International Journal of Wireless and Mobile Computing (32 papers in press)
Expected-mode augmentation method for group targets tracking using the random matrices
by Yun Wang, Guoping Hu, Hao Zhou
Abstract: In order to improve the estimation performance of interactive multiple models (IMM) tracking algorithm for group targets, a new EMA-VSIMM tracking algorithm is proposed in this paper. Firstly, by using the expected-mode augmentation (EMA) method, a more proper expected mode set has been chosen from the basic model set of group targets, which can make the selected tracking models match up to the unknown true mode availably. Secondly, in the filtering process of variable-structure interactive multiple model (VSIMM) approach, the fusion estimation of kinematic state and extension state have been implemented by using classical weighting method and scalar coefficients weighting method, respectively. We use the trace of the corresponding covariance matrix of extension state to calculate the weight coefficient. We calculate the prediction value of the extension state parameter by using a fuzzy reasoning approach to improve the estimation accuracy of the covariance matrix, which takes the elliptical area of extension and its change ratio as the input of the fuzzy controller. The performance of the proposed EMA-VSIMM algorithms is evaluated via simulation of a generic group targets manoeuvring tracking problem.
Keywords: interactive multiple models; expected-mode augmentation; group targets; maneuvering tracking.
Low financial cost with ant colony optimisation in intelligent agriculture
by Xu Gaofeng
Abstract: With the development of wireless sensor networks, industrial automation and other computer and information related high technologies, a lot of practical IoT applications have greatly increased the productivity. Currently, more and more capital is being invested in IoT, especially intelligent agriculture as many countries begin to pay more attention to basic and intelligent agriculture. For large intelligent agriculture systems, it will cost a lot of time and energy (which further will cost investors' money) for the mobile sink to collect all the data of the sensing system with the help of cluster head node. In this paper, we try to solve this issue that minimizes the data collection path of the mobile sink, with the help of the ant colony optimisation algorithm. We implement the algorithm in Python and conduct two experiments that show that we can get the best path of the given example and show how the efficiency changes when the numbers of ants and loops increase. The better the optimal path becomes, the less financial cost we can achieve.
Keywords: financial cost; wireless sensor network; ant colony optimisation; intelligent agriculture.
Demand estimation of water resources via bat algorithm
by Xiangdong Pei, Youqiang Sun, Yeqing Ren
Abstract: In the process of urban water resources planning, the demand estimation of urban water consumption is one of the important basic contents. In this paper, a hybrid of a linear estimation model and an exponential estimation model is proposed to forecast the water consumption. The bionic intelligent algorithms are widely used in industrial engineering, so we use intelligent algorithms to solve the proposed model including Bat Algorithm (BA) and modified Bat Algorithm (FTBA). FTBA improves the global search capability, and the improvements increase the probability of solving the optimal value. In the simulation experiments, we use the data from Nanchang city during 2003 to 2015. The data from 2003 to 2012 are used to find the optimal weights, and the remaining data (2013-2015) are used to test the model. Simulation results show that the modified BA (FTBA) is superior to the standard algorithm and achieves higher accuracy in prediction.
Keywords: demand estimation; water resource; hybrid model; bat algorithm.
A hybrid swarm algorithm and its application in vehicle route guidance system on wireless networks
by Qi Xin, Han Chenghao
Abstract: In this paper, a hybrid optimisation algorithm for the vehicle route guidance system on wireless network is presented. Applying the wireless sensor network technology to intelligent transportation systems has become a research hotspot in recent years. Research on the shortest path algorithms and improving vehicle route guidance system has been the core of the study on the intelligent transport systems. The goal of our approach is to reduce the computing time and get better solutions. By a thorough analysis of several existing optimal routing choice algorithms, we propose an optimal routing choice algorithm based on artificial fish swarm algorithm. Our simulations show that the speed of convergence of the improved algorithm is enhanced greatly compared with other optimal choice algorithms, and the new algorithm obtained a better result. The results show that the improved algorithm can be applied to intelligent transport systems.
Keywords: vehicle route guidance system; swarm intelligence; artificial fish swarm algorithm; transportation.
