square error (RRMSE) between the experimental outcomes and the results predicted by the ZOH-Tustin-ZOH transformations varies between 0.54-0.83, while for the half-sample delay approximation it varies from 1.21-3.2.
Keywords: multi-rate controller; impedance haptic interfaces; stability bounds.
Gear faults identification based on big data analysis and Catboost model
by Yongsheng Qi, Xiaoda Zhang, Jianxin Zhang
Abstract: The gear faults identification based on big data analysis and Catboost is investigated. The big data sets with nine and ten features for five gear faults are constructed, respectively. The Catboost model based on the above two data sets is constructed and trained. The testing results show that Catboost, XGBoost, and LGBM models based on the data set with ten features are better than ones with nine features, and the fault identification accuracy and time obtained by Catboost are better than the other two models. By calculating the influence of features to the identification results, it can be found that four features play the crucial roles. The Catboost based on the data set with the above four characteristics and five faults is verified to achieve identification accuracies and times of are 100% and 680 s, respectively, which are better than ones obtained by using XGBoost and LGBM.
Keywords: gear faults identification; big data analysis; Catboost; classification prediction; feature importance.
Displacement tracking of uncertain nonlinear cardiovascular muscle using Lyapunov function with disturbance observer-based control
by Soumyendu Bhattacharjee, Sourish Sanyal, Madhabi Ganguly, Aishwarya Banerjee, Biswarup Neogi
Abstract: The nature of cardiovascular muscle was modelled primarily using some very basic mechanical elements, among which some were considered as a linear element so that the design can be easily understood. Owing to non-zero reaction time of cardiovascular muscle, the overall model becomes nonlinear. Nonlinear behaviour of the proposed model has been explained using some very common types of nonlinearity, such as dead-zone and saturation. Robust control of uncertain nonlinear human cardiovascular muscle dynamics is investigated using Lyapunov stability theory along with an observer-based control system. In this work, a nonlinear controller has been designed in the absence of uncertainty and any other disturbances to track the displacement of muscle dynamics. To achieve good tracking performances, stability analysis of the plant-sensor system has successfully been done in more than one time considering different situations. Finally, an asymptotical stability has been found towards the proposed nonlinear system.
Keywords: DOBC approach; Lyapunov stability; non-linearity; robust control design; uncertainity.
An adaptive disturbance multi-objective evolutionary algorithm based on decomposition
by Yanfang Shi, Jianguo Shi
Abstract: In solving multi-objective optimisation problems, the uniformly distributed weight vector of decomposition based multi-objective evolutionary algorithm (MOEA/D) is not completely suitable for the non-uniformly distributed Pareto front (PF). In order to solve the situation above, this paper proposes an Adaptive Disturbance Multi-Objective Evolutionary Algorithm based on Decomposition (AD-MOEA/D), the proposed algorithm introduces the disturbance individuals and disturbance weight vectors during the evolution. The disturbance individuals maintain the population diversity and improve convergence accuracy. The disturbance weight vectors assist the weight vectors to adjust adaptively and improve the distribution of PF. Besides, both disturbance individuals and disturbance weight vectors are produced according to the actual evolution, which will not participate in evolution when it is not necessary. The experimental results on multi-objective test functions show that the PF optimised by AD-MOEA/D has better convergence and distribution.
Keywords: multi-objective evolutionary algorithm; disturbance individuals; disturbance weight vectors; decomposition.
Strong Wolfe condition based variable stacking length multi-gradient parameter identification algorithm
by Yiqiao Shi, Shaoxue Jing
Abstract: This paper considers the acceleration of the gradient algorithm for the linear models. The traditional stochastic gradient algorithm requires less computation, but it converges to the true parameter slowly. To accelerate the gradient algorithm, a novel gradient algorithm using several gradients is proposed. One important issue of the proposed algorithm is how to determine the stacking length. The stacking length defines the number of gradients used in each recursion. A variable stacking length based on the strong Wolfe condition is presented to enable the algorithm to converge faster. The stacking length obtained by using the strong Wolfe condition can ensure that the proposed multi-gradient algorithm converges faster. Several experiments are made to validate the proposed algorithm.
Keywords: parameter estimation; stochastic gradient; multi-gradient; strong Wolfe condition; convergence speed.
Four generations of control theory development
by T.C. Yang
Abstract: In our control community, in particular in our teaching, we often use the terms classical control theory and modern control theory. History moves forward. The word modern here is not appropriate. Todays modern is futures classical. Nevertheless, behind the ambiguous words they are meaningful terms: transfer function based for classical control, and state-space based for modern control. Looking back and forward, and to give an overall overview, this short article presents an opinion that control system study up to date can be divided into four generations; namely, 1) transfer function based; 2) state-space based; 3) networked control systems; and 4) control in the new AI era.
Keywords: control theory development; four generations; control in the new AI era.
3D indoor reconstruction using Kinect sensor with locality constraint
by Peng Zhu, YanGuang Guo
Abstract: In this paper, an indoor 3D construction is proposed based on RGB-D measurement. It is intentionally designed to solve the traditional issues, such as cloud registration inaccuracy, large computational time. Firstly, potential candidates are extracted by Harris detector, and the SURF method is used to generate the feature descriptors. Afterwards, the correct functional match is selected by RGB and depth measurements with neighbouring constraint. Lastly, 3D clouds are formed through graphical optimization. In the experiment, the RGB-D sensor is rigidly fixed on the mobile platform to reconstruct the indoor 3D scene, which shows comparable performance in terms of computational time and accuracy.
Keywords: RGB-D; 3D indoor reconstruction; Kinect; point cloud; SURF method.
HMM-based IMU data processing for arm gesture classification and motion tracking
by Danping Wang, Jina Wang, Yang Liu, Xianming Meng
Abstract: This paper investigates Inertial Measurement Unit (IMU) data processing methods for human gesture classification and arm motion tracking in Wireless Body Sensor Network (WBSN). The method is adopted that consists of two main stages. In the training stage, the supervised learning method is adopted to obtain the HMM model and the Viterbi algorithm is used to obtain the optimal hidden state sequence in the testing stage. HMM also complements the intuitional evaluation for arm motion recovery. We take advantage of the twists and exponential maps to recover the arm motion process. In addition, visual tracking device-VICON is used to validate the accuracy of the inertial tracking system. The experimental results show that the HMM algorithm gesture classifier achieves up to 96.63% accuracy on five commonly used arm gestures and visual assisted tracking outcomes verify the robustness and feasibility of the IMU tracking device.
Keywords: gesture classification; arm motion tracking; inertial measurement unit; computer network; vision motion tracking.
Multi-robot Cooperative Exploration Based on Geometric-topological Scene Map
by Yanqing Wang, Yongquan Li, Wenjun Lu
Abstract: In view of the advantages and disadvantages of node geometry information and characteristic scene information, this paper proposes multi-robot cooperative exploration based on geometric-topological scene map. Our algorithm adds the geometric information of nodes to the feature scene map, so as to detect and recognize the node loop. Considering the spatial relationship between a single node and different nodes, we use the Hidden Markov Model (HMM) to locate nodes accurately, and use the geometrical-topological feature scene matching method to perform map fusion. The local map fusion of each robot is turned into a problem of geometrical-characteristic scene matching. Finally, the multitask assignment method of market method is used to realize the multi-robot cooperative exploration task on the simulation platform.
Keywords: market approach; geometric-feature scene matching; loop-back detection; HMM; map fusion; node geometry information; characteristic scene map; contract network agreement; multi-robot cooperative; SuperPoint scene feature information.
A novel adaptive variable speed control strategy for wound rotor induction motors
by Dieudonné Ekang, Donatien Nganga-Kouya, Aime Francis Okou
Abstract: A new approach is proposed for the design of an adaptive variable speed control for an induction motor. This design approach is based on a new model for induction motors in the (/) reference frame. The model state variables are constant in steady state and therefore enable the application of adaptive backstepping control design techniques to find controller equations and adaptation laws that insure that the rotor speed and flux track their reference values despite signification changes in machine resistances and inductances due to temperature and magnetic saturation. The proposed controller is tested in simulation. Results show robust steady state and transient performances.
Keywords: wound rotor of induction motor; adaptive backstepping control; speed control; rotor flux control; stability analysis.
Map matching navigation method based on scene information fusion
by Yanqing Wang, Yongquan Li, Chuang Xu, Chaoxia Shi
Abstract: Because of the similarity of the geometric information of nodes in traditional topological maps, the location error of nodes using geometric information will be larger and larger, which will reduce the precision of map creation. In this paper, a new representation method of the topological map environment is proposed, in which scene information is added to the traditional topological map; this information is used for scene matching and node localisation to provide more reliable support for map creation. Moreover, in the campus environment, the image information of each road is used to calibrate and generate the data set. The experiment proves that the method can effectively guarantee the robot to complete the exploration task.
Keywords: superpoint network; topological map; bag of words model; image matching; self-supervised; simulation platform; multi-robots; feature extraction; environmental exploration; continuous exploration.
Experimental parameter estimation methodology based on equivalent output injection
by David Rosas, Karla Espinoza, Karla Velazquez
Abstract: This work proposes a methodology to estimate parameters for linear and nonlinear dynamical systems, with partial state measurement, that satisfy the property of parameter linearity. This methodology is experimental, off-line, and recursive. It uses discontinuous state observers to estimate all state variables and the disturbance terms needed in the estimation processes. Because the equivalent output injection corresponds to the disturbances produced by the parameter uncertainties, the methodology allows us to obtain the best parameter estimation by minimizing an index related to the power of the equivalent output injection; a smaller value represents a better estimation. With this parameter estimation, we can establish a model that facilitates the design and implementation of many control algorithms, including robust controllers. We validate the methodology through numerical simulations and experiments with linear, nonlinear, and discontinuous systems. Based on the experimental results, we conclude that the proposed algorithm's performance is better than other methodologies.
Keywords: identification; modelling; equivalent output injection; discontinuous observers; least squares algorithm.
A continuous-time fault-tolerant predictive control approach for wind turbines
by Rime Elhouti, Selma Sefriti, Ismail Boumhidi
Abstract: This paper present a developed fault-tolerant control technique for the wind turbine system. The proposed approach is a combination of an additive sliding mode with a predictive model based controller. The online continuous-time model predictive controller (CMPC) is designed to be a nominal controller for maximum power tracking and tolerates the actuator loss of faults, while the additive term is responsible for ensuring more robustness with respect to the various actuator faults. The considered controller uses the full state of the system, so, a robust observer is proposed for estimating the full state needed in the control law. The obtained results in the simulation part show the efficiency of the proposed method for both maximum power tracking and handling the system actuator faults.
Keywords: predictive control; fault-tolerant control; sliding mode observer; wind turbine.
Parameter identification for a model of gas exchange dynamics during cycling
by Nadia Rosero, Maxime Chorin, John Jairo Martinez Molina
Abstract: This paper presents the modelling and parameter identification of the gas exchange dynamics during cycling. A discrete-time linear parameter-varying model is proposed, which relates the dynamics of oxygen consumption and carbon dioxide production with the developed pedal power. A state-dependent non-linear function is used for modelling the excess carbon dioxide production. The approach proposed for parameter identification is based on specific exercise scenarios tailored to the considered individual, which narrows the data acquisition process. The parameter identification process is performed as a solution of a sequence of non-linear unconstrained optimization problems using measured data from different cycling scenarios. An illustration of the methodology used for the identification and the validation of the model is also presented.
Keywords: model identification; gas exchange; cycling; model checking; physiology.
Optimal control strategies-based maximum power point tracking for photovoltaic systems under variable environmental conditions
by Sally Abdulaziz, Galal Attlam, Gomaa Zaki, Essam Nabil
Abstract: To increase the efficiency of photovoltaic (PV) array output under variable environmental conditions, maximum power point tracking (MPPT) of the solar arrays is needed. This paper proposes Fuzzy Logic Controller (FLC) based MPPT, Artificial Neural Network (ANN) based MPPT, Neuro-Fuzzy (NF) based MPPT, Particle Swarm Optimisation (PSO) based MPPT, and Cuckoo Search (CS) algorithm based MPPT to combine an adaptive controller and an optimisation, to guarantee global stability and a constant settling time for all operation conditions. This combination enables an increase in the power generated in comparison with conventional MPPT techniques. Simulation results show that the proposed photovoltaic/storage generator is able to supply the suggested dynamic loads under different conditions, and achieve good performance. It is also noticed that operating the photovoltaic array based on maximum power point tracking conditions gives about 43% extra power generation than in the case of normal operation.
Keywords: DC_DC power converters; fuzzy control; fuzzy neural controller; maximum power point trackers; photovoltaic systems; particle swarm optimisation; renewable energy sources.
Impulse-based controller synthesis for a class of polynomial systems
by Qian Ye
Abstract: This paper develops a time-triggered impulsive control strategy for stabilisation problem of a class of polynomial systems. The proposed hybrid control method combines time-triggered impulsive control and state-feedback control. To obtain the control feedback gain, the matrix sum-of-squares programming is employed to solve the feasible solutions of constructed positive polynomial problems. Compared with a single time-triggered impulsive control or the pure state-feedback control, the proposed control method can effectively combine the advantages of the two control methods. In particular, the design of the controller is more flexible and the control effect is greatly improved. Finally, two simulation examples including the Chua's chaotic system are provided to illustrate the effectiveness and superiority of the proposed method.
Keywords: polynomial system; impulse-based control; state-feedback control; sum-of-squares programming.
Boundary control of a flexible beam with output constraint under saturation input
by Chuyang Yu, Haojie Lin, Xuyang Lou, Jiajia Jia
Abstract: This paper focuses on the problem of saturated boundary control for a flexible beam system with unknown disturbances and output constraint. It is aimed to suppress the vibration of the flexible beam system with unknown disturbances and input saturation, and make the pitch of the boundary output remain in the constraint. In order to satisfy the system constraint, an auxiliary term is established and a barrier Lyapunov function is applied. Two boundary controllers with a barrier term and an auxiliary term are constructed to prove the uniform ultimate boundedness of the state of the flexible beam. Numerical simulations are given to verify the effectiveness of the proposed control methods.
Keywords: flexible beam system; input saturation; output constraint; boundary control; unknown disturbance.
Guidelines for choosing hyperparameters of echo state networks for system identification: Two case studies
by Thiago Ushikoshi, Luis Aguirre
Abstract: Echo state networks (ESN) can be used to model dynamical systems in the context of reservoir computing using standard regression algorithms, which is one of its main advantages. However, there are some hyperparameters that need to be carefully chosen and there is no general recommendation on how to perform this important step. After setting the ESN paradigm for system identification, this paper describes the choice of hyperparameters in the context of two case studies: one using experimental data of a pilot heater and the other using the Duffing-Ueda oscillator with chaotic dynamics. The main findings are: (i) ESNs can reproduce the chaotic regime of the Duffing-Ueda oscillator for a specific region on the hyperparameter space, (ii) some hyperparameters may not be critical from a statistical perspective but can still drastically affect the dynamical regime, and (iii) the ESN initialization is not critical when the hyperparameters are adequately chosen.
Keywords: reservoir computing; echo state networks; ESN; guidelines for echo state networks; hyperparameters of echo state networks; nonlinear models; system identification; dynamical system identification; modelling; chaotic oscillators.
Joint variable and variable projection algorithms for separable nonlinear models using Aitken acceleration technique
by Lianyuan Cheng, Jing Chen, Yingjiao Rong
Abstract: This paper proposes a joint variable based gradient descent algorithm (Joint-GD) and a variable projection (VP) based gradient descent algorithm (VP-GD) for separable nonlinear models. The VP algorithm takes advantage of the separability property of variables to reduce the dimensionality of the parameters, which makes the convergence rates faster. In order to speed up the convergence of the gradient descent algorithm, the Aitken acceleration technique is introduced in the algorithms, which is second-order convergent. Moreover, the Aitken based methods are robust to the step-size, therefore they can be widely used in engineering practices. The numerical simulation shows the effectiveness of the proposed algorithms.
Keywords: variable projection algorithm; joint variable algorithm; gradient descent
algorithm; separable nonlinear model; Aitken acceleration technique.