Lean evaluation system of manufacturing enterprises based on full lifecycle
by Le Yang, Guozhang Jiang, Zhongyuan Li, Gongfa Li, Chao Xu
Abstract: The production of manufacturing enterprises is a complex process. In lean production it is hard to achieve lean weight reduction. It is very important to propose a lean life assessment framework. This paper evaluates the lean production level of the whole lifecycle of the company by using the fuzzy theory and the neural network technology, establishes a scientific and systematic framework of the lean evaluation index of the whole lifecycle, and establishes the evaluation model of the fuzzy neural network. The example analysis shows that the actual output value of the fuzzy neural network model is not much different from the predicted output value, which indicates that the model has high prediction accuracy. The experimental results further verify the reliability and validity of the model. The system framework can well plan and evaluate the whole lean production level. Accurate assessment of lean production level is an effective way to improve lean levels.
Keywords: whole lifecycle; system architecture; fuzzy theory; neural network; evaluation system.
An actual traffic prediction method based on particle swarm optimisation and wavelet neural network
by Ke Chen, Zhiping Peng, Wende Ke
Abstract: For the congestion phenomena of network, this paper provides a new prediction method for service flow, based on particle swarm optimisation and wavelet neural network prediction (PSOWNNP). Firstly, this method uses wavelet exchange to resolve the service flow, and uses its wavelet coefficient and metric coefficient as the sample data. Secondly, the sample data is trained using the neural network method of the particle swarm optimisation, in which the wavelet model is applied for construction, and the prediction data for service flow are obtained from this. At the same time, the prediction methods of wavelet neural network and BP neural network for particle swarm optimisation are analysed and compared through simulation experiments, and the result for indicating the performance of AWNNP method is relatively good, with a tolerance of 17.21%.
Keywords: congestion; prediction; particle swarm; neural network.
An anomaly detection technique based intrusion detection system for wireless sensor networks
by Sushant Kumar Pandey
Abstract: In wireless sensor networks (WSN), from research it has been found that to figure out a compromised node is a very difficult task. A compromised node is one that doesn't have features like other normal nodes. A compromised node can be an intruder for the network and can be harmful for the network. It can steal information or can interrupt data communication. Sometimes it can corrupt other nodes in a network. In WSN, there is limited power and very limited processing capability. Moreover, in the WSN homogeneous network nodes have similar behaviour and this is the main reason why to figure out a compromised node is a difficult task. We have used an anomaly technique for detection of compromised nodes, in which we have compared the behaviour of anonymous node with normal node. There are some certain parameters on the basis of which it is decided that behaviour will get abnormal or not. We have chosen some operators called IDS operators from all the set of normal nodes. This selection is done by internal congestion within a node. The IDS operator will be selected in such a way so that it will cover the full network architecture. Every node should be in the radio range of IDS operator. Learning will be provided to the IDS operator, so that it can figure out the compromised node from the network. This learning is based on parameters, which are related with behaviour of the normal sensor nodes. After detection of a compromised node, the IDS operator will notify the sink node about it. The IDS operator is behaving as normal nodes. If any suspicious node is in the radio range of it, then it will be tested by the IDS operator.
Keywords: anomaly technique; intrusion detection system; IDS controller node; WSN; behaviour parameters.
Evaluating the performance of asynchronous MAC protocols for wireless sensor networks
by Abdelmalek Bengheni, Fedoua Didi
Abstract: Medium Access Control (MAC) protocols for duty-cycling, in Wireless Sensor Networks (WSN), are responsible for arbitrating access to a shared medium, for coordinating the access from active nodes and conserving power energy. These types of protocol can be classified into two categories: synchronous and asynchronous MAC protocols. In synchronous MAC protocols, each node is synchronised with the neighbourhood where each transmitter node can transmit a packet to the corresponding receiver during their period of listening, through the use of the control messages. The main goal of these protocols is to reduce energy loss caused by idle-listening, collision, overhearing and control overhead. On the other hand, asynchronous MAC protocols reduce energy consumption without the use of control messages. The main objective of this paper is to evaluate the performance of an experimental WSN using two different types of asynchronous MAC protocol; the first uses the transmission of the preamble frame such as B-MAC (Berkeley MAC) and X-MAC (A Short Preamble MAC) and the subsequent based on beaconing paradigm that uses the transmission of the beacon frame as RI-MAC (Receiver Initiated MAC). In this experiment, we used OMNeT++/MiXiM network simulator to create WSN and then evaluate its performance in terms of the packet delivery ratio, mean latency and energy consumption. The simulation results show that RI-MAC performs better than X-MAC and B-MAC and that X-MAC is ranked second best before B-MAC.