Nonlinear system identification using butterfly optimisation algorithm and Hammerstein model
by Sandeep Singh, Tarun Kumar Rawat, Alaknanda Ashok
Abstract: This paper focuses on the nonlinear system identification using butterfly optimisation algorithm (BOA) optimised with adaptive Hammerstein model which is the cascade of nonlinear second-order Volterra (SOV) and linear finite impulse response (FIR) systems. Generally, gradient-based methods have been applied for solving such problems. However, these methods may face the problem of getting trapped in local minimum solution. In this paper, a novel butterfly optimisation algorithm is used to identify the nonlinear system by using three different models namely, Hammerstein model, memoryless polynomial nonlinear (MPN)-FIR and SOV models. Furthermore, to measure the accuracy of the employed BOA, mean square error, coefficient estimation and convergence speed are considered. To prove the efficacy of the proposed BOA, the simulated results has been compared with that of the antlion optimisation algorithm and dragonfly algorithm. The simulated results confirm that Hammerstein model with SOV-FIR optimised with BOA is able to outperform the other models and algorithms.
Keywords: nonlinear system identification; Hammerstein model; meta-heuristic algorithms; butterfly optimisation algorithm; antlion optimisation algorithm; dragonfly algorithm.
Stochastic pointwise second-order maximum principle for optimal continuous-singular control using variational approach.
by Nour El Houda Abada, Mokhtar Hafayed
Abstract: In this paper, we establish a second-order necessary conditions for optimal continuous-singular stochastic control, where the systems is governed by nonlinear controlled It
Keywords: optimal control; stochastic continuous-singular control; pointwise second-order necessary conditions; variational method.
New decentralized control based on T-S fuzzy logic approach of an electrical wind-source integrating grid
by Mohsen Ben Ammar, Wissem Bahloul, Mohamed Ali Zdiri, Hsan Hadj Abdallah
Abstract: The present work is designed to advance a new methodology that allows the implementation of a decentralized control system within a multi-machine grid. The design consists in determining the grid Thevenin equivalent, as conceived by each generator node. Such a process should help in transforming the grid into n machines, whereby, each single machine must be connected to an Infinite Bus (SMIB). Relying on a Blondel diagram, we have been able to define a complete model relevant to each of the grid-associated machines. Given the system non-linearity, application of a Takagi-Sugeno (T-S) fuzzy logic turns out to yield satisfactory results. Noteworthy, is that the investigated test grid has been equipped with a wind turbine, while the considered disturbances are power injected variations by this renewable source. The simulation results, was implemented on the 9-node Western System Coordinating Council (WSCC) grid test, proved the remarkable robustness of the applied control in terms of disturbance reduction.
Keywords: electrical grid; wind energy; AVR and PSS controllers; Thevenin model; decentralised control; fuzzy logic; electrical machine; WSCC 9 nodes.
Identification of the three-axis pedestal using Euler-Lagrange method using mathematical approach
by S.Mohammadreza Ebrahimi, Behrooz Rezaei, Mehdi Tavan
Abstract: Pedestals are considered as an applicable tool to hold and rotate various instruments. One of their important applications is carrying and rotating antenna to precisely track satellites in space so that it can be possible to receive signals sent from the satellites and send information to the satellites as well. The quality of received and sent signals depends on different factors among which selecting an appropriate pedestal and a robust controller play a vital role. Selecting an inappropriate pedestal can lead to losing a part of the information in keyhole of the pedestal. Also, choosing an unsuitable controller causes the pedestal cannot track the satellites precisely or conserve the stability of the nonlinear pedestal system, which results in decreasing performance of the pedestal. Three-axis pedestal is one of the best pedestals ever constructed and can be used in diverse fields due to not having any keyhole. However, it has not been proposed an accurate model for it until now. In this article, an accurate model of the three-axis pedestal system has been extracted. For extracting motion equations of the pedestal in which the Euler-Lagrange method has been used, both rotational movement and transitional movement have been considered to obtain a precise and comprehensive model of the system. Because of the inaccessibility of the actual model and also saving time and cost, the proposed model has been compared with the model simulated in SolidWorks and MATLAB software to carry out validation. The results of simulation and experiments have shown the validity of the modelling.
Keywords: mathematical model; mobile antenna; XYZ pedestal; modeling; validation; simulation;.
Sliding mode switched tracking control of space robot manipulator
by Juan Wang, Quanze Zhao, Liangliang Sun
Abstract: This paper investigates the tracking control of space robot manipulator by sliding mode switched method. According to the difference between space gravity environment and ground gravity environment, a sliding mode switching control strategy based on dwell time was proposed. Since the gravity environment of space robot manipulator is different, the space robot manipulator is modelled as a multi-mode switching system. It is divided into ground subsystem and space subsystem, and different sliding mode controllers are designed respectively. The stability of the switching system is proved by multiple Lyapunov function method and the trajectory track problem of space robot manipulator is realised in the framework of switching control. Finally, the simulation example shows the effectiveness of the proposed control method and the comparative simulation demonstrates the superiority of the proposed control method.
Keywords: space robot manipulator; switching control; sliding mode control; dwell time.
Identification of Hammerstein-Wiener time delay model based on approximate least absolute deviation
by Baochang Xu, Zhichao Rong, Yaxin Wang, Likun Yuan
Abstract: Nonlinearity, time delay and spike noises widely exist in industrial processes. Compared with the linear model, the typical nonlinear Hammerstein-Wiener (H-W) model can describe nonlinear characteristic of industry processes more accurately. In order to overcome the effect of spike noise on the identification results, we propose a stochastic gradient algorithm based on the least absolute deviation in this paper. To solve the non-differentiable problem of the least absolute deviation, an approximate least absolute deviation objective function is established by introducing a deterministic differentiable function to replace the absolute residual. Experiments show the proposed algorithm can suppress the influence of the spike noise on the identification results, and has high identification accuracy and strong robustness.
Keywords: Hammerstein-Wiener model; approximate least absolute deviation; stochastic gradient; spike noise; time delay.
Design of a PEM fuel cell powered autonomous quadcopter
by Soham Prajapati, S. Charulatha
Abstract: The purpose of this paper is to design a PID controller for a detailed nonlinear model of a custom-made quadcopter and utilize the power requirements in designing a proton exchange membrane (PEM) fuel cell system. The first principles methodology to commence with unmanned aerial vehicle architecture is delineated. The empirical actuator response of the brushless DC (BLDC) which further adds accuracy to the mathematical model of the system. The paper also presents simulations for nonlinear open loop and closed loop feedback with PID controller. The controller gains are obtained from the simulations for hover flight mode. The experimental implementation of battery powered quadcopter is validated with simulation results of PID control design. Finally, a mathematical model of a PEM fuel cell system for proposed quadcopter model is presented. The key limitation of this paper is the absence of experimental data for fuel cell model due their high cost and low-availability.
Keywords: drone CAD model; quadcopter dynamics; PID control; fitting thrust vs RPM curves; hover flight mode; PEM fuel cells; hydrogen powered; zero-emission technology.
Force/Position Control of Constrained Reconfigurable Manipulators with Sliding Mode Control based on Adaptive Neural Network
by Ruchika , Naveen Kumar
Abstract: A reconfigurable manipulator can achieve proficient end effector
and elongate workspace. However, deformable link causes frequent
changes in shape and therefore brings difficulties to model and
control the manipulator. In view of distinctive behaviour because of
bending operation, a sliding mode based mechanism with no prior
dynamic information is introduced for validated control operation.
The nonlinear terms are included in the sliding mode to improve the
convergence rate. Moreover, we show that fast terminal sliders
reinforce parametric uncertainty as compared to conventional
sliders. The neural network system is adopted for the estimation of
nonlinear components whereas the friction term and constraint force
of each joint are compensated with the help of adaptive control. The
Lyapunov theory proves the stability of a closed-loop system.
Finally, simulations are performed in a comparative manner with two
different configuration controls that will provide the benefit of
the design method.
Keywords: finite time convergence; RBF neural network; adaptive bound; reconstruction error; terminal sliding mode control.
Universal activation function for data-driven gait model
by Bharat Singh, Suchit Patel, Ankit Vijayvargiya, Rajesh Kumar
Abstract: Gait generation for the biped robot is a very tedious task owing to higher degrees of freedom and an uncertain environment. Deep learning approaches can be employed for the modelling of real human kinematics, which can be further applied as a reference to the biped robot. However, choosing the right activation function is a very challenging task. This research work proposed the universal activation function for the kinematic modelling which is adaptive in sense of application. Twenty-five different activation function from the literature is compared with the presented activation function in term of mean and maximum model prediction error along the gait trajectory. It shows that the universal activation function-based gait model outperforms others by large margins. Additionally, the parameter sensitivity of the presented activation function is discussed in detail. Furthermore, two cases of 5% and 10% variation in the input are analysed to evaluate the prediction ability of the developed gait model with a 95% prediction interval.
Keywords: gait model; activation function; prediction interval; data-driven; biped robot.
Extended linear quadratic regulator control and its application in trajectory following control of autonomous vehicles
by Jianwei Wu, Lin Chen, Yang Zhou, Beibei Sun
Abstract: Owing to the limitation that the linear quadratic regulator (LQR) method cannot consider the weight of input rate, we propose an extended linear quadratic regulator (ELQR) method, and further extend the application of the LQR. Considering that the standard Riccati equation cannot be obtained after adding the weight term of input rate in the quadratic performance index, it cannot be solved by the traditional matrix algebra equation method. Therefore, an optimisation model is constructed, and is solved by the genetic algorithm. A simulation example from the trajectory following control for autonomous vehicles, which need to consider the limitations on the angular velocity of front steering to ensure safe driving, is given to illustrate the effectiveness of the ELQR in this paper. The results show that both the LQR and ELQR can achieve the expected control effects. Compared with the LQR, the ELQR considering the weight of input rate has obvious advantages, which avoids exceeding the limitations on the angular velocity of front steering and thus improves safety and comfort of driving.
Keywords: ELQR; LQR; weight of input rate; genetic algorithm; algebraic Riccati equation.
Equal-weight and rank-sum-weight based systematic diminution of higher order continuous systems using grey-wolf-optimization
by Umesh Kumar Yadav, Naresh Patnana, V. P. Meena, V. P. Singh
Abstract: The order-diminution techniques are adopted in various fields of engineering and applications to reduce the order of systems from higher order to desired lower order. In this research proposal, diminution of higher-order-continuous-systems (HOCSs) is done by incorporating systematic procedures for determination of weights. The errors between time-moments and Markov-parameters of HOCS and desired reduced-order-model (ROM) are utilized to frame the objective function. In the objective function, associated weights are determined using systematic procedures. The systematic procedures exploited in this work are equal-weight method and rank-sum-weight method. The minimization of framed objective function is done by grey-wolf-optimization algorithm. The effectiveness and superiority of proposed method is claimed with the help of tenth-order and seventh-order system by considering as test cases. The comparative analysis is done by tabulating the time-domain-specifications and error-indices. The responses of the HOCS and ROMs are also presented, to prove the efficacy and effectiveness of the proposed method.
Keywords: Higher-order-continuous-systems; Grey-wolf-optimization algorithm; Order-diminution; Reduced-order-model; Systematic procedures.
A collaborative channel gain and delay estimation algorithm based on dynamic state space model
by Danping Wang, Yang Liu, Yanhui Wang
Abstract: In order to overcome the serious impact of the presence of fading and sensing delay factors in the channel on the performance of spectrum sensing, a novel joint channel gain and sensing delay algorithm is proposed in this paper. A dynamic state space model is constructed to establish the relationship between the PU state, the dynamic fading channel state and the perceived delay. The channel gain can be obtained by means of hidden Markov chain and maximum a posteriori probability. Moreover, the perceived delay value can be obtained by random walk model and sequential particle filtering method. The performance evaluation performed by simulation shows that the algorithm is able to eliminate the uncertainty information in the signal, and the spectrum sensing performance and the sensing delay are significantly improved.
Keywords: spectrum sensing; joint estimation; perceptual delay; channel gain.
Anti-local occlusion intelligent classification method based on Mobilenet for hazardous waste
by Jinxiang Chen, Yiqun Cheng, Jianxin Zhang
Abstract: Anti-local occlusion intelligent classification methods based on Mobilenet and VTM for hazardous waste are investigated in this paper. Three image data sets with ten kinds of hazardous waste and 5000 samples are constructed, which include the image data set with without occlusion, the image data set with 15% occlusion, and the image data set with random occlusion. Based on them, the Mobilenet and VTM intelligent classification model are constructed, trained, and tested. It can be seen from the testing results that the classification accuracies of VTM and Mobilenet are very high for the image data set with and without occlusion. But as occlusion areas on images go up or randomly change, the classification accuracies of VTM and Mobilenet go down for 15% and random occlusion cases. The testing results show that classification accuracy of Mobilenet model is better than that of VTM model for hazardous waste with or without occlusion.
Keywords: hazardous waste classification; occluded target identification; VTM; Mobilenet.
Surface detection method of glass fibre composites based on computer vision
by Yanfang Shi, Jianguo Shi
Abstract: Considering the high cost, low efficiency and poor real-time performance of manual inspection methods in detecting surface defects such as glass fibre imprints, resin build-up and wrinkles. Therefore, in this paper, a machine vision-based method is proposed to detect surface defects of glass fibre composites. The method designs an automatic inspection platform using two high-resolution line scan cameras for image acquisition. The eight directional templates of the kirsch operator are used to convolve the derivatives of the image pixel points respectively, and the largest template is selected to determine its edge direction, and the detection of surface defects is achieved by combining with the canny operator. The experimental results show that the proposed algorithm can well suppress noise interference, improve the accuracy of edge localization and detection, and well retain edge information while avoiding pseudo-edges.
Keywords: surface defect detection; computer vision; glass fibre composites; Canny edge detection; Kirsch operator.
Analysis and control for ultralow frequency oscillation damping caused by asynchronous networking mode
by Renqiu Wang, Shangyu Tian, Hongjian Shi, Feiao Li, Yuanyi Kang
Abstract: The ultralow frequency oscillation damping problem caused by asynchronous networking mode is investigated for Chongqing-Hubei asynchronous grid. The negative damping problems offered by governors are investigated. The electromechanical transient model of the voltage source converter (VSC)-high voltage direct current (HVDC) for the Chongqing-Hubei asynchronous grid is constructed. Fault modes that cause ultralow frequency oscillation after VSC-HVDC are analysed by using the Power System Analysis Software Package (PSASP), because faults in asynchronous grid often result in the ultra-low frequency oscillation damping. An additional damping controller of the VSC-HVDC and its small signal model are proposed, and the mechanism of the additional controller which increases the system damping is also analysed. The simulation results show that low frequency oscillation can be effectively suppressed by using the provided controller in this paper.
Keywords: ultralow frequency oscillation; asynchronous grid; additional damping controller; VSC-HVDC electromechanical model.
Identification of residual disease followed by trade-off analysis between drug optimisation, MRD and sustenance of normal haematopoiesis under maintenance chemotherapy in childhood acute lymphoblastic leukaemia
by Durjoy Majumder
Abstract: Acute Lymphoblastic Leukaemia (ALL) is a commonly occurring cancer in children, and relapses in many cases. Hence to remove (minimal) residual leukaemia or disease (MRD), a maintenance chemotherapy schedule is conducted for two years after intensive chemotherapy. MRD detection occasionally fails owing to the mutability behaviour of leukemic cells or their aberrant marker expression, and the presence of MRD in very minor amount enhances the chances of relapse in the long term. Application of higher drug dose during the maintenance phase may remove MRD, but produces drug-related toxicity. Bone marrow biopsy is required for MRD detection. Here, a peripheral blood based control theoretical model is proposed to detect the presence of MRD. Moreover, the model-based eigen and trade-off analysis could provide a guidance in clinical decision making to optimise the maintenance chemotherapeutic regime (both dose and duration) for individual ALL patients.
Keywords: delay ordinary difference equation; drug application control; drug optimisation; chemotherapy in leukaemia; clinical decision support system.