Keywords: performance; wireless sensor network; asynchronous MAC protocols; OMNeT++; MiXiM.
Enhancement of the dynamic bandwidth channel access for IEEE 802.11ac WLANs
by Fadhila Halfaoui, Mohand Yazid, Louiza Bouallouche-Medjkoune
Abstract: The aim of the new IEEE 802.11ac amendment is to significantly increase the throughput within the Basic Service Set (BSS) by improving the physical data rate. IEEE 802.11ac provides a maximum Multi-Station (Multi-STA) throughput of at least 1 Gbps and a maximum single link of at least 500 Mbps. This achieved throughput is obtained by adding many enhancements for the PHY and MAC layers. One of the main features that allows 802.11ac to reach gigabit transmission rates is a static and dynamic channel bonding technique. In this paper, we first overview the static 80 MHz and the dynamic 20/40/80 MHz bandwidth channel access schemes defined in 802.11ac. We then present the proposed enhancement of the dynamic access for an 80 MHz wide channel. The simulation results show that the enhanced dynamic access offers better throughput compared with the two methods used by 802.11ac stations.
Keywords: IEEE 802.11ac; multi-channel access; simulation and performance analysis.
Transport layer dependability benchmarking: TCP congestion control with fault injection
by Maroua Belkneni, M.Taha Bennani, Samir Ben Ahmed
Abstract: Previous approaches have used fault injection to evaluate the distributed system's dependability by targeting the application and hardware layers. In this paper, we delve on the communication layer services which are located between the two previous layers. We provide the specification of the assessment components: fault injection, architectural structure of the workload, and dependability measures. We have examined in this paper, the use of fault injections to evaluate the dependability of a transport layer protocol (i.e., TCP) with a random study. We have focused on the congestion control service, then, we have defined workload, faultload and dependability measures. We have shown that the communication is interrupted after exceeding a given tolerance threshold (below this limit, the service employs a recovery time to resume execution). We conclude that congestion control is robust to port fields; however, fault injections to the window size and flag fields hang the service. The test bed is carried out through the simulator NS-3.
Keywords: wireless sensor networks; congestion control; transmission control protocol; NS-3; dependability; assessment; fault injection.
Anti-conspiracy attack threshold signature model and protocol
by Xiaoping Wang, Zhenhu Ning, Wei Wang, Yongli Yang
Abstract: Threshold signature has security and robustness, so more and more scholars pay close attention to it. Aiming at the disadvantage that threshold signature cannot prevent conspiracy attack, this paper makes full use of tracing set and other technologies, proposes a threshold signature model that can withstand the conspiracy attack, and resolves conspiracy attack problem successfully. Moreover, the model can prevent replay attack, has forward security and does not need the support of the trusted centre. Finally, this paper proves that in the standard model, the model is unforgeable under adaptive chosen message attack.
Keywords: threshold signature; conspiracy attack.
Intelligent task scheduling strategy for cloud robot based on parallel reinforcement learning
by Xue Fei, Su Qinghua
Abstract: Cloud computing is an emerging computing model that has been developed on the basis of grid computing. Its powerful computing capabilities have evoked great vigour in its integration with various industries. The integration of cloud computing and robotics has created the concept of cloud robots. This paper proposes an intelligent task scheduling strategy for cloud robot based on parallel reinforcement learning. Firstly, the cloud computing platform is used to divide the complex reinforcement learning problem into several sub-problems. Then the task-scheduling strategy is used to assign sub-tasks to the robot to learn and summarise the results. Finally, Cloudsim builds a cloud computing platform for simulation experiments to verify the effectiveness of the proposed method. The experimental results show that the proposed method reduces the time for the robot to learn the whole problem and improves the learning efficiency.