Adaptive modified super-twisting sliding mode control based on PSO with neural network for lateral dynamics of autonomous vehicle
by Rachid Alika, El Mehdi Mellouli, El Houssaine TISSIR
Abstract: In this article, we have developed a strategy for controlling the lateral dynamics of an autonomous vehicle. The bicycle model of the autonomous vehicle is used. In order to improve the systems performance, we take a new dynamic surface of the sliding mode and a novel expression of the super twisting part of the controller. The parameters of the controller are determined using the particle swarm optimisation (PSO). The objective of this strategy is to follow the reference trajectory of the autonomous vehicle while reducing the lateral displacement error. The steering angle is the control input, the output of this system are the lateral displacement and the yaw angle. The radial basic function neural network (RBFNN) is used to approximate the unknown nonlinear dynamic. Simulation results show some improvements over the literature.
Keywords: autonomous vehicles; STSMC; PSO; RBFNN; nonlinear dynamic; path planning; Lyapunov’s stability theory.
A fuzzy enhanced adaptive PID control algorithm for Quadrotor aircraft
by Wei Li, Kai Zhang, Chunpeng Zhang, Qiang Wang, Yi Zhang
Abstract: A fuzzy enhanced adaptive PID control algorithm is designed for quadrotor Unmanned Aerial Vehicles (UAVs). An ideal quadrotor dynamic model is established through the dynamics analysis of a quadrotor first. To verify the effectiveness of the proposed control strategy, both the hovering and trajectory tracking simulations are carried out with this method. Experiments and simulations show that the designed controller can perform well in various conditions, and the tracking error can be limited to 0.41 meters under a disturbance condition. The comparison results with the traditional PID control algorithm also prove the overall dominance of the proposed controller.
Keywords: fuzzy enhanced adaptive, quadrotor aircraft, under actuated, double closed-loop PID
Chaotic Harris Hawks optimisation based fuzzy lead lag TCSC and PSS with coordinated control design for enhancement of power system transient stability
by Asit Kumar Patra, Sangram Mohapatra
Abstract: This paper investigates the application of the chaotic Harris Hawks optimisation (CHHO) technique for the tuning of a coordinated control of fuzzy lead lag based TCSC controller with fuzzy PSS power system. The eigenvalue and simulation results of the proposed CHHO based optimised TCSC controller are presented and compared with a coordinated control of lead lag TCSC controller with PSO, DE and GSA optimised lead lag controller under various cases of operating conditions and disturbances in SMIB power system. The proposed CHHO-based fuzzy coordinated controller is compared with CHHO-based lead lag TCSC coordinated control of same power system to check the effectiveness and robustness analysis. Finally, the proposed design approach is extended to a multi-machine test model system to demonstrate how the coordinated control of fuzzy lead lag TCSC damping controller with fuzzy power system stabilises and damps out oscillations in power system to improve transient stability performances.
Keywords: TCSC; power system stability; fuzzy lead lag damping controller; multi-machine power system. chaotic Harris Hawks optimisation.
Controlling and Stabilization of Remotely Operated Underwater Vehicle (ROV)
by Fahad Farooq, Noman Ahmed Siddiqui, Amber Israr, Zain Anwar Ali
Abstract: Remotely operated underwater vehicles (ROVs) play a significant role in deep and shallow water missions for exploration, inspection, and extraction. The motions of ROV are guided and controlled by a human pilot present on a surface through a single cord providing power. This study presents the mathematical modelling, kinematic model, and hydrodynamic model of the designed underwater vehicle. It also designs proportional, integral, and derivative (PID) with the non-linear observer model for ROV which helps in controlling and stabilizing its position. The PID controller helps in controlling the altitude of the vehicle while a non-linear observer model with PID controls and stabilizes the attitude. The simulation results show that the designed control scheme is highly accurate and effective. It also shows higher stability and better transient response (TR).
Keywords: ROV controlling, underwater vehicle, PID
Robust adaptive SMC for uncertain singular delayed systems via observer
by Xiaoliang Tang, Zhen Liu
Abstract: This article is focused on state-estimation-based adaptive control design for uncertain singular systems subject to state delay and uncertain nonlinearity by employing sliding mode technique. Firstly, the unmeasured state variables are generated by a particular observer without any inputs, and a new switching surface function of linear type is presented. In view of linear matrix inequality technique, the motion of the closed-loop system on the sliding surface is analysed, and new admissibility criteria are deduced. Then a switching controller with adaptive rules is synthesised to ensure the established sliding surface can be attained in finite moment. Finally, an illustrative example is proposed to demonstrate the feasibility of the theoretical method.
Keywords: singular delayed systems; state-estimation; adaptive sliding mode control; admissibility.
Optimisation of group batch scheduling in flexible flow shop based on multi-player cooperative game
by Zhonghua Han, Xusheng Bian, Ziyao Ding, Dechang Sun
Abstract: In a flexible flow shop, if there are group batch processes, the processing task needs to form multiple batches for production in the batch operation. In order to make up batch processing, there will be job-waiting time thereby extending the total working hours of the production process, and the conflict between various production indexes is further intensified. Therefore, this paper first establishes the group batch scheduling of flexible flow shop (GBSFFS) mathematical model, and proposes a local scheduling method based on multi-player cooperation and complete information static game, establishes a method of game elements such as game information, game player, game strategy, game payoff for group batch processing, and uses the method for behaviour prediction of various game players, that is, the valuation indexes, which reduce the evaluation indexes conflict in the GBSFFS problem. On the premise of reducing the waiting time, other evaluation indexes have been increased.
Keywords: flexible flow shop; production jam; multi-player static cooperative game; batch scheduling.
Distributed control of linear partial integro-differential equations based on the input-output linearisation approach
by Ahmed Maidi, Jean-Pierre Corriou
Abstract: In this paper, the input-output linearisation control approach is extended to distributed parameter systems whose dynamical behaviour is described by a partial integro-differential equation. The design of the infinite dimensional state feedback controller is achieved using the late lumping approach, i.e., using the partial integro-differential equation model without any prior reduction or approximation. Thus, based on the notion of the characteristic index as a generalisation of the relative degree, a distributed state feedback controller is designed by evaluating the successive time derivatives of the controlled output. The designed controller yields in closed loop a first order lumped parameter system where the time constant is a design parameter that fixes the desired dynamical behaviour. The stability of the closed loop system is investigated, based on semi-group theory, by employing the perturbation theorem of the bounded linear operators, and the sufficient condition for exponential stability in L2-norm is derived. This condition yields the upper bound for the design parameter, i.e., the time constant. Both output tracking and stabilisation capabilities of the developed state feedback are demonstrated through numerical simulation by considering three application examples: Volterra, Fredholm and Fredholm-Volterra PIDEs. The effectiveness of the developed controller is shown by simulation.
Keywords: distributed parameter system; partial integro-differential equation; input-output linearisation; semi-group theory; perturbation theorem; exponential stability.
Hover autopilot design for an un-crewed helicopter using static output feedback controller
by Femi Thomas, S.J. Mija
Abstract: Design of linear matrix inequality-based static output feedback controller for an un-crewed helicopter in hover flying mode is presented. The six degrees of freedom linear time-invariant state-space model of the vehicle is developed analytically from first principles considering the force and moments acting on it without the usual simplifying approximations. Since access to full state information is not a situation in practice, output feedback-based controller is an appealing solution for the autopilot design. Here, two separate static output feedback controllers are developed for the fast inner-loop and the slow outer-loop dynamics of the vehicle. As the number of variables to be fedback is reduced, the proposed scheme is simpler compared to conventional state feedback controller. Numerical simulation studies validate that the proposed controller exhibits fast transient performance and robustness when subjected to wind disturbances acting in the three fuselage axes during the hover flight.
Keywords: un-crewed helicopter; UCH; hover flight; static output feedback controller; SOFB; static state feedback controller; SSFB; linear matrix inequality; LMI; Lyapunov equation.
The statistical analysis in the era of big data
by Zelin Wang, Xinke Liu, Weiye Zhang, Yingying Zhi, Shi Cheng
Abstract: In the big data environment, the traditional machine learning algorithm for data processing is somewhat inadequate. Therefore, machine learning algorithms adapted to big data environment have become a research hotspot. At the time of the marriage of big data and machine learning, it is necessary to predict the related challenges and opportunities. This paper mainly analyses and summarises the current research status of machine learning algorithms for processing big data, and discusses the new opportunities and challenges that machine learning paradigm will face in the era of big data. It also explores the new technology breakthrough that machine learning will produce in the era of big data.
Keywords: big data; machine learning; deep learning; integrated learning; transfer learning.
Hyperparameters optimisation of ensemble classifiers and its application for landslide hazards classification
by Jiuyuan Huo, Hamzah Murad Mohammed Al-Neshmi
Abstract: Along with assessing the hazards of landslides taking into consideration the faced difficulties and the consumed time when determining the algorithms configurations and parameters manually, the primary aspiration of this study is to optimise the parameters of two ensemble-based machine learning algorithms using particle swarm optimisation, genetic algorithm, and Bayesian optimisation so that the optimised algorithms can identify and classify landslides more efficiently and accurately. Random forest classifier and XGBoost models were used and the ADASYN was implemented to overcome the shortage of imbalanced data. In the experiments, it was clearly shown that the hypered ensemble-based models along with the PSO and GA successfully surpassed the single models on classifying the landslides' triggers, sizes, and types. The experimental results demonstrated that the hyperparameters optimisation can greatly improve the accuracy of the ensemble classifiers, thus it can provide accurate classification results and decision support for the disaster prevention and mitigation management departments.
Keywords: landslide; optimisation; random forest; extreme gradient boosting; XGBoost; adaptive synthetic; ADASYN.
Maximum power harvesting from a PV system using an improved two-stage MPPT scheme based on incremental conductance algorithm and integral controller
by Mostufa Atia, Noureddine Bouarroudj, Aimad Ahriche, Abdelhamid Djari, Yehya Houam
Abstract: This article deals with a novel strategy called two-stage incremental conductance (INC) algorithm for maximum power point tracking (MPPT) for photovoltaic (PV) systems. This method is proposed to tackle the shortcomings of classical incremental conductance (INC) method such as oscillations, imperfect MPPT, slow tracking of MPP, etc. These demerits stem from the use of a one-stage INC algorithm with a fixed duty cycle step size, which cannot provide the optimal duty cycle especially under changing weather conditions (temperature and irradiance) and resistive load. The first stage of the proposed approach is used for providing the reference voltage using the INC algorithm; and the second one is an integrator controller tuned by Routh's criterion used to ensure the stability of the voltage closed loop control. Simulation and numerical results in different cases confirm the superiority of the proposed two-stage INC algorithm over the classical one based on one stage, with an efficiency of more than 93.75%.
Keywords: PV-module; boost converter; INC algorithm; one stage; two stages; maximum power point tracking; MPPT; Routh's criterion.
Special Issue on: Machine Learning in Bio-Signal/Image Analysis
Classification of magnetic resonance images of brain using concatenated deep neural network
by Abhishek Das, Mihir Narayan Mohanty
Abstract: Medical image classification is an ongoing research topic in the field of medical science. Deep learning is playing a vital role in image analysis owing to its ability of auto feature extraction. Various works have been developed for brain image classification with models with high complexity but less performance. In our work, we have explored the deep learning techniques in the ensemble and stacking approach with less complexity and improved performance. Convolutional Neural Network, Recurrent Neural Network, and Long Short Term Memory are used as base classifiers for feature extraction and first stage classification. The predictions of the base classifiers are fed to the Multilayer Perceptron model for second stage training and classification. The performance of the proposed model is verified with a brain magnetic resonance image dataset online available at Kaggle. F1-score, recall, precision, sensitivity, specificity, and accuracy are calculated for evaluation of the proposed method and the effectiveness of choosing this method is discussed in the result section. Classification accuracy of 97% is achieved in the proposed method on the brain MRI dataset, which is representing a competitive result concerning the state-of-the-art methods to the best of our knowledge.
Keywords: brain image classification; convolutional neural network; recurrent neural network; long short term memory; multilayer Perceptron; ensemble learning.
Relevant gene selection using ANOVA-ant colony optimisation approach for malaria vector data classification
by Micheal Olaolu Arowolo, Joseph Bamidele Awotunde, Peace Ayegba, Shakirat Oluwatosin Haroon-Sulyman
Abstract: Recent progress in gene expression data research makes it possible to quantify and identify several thousand gene expressions simultaneously. For malaria infection and transmission, gene expression data classification using dimensionality reduction is a standard approach in gene expression data analysis and proposed for this study. A major problem occurs in the reduction of high dimensional data, it plays a significant role in improving the precision of classification, allowing biologists and clinicians to correctly predict infections in humans by choosing a limited subclass of appropriate genes and deleting redundant, and noisy genes. The combination of a novel Analysis of Variance (ANOVA) with Ant Colony Optimisation (ACO) approach as a hybrid feature selection to select relevant genes is suggested in this study to minimize the redundancy between genes, and SVM is used for classification. The proposed method's efficacy was shown by the experimental outcomes based on the high-dimensional of gene expression data.
Keywords: malaria vector; gene expression; analysis of variance; ant colony optimisation; support vector machine; machine learning.
A novel framework for segmentation of uterus fibroids in ultrasound images using machine learning models
by Dilna K T, Anitha J, Jude Hemanth
Abstract: A tumour of non-cancerous structure that appears in the uterus during child-bearing years is called uterine fibroids. Thus, it is necessary to design a fibroid detection system for the fibroid ablation. Various methods developed for the detection of fibroids are easily affected by the image artefacts as they do not take into consideration the spatial information and have lower efficiency problems for fibroid segmentation. This paper puts forward a method for segmentation for fibroid detection. The proposed segmentation model overcomes the drawbacks of existing methodologies of fibroid detection in all stages. Here, the speckle noise existing in the noisy input image can be removed by using IGDT-DWT method and EMD-GCLAHE method. After contrast enhancement, the segmentation of the contrast-enhanced image is done using a novel clustering algorithm, namely PC-KMA. The proposed segmentation algorithm effectively detects the fibroids, which is experimentally proved by comparing it with existing classifiers.
Keywords: uterus fibroid; ultrasound scanned images; DWT; K-means algorithm.
A hybrid model for the identification and classification of thyroid nodules in medical ultrasound images
by Rajshree Srivastava, Pardeep Kumar
Abstract: Ultrasonography (USG) is one of the leading diagnostic methods for accurately distinguishing the early stage of thyroid nodules. ANN-SVM hybrid model is proposed for the identification and classification of thyroid nodules in medical ultrasound images. After feature extraction using grey level co-occurrence matrix method, two experiments are performed. In experiment 1, five different machine learning (ML) classifiers Random forest (RF), Support vector machine (SVM), Decision tree (DT), Artificial neural network(ANN) and K-nearest neighbour (KNN) are used for classification. In experiment 2, the two best classifiers based on the performance are hybridised together. The proposed hybrid model has achieved 84.12% accuracy, 85.14% sensitivity and 82.95% specificity on a public dataset with 295 USG images, and 90% accuracy, 91.66% sensitivity and 87.5% specificity on the local dataset having 654 thyroid USG images. It has shown an improvement of 2% to 5% in the performance evaluation in comparison with the other state-of-the-art methods.
Keywords: ultrasonography; artificial neural network; ANN-SVM hybrid; machine learning; thyroid nodule.
Mathematical modelling for prediction of spread of COVID-19 and AI/ML based technique to detect SARS-CoV-2 via smartphone sensors
by Digvijay Pandey, Sumeet Goyal, Harjinder Singh, Joginder Singh, Rahul Kakkar, P. Naga Srinivasu
Abstract: An infectious and communicable disease named COVID-19 is a novel coronavirus which originated from China. The virus is known as severe acute respiratory syndrome coronavirus 2 abbreviated as SARS-CoV-2 and generally known as COVID-19. Its epicentre is in the Wuhan city of China. In this paper, a mathematical model (SEIR) for the prediction of infectious diseases is described and modelled, which can be used to predict the cases, and a framework to detect SARS-CoV-2 from home is also proposed based on artificial intelligence, machine learning and smartphone embedded sensors. There are various sensors embedded in the smartphones, such as proximity sensor, light sensor, accelerometer, gyroscope and fingerprint sensors, which have very fast processors and memory space thus making it easy to read and compare the symptoms or activity and scan the CT images, and can be used to detect COVID-19. The proposed framework is based on AI, cloud and ML.
Keywords: SARS-CoV-2; modelling; artificial intelligence; machine learning; sensor.
Quantum grey wolf optimisation and evolutionary algorithms for diagnosis of Alzheimer's disease
by Moolchand Sharma, Shubbham Gupta, Himanshu Aggarwal, Tarun Aggarwal, DEEPAK GUPTA, Ashish Khanna
Abstract: Alzheimer's disease (AD) is a type of brain cancer, similar to coronary artery disease. AD is a progressive neurological disorder that impairs memory, thinking abilities, and behaviour. Thus, early detection of the condition is critical, as there is no cure. We conducted a comparative analysis of various evolutionary algorithms for extracting meaningful information from the Alzheimer's dataset, which is then used to predict whether or not a patient has the illness. We attained an accuracy of 78-85% using machine learning methods. When we used various evolutionary algorithms to perform feature selection, we observed an increase in accuracy of 5-10%, with Grey Wolf and Quantum Grey Wolf optimisation (qGWO) achieving the highest accuracy of 92.8% and 94.5%, respectively, using a random forest classifier. The model was evaluated using three metrics: the increase in accuracy, the time required to execute, and the number of features eliminated. Additionally, the testing revealed that certain characteristics are replicated across multiple models and might be regarded as critical in the process of identifying AD.
Keywords: Alzheimer's disease; machine learning; neurodegenerative disease; bio-inspired algorithms; qGWO.
An improved ensemble learning approach for the prediction of cardiovascular disease using majority voting prediction
by Sibo Prasad Patro, Neelamadhab Padhy, Rahul Deo Sah
Abstract: Coronary heart disease (CHD) is one of the most common heart disease types in the world. Heart disease is a critical health issue today. It is one of the most frequent causes of mortality owing to a lack of proper medical diagnosis, technology, and a healthy lifestyle. Machine learning is a typical application to predict an outcome from existing data. The machine learns the patterns from an existing dataset, and it applies different rules to the dataset to predict the outcome. Classification becomes a powerful machine learning technique for prediction. Few classification algorithms give satisfactory results, and some others produce limited accuracy. In this work, we propose a new ensemble classification model by combining multiple classifiers for improving the accuracy of weak algorithms. An ensemble classifier was applied by using a majority vote-based technique for cardiovascular disease prediction and classification. The performance of this model is implemented on the Cleveland dataset from the UCI repository has applied a three-dimensionality approach on the dataset, and the average accuracy of each method is calculated as PCA(0.8636), K-PCA(0.8630), and LDA(0.90). As the PCA and K-PCA provide the same accuracy, whereas LDA gives higher average accuracy. So LDA is used as the best dimensionality reduction technique. The results show that Model-6 based on AdaBoost with base estimator SVC gives an accuracy of 0.9230 and Model-7 based on AdaBoost with base estimator RFC gives testing accuracy 1.0 respectively.
Keywords: ensemble methods; voting classifier; coronary heart disease; bagging classifier; stacking classifier; Adaboost classifier.
An ensemble-based approach for image classification using voting classifier
by Bhoopesh Singh Bhati, Achyut Shankar, Srishti Saxena, Tripti Saxena, Anbarasi M, Manoj Kumar
Abstract: This paper presents a proposed scheme for image classification. Image classification is used in different areas to identify images of people, places, and objects accurately. There are many different image classification models, but the proposed scheme of using the voting classifier provides better results with high performance. The paper describes how the drawback of machine learning algorithms is overcome by using voting classifier, as it helps to improve the results by combining multiple machine learning algorithms. The experimentation is carried out on the Fashion MNIST dataset, which is a benchmark for image classification. Three machine learning algorithms: K-Nearest Neighbour, Random Forest, and Decision Tree, are used to evaluate the proposed scheme. The proposed scheme gives an accuracy of 87.39%. The ensemble method is better than the individual accuracies of the three algorithms.
Keywords: machine learning; image classification; voting classifier; ensemble learning.
Special Issue on: Soft Computing for Data Analytics, Image Classification and Control
DETECTION OF CORONARY ARTERY DISEASE USING MACHINE LEARNING ALGORITHMS
by Kriti Vashistha, Anuja Bokhare
Abstract: One of the most difficult tasks in medicine is predicting heart disease. Every minute, roughly one person dies from heart disease in the modern era. The heart is the second most important organ in the human body after the brain. Predicting the occurrence of heart diseases is the most important work in the medical industry. This is where machine learning and data analytics comes into play. Moreover, the medical industry is able to collect huge amount of data on a monthly basis. This information can be used to forecast the occurrence of future diseases. According to the previous work for this research authors have mostly worked on algorithms like KNN, SVM and Nave Bayes. In this study, the proposed technique analyses three different algorithms: decision trees, random forests, and logistic regression. After correctly training and evaluating the models, we noticed that random forest had the highest accuracy of 83 percent, followed by logistic regression with 81 percent, and decision tree with 77 percent. The most important factors in prediction were found to be age, Trestbps, Cholesterol, and Oldpeak. For future work we would enhance the accuracy of our model which will hopefully one day be able to help battle the ever-growing problem of coronary artery disease.
Keywords: medical industry; heart disease; random forest; decision tree; logistic regression; machine learning.
Optimization of Target Coverage in Wireless Sensor Network Using Learning Automata Approach
by Haribansh Mishra, Anil Kumar Pandey, Bankteshwar Tiwari
Abstract: Wireless Sensor Networks (WSNs) technology is employed in multiple areas like battleground surveillance, home security etc. In WSN, most algorithms are based on the Maximum Cover Set (MCS) for energy-efficient target coverage (TC). But it generates the NP-complete problem of constructing maximal CS. These Covers formations consume more energy because each node participates in the building of Sets. To reduce the average energy consumption of networks, we propose Learning Automata based on a scheduling algorithm called Self-Adaptive Minimum Energy Consumption algorithm (SAMECA). The SAMECA assists each sensor to choose the proper state (active or sleep) at any given time. The purpose of SAMECA is to increase the network lifetime by maximizing the sleep state presence of nodes. Besides, it ensures that fewer sensors are required to cover all the targets. The results indicate that the SAMECA is a decent option to analyze all the targets by consuming less energy power.
Keywords: Learning automata; Lifetime; Sensor; Wireless sensor network.
Classification of Imbalanced Hyperspectral Images using Ensembled Kernel Rotational Forest
by Debaleena Datta, Pradeep Kumar Mallick, Mihir Narayan Mohanty
Abstract: Hyperspectral image classification suffers from an imbalance in the samples belonging to its different classes. In this paper, we propose a two-fold novel approach named Oversampler+Kernel Rotation Forest (O+KRoF). First, Synthetic minority oversampling (SMOTE) and Adaptive synthetic oversampling (ADASYN) techniques are employed on original data to balance it due to their adaptive nature in the majority and minority samples. Finally, the ensembled KRoF classifier is applied, a combination of unpruned Classification and Regression Trees (CART) as its base algorithm and kernel PCA for feature reduction and most significant nonlinear spatial-spectral feature selection. Furthermore, we designed a comparison study with frequently used oversamplers and related state-of-art tree-based classifiers. However, it is found that our ensemble model is suitable and performs better as compared to earlier works as it attains 90.92%, 97.1%, and 93.39% overall accuracies when experimented on the benchmark dataset, Indian Pines, Salinas Valley, and Pavia University, respectively.
Keywords: Hyperspectral Images; Resampling; Synthetic Oversampling; Tree-based classifiers; Modified Rotation Forest.
An Efficient Data Retrieval Method for Grid Blockchain
by Caijun Zhang, Qianjun Wu, Jiayi Lang, Huafei Yang, Xiaolong Wang, Kaiqiang Xian, Jingqiu Zhang
Abstract: The blockchain-based power grid integration business system (PG-IBS) are increasing rapidly. However, due to the limitation of blockchain, these systems have problem of low data retrieval efficiency. To solve this problem, through careful investigation and analysis, an efficient data retrieval method for power grid blockchain (EDRM-PGB) is proposed in the paper. EDRMPGB rebuilds an efficient retrieval index structure TIS (Transaction Index Structure) for a PG-IBS, while maintaining compatibility with the original system. TIS index structure is built on two data structures BABF (Blockchain Account Bloom Filter) and BTTI (Binary Tree with Transaction Information). Based on the structure, EDRM-PGB efficient retrieval algorithm is designed. EDRMPGB's feasibility is verified by the prototype system implementation and performance simulation. Simulation result shows that compared with the traditional retrieval method, EDRM-PGB can greatly improve the data retrieval performance of PG-IBS. Meanwhile, it also has advantage of sharing of index files easily.
Keywords: data retrieval; blockchain; power grid; retrieval algorithm.
Special Issue on: Recent Advances on Learning-Based Control Theory and Application
A Survey on Modern Trends of Low Power Long Range Network applied on IoT applications
by Muhammad Aamir Khan, Zain Anwar Ali, MUHAMMAD SHAFIQ
Abstract: In recent years, Long Range (LoRa) network are gaining popularity in all areas of Engineering fields and also demands for minimized the structure of the network to cover a wide geographical area with extremely low power consumption. LoRa network is designed for the broad range communications capacity especially suitable for Internet of Things (IoT) applications. In the wide context of communication channels, LoRa has the significant support applications for long distance multi-hop network with the minimize packet size and low latency. This paper presents the recent advancements and technical analysis of LoRa network in different IoT applications. The paper also reviews performance and challenges faced by LoRa networks under different scenarios. The paper also involves the findings and restrictions of the proposed work to help research scholars for the network optimization in order to improve the performance parameters for any environment.
Keywords: Long Range Network; Low Power Consumption; Internet of things and communication channels.
Experimental validation of an output feedback controller based on an integral and adaptive backstepping technique for a fuel-cell power system.
by Soukaina Nady, Hassan EL Fadil, Fatima Zahra Belhaj, Abdessamad Intidam, Mohamed Koundi, Zakariae El Idrissi
Abstract: The present work establishes a comparison between two controllers based on a backstepping approach for a fuel-cell power system. The load resistance representing the impedance of the DC bus is assumed to be unknown and can change. Besides, the internal fuel-cell voltage is not accessible for measurement. Therefore, to cope with these two issues, two output feedback controllers are designed using a backstepping technique. The first controller uses an integral action while the second one is an adaptive version of the former. It is formally shown using theoretical analysis and simulation that the obtained controllers achieve all control objectives. A comparison between the two controllers shows that, when they are correctly tuned, both behave almost similarly. Nevertheless, we noted the weak supremacy of the adaptive version over the integral version in terms of rapidity. A laboratory prototype is built to show the effectiveness of the proposed control approaches.
Keywords: Fuel cell; dc-dc buck power converters; nonlinear control; adaptive control; Lyapunov stability; output feedback; backstepping technique.
Special Issue on: ICECOCS'20 Intelligent Control for Future and Complex Systems
Digital implementation of model predictive control of an inverter for electric vehicle Applications
by Khawla Gaouzi, Hassan El Fadil, Zakariae El Idrissi, Abdellah Lassioui
Abstract: The control of DC-AC converters with output LC filter has a special importance in applications intended for electric vehicles. However, the controller design becomes more complicated. This paper proposes a Model Predictive Control (MPC) of DC-AC power converters for electric vehicle applications. The control objective is to regulate the output voltage of a three-phase inverter with output LC filter to its desired constant values. Once the control law is elaborated, several simulations are performed using Matlab/Simulink tools, and the results show that the output voltages perfectly track its references. Experimental results are also given, which show the effectiveness of the model predictive controller.
Keywords: model predictive control; three-phase voltage source inverter; electric vehicle; Matlab/Simulink.
Experimental and numerical study of the influence of FFF process parameters on the flexural properties of 3D printed medical devices and personal protective equipment
by Mohamed Abouelmajd, Ahmed Bahlaoui, Ismail Arroub, Manuel Lagache, Soufiane Belhouideg
Abstract: 3D printers are increasingly used after the spread of the Covid-19 pandemic. This machine was used to overcome the lack of medical devices and personal protective equipment. In order for printed objects to be useful in the medical field, the mechanical properties of these objects must be known. The main objective of this study is to assess the mechanical properties of parts manufactured in polylactic acid by a 3D printer machine. The mechanical properties are determined from the experimental results of the three-point bending test. The results obtained show that the mechanical properties depend on the process parameters. The analysis of variance was used to determine the printing parameters that have a statistically significant effect on the mechanical properties. The optimal printing parameters are determined to manufacture parts with high mechanical performance so they can be used in complete safety. The finite element analysis was performed using ANSYS Mechanical APDL software.
Keywords: fused filament fabrication; mechanical properties; polylactic acid; design of experiments; analysis of variance; response surface methodology.
Black widow optimisation based controller design for Riverol-Pilipovik water treatment system
by Nitin Mathur, Veerpratap Meena, Vinay Pratap Singh
Abstract: In this paper, a black widow optimisation (BWO) algorithm based proportional-integrative-derivative (PID) controller is proposed for the Riverol-Pilipovik (RP) water treatment plant. The plant is a two-input-two-output (TITO) system with two interacting loops, hence a decoupler is deployed to convert these interacting loops into non-interacting loops. The PID controller is designed by formulating an integral-square-error (ISE) and minimising it. In this work, the BWO algorithm is used for minimising the ISE. The interval model of the RP water treatment plant is considered and the controller is designed for both the lower interval as well as the upper interval model by using BWO and Jaya algorithm. The results obtained affirm the fidelity of the proposed BWO-based PID controllers. Further, the controllers designed are efficient and stable.
Keywords: black widow optimisation; controller design; reverse osmosis; Riverol-Pilipovik water treatment system; system modelling.
Robust hybrid controller for quadrotor UAV under disturbances
by Hamid Hassani, Anass Mansouri, Ali Ahaitouf
Abstract: With the increasing use of autonomous quadrotors in daily activities, the development of robust tracking controllers which allows these vehicles the achievement of the planned location has become extremely important. In this paper, a novel hybrid control strategy is designed for an uncertain quadrotor affected by external disturbances. A robust sliding mode controller is used to stabilize the vehicle orientation. In addition, an adaptive rule is developed for the online tuning of the switching gains. Whereas, a finite-time control law is designed for the quadrotor position, in which a nonsingular terminal sliding mode control (NTSMC) is combined with the supertwisting algorithm (ST), to obtain fast convergence with reduced chattering influence. To prove the stability of the entire system and calculate the adaptive laws, Lyapunov stability concept is used. Using the proposed hybrid controller, fast convergence, reduced tracking errors and strong robustness are all ensured. Finally, the performance of the proposed control scheme is tested under the influence of complex disturbances. Simulation results show the efficacy of the suggested method in term of path tracking and robustness.
Keywords: hybrid controller; ASMC; finite-time control; NTSMC; supertwisting.
A study on photovoltaic charging strategy of electric vehicles with multi-objective constraints.
by Wu Haowen, Wang Chong, Zhou Yudi, Xie Wenwang, Chen Chen
Abstract: Installing photovoltaic panels (PV) in the workplace to charge electric vehicles (EV) can realise the effective coupling of clean energy and EV. The disorderly charging behaviour of EV photovoltaic charging stations will aggravate the fluctuation of the power grid distribution side, and make the energy dispatching among electric vehicles, photovoltaic power generation and power grid unreasonable, which will lead to the increase of charging cost. In order to solve the above problems, a photovoltaic charging strategy considering multi-objective constraints is proposed. By modelling the photovoltaic charging scene of EVs, extracting the constraints, such as charge power and charging time of EVs, a mixed integer linear programming formula considering multi-objective constraints is constructed. The result of solving the formula is the charging strategy of EVs. The simulation results verify the effectiveness of the method in reducing the charging cost.
Keywords: electric vehicle charging; photovoltaic power generation; multi-objective constraints; hardware-in-the-loop; mixed integer linear programming.