Keywords: cloud robots; reinforcement learning; parallel computing; task scheduling; Cloudsim.
Joint relay selection and power allocation protocol for two-way relaying system
by Hui Zhi, Jun Zhu, Yanjun Hu
Abstract: The joint relay selection and power allocation (JRSPA) for amplify-and-forward (AF) relaying (JRSPA-AF) combines power allocation (PA) and relay selection (RS) very well in two-way relaying (TWR) system. In order to explore the system performance of JRSPA-AF protocol, the outage probability and diversity-multiplexing tradeoff (DMT) of JRSPA-AF protocol are analysed. One upper bound, two lower bounds and approximate expressions of outage probability along with the DMT are derived. Based on this, the idea of JRSPA is introduced into decode-and-forward (DF) relaying TWR system, and the JRSPA-DF protocol is proposed. Moreover, the exact expressions of outage probability and DMT for JRSPA-DF are derived. Simulations are developed to validate the analytical results, the results show that the approximation outage probability of JRSPA-AF can be used well to evaluate system outage probability over the entire transmission SNR region, and the analysis results of outage probability of JRSPA-DF are consistent with its simulation results. In addition, the simulation results indicate that JRSPA-DF can achieve lower outage probability than JRSPA-AF and other protocols.
Keywords: joint relay selection and power allocation; two-way relaying; outage probability; diversity-multiplexing tradeoff.
Particle swarm optimisation with multi-strategy learning
by Guohan Lin, Jing Sun
Abstract: To ease the conflict between diversity and convergence rate encountered by PSO, a multi-strategy learning particle swarm optimisation algorithm (Multi-strategy Learning PSO, MSLPSO) is proposed. The proposed method can effectively preserve the heuristic information, a modified differential mutation is combined with PSO to expand the search range and to increase the diversity of the population. If the population is trapped into local optimum, the inferior particle adopts opposition-based learning, This mechanism can improve the diversity and can help the particles to move away from the local optimum. Gaussian disturbance is applied to elite particle to further improve the diversity of particle and to the proposed MLSPSO's exploration ability. Twelve benchmark function tests from CEC2005 are used to evaluate the performance of the proposed algorithm. The results show that the proposed multi-strategy learning has performed consistently well compared with other state-of-art PSO algorithms
Keywords: particle swarm optimisation; learning strategy; differential mutation; perturbation strategy; numerical optimisation.
An improved sum-of-sinusoids based Rayleigh and Rician fading channel simulation model
by Mohit Sharma, Manpreet Dhanjal
Abstract: Sum of sinusoids has been considered as a common approach to model the Rayleigh as well as the Rician fading channel in the multipath environment. A number of models have been suggested by various authors as modification of Clarke's model to best fit the theoretical results. In this paper, a modification is proposed to the angle of arrival of the signal at the receiver to address the situations where scatterers are scarcely present in the proximity of the receiver. Therefore, the probability of reception of the signal is assumed high in pi/2 to +pi/2 than the rest of the range. The proposed model is a generalised model for Rayleigh and Rician channels adaptable to wide range of scenarios. Simulation results of the proposed model in terms of autocorrelation function, probability density function, and power spectral density function are in agreement with the reference Clarke's model.
Keywords: fading channel simulator; Rayleigh fading channel; Rician fading channel; sum Of sinusoids.
Improved Bayesian inverse reinforcement learning based on demonstration and feedback
by Hengliang Tang, Anqi Wang, Xi Yang
Abstract: A major obstacle to traditional reinforcement learning is that rewards need to be artificially set to have a strong subjectivity. The inverse reinforcement learning algorithm solves this problem. Traditional inverse reinforcement learning requires an optimized demonstration, which is often not met in reality. Therefore, an interactive learning method is proposed to enhance the learned reward function by combining the feedback with the demonstration and using the improved Bayesian rule iteration of the imagery to improve the agent strategy. The proposed method was tested in experimental and simulation tasks. The results showed that the efficiency of the method was significantly improved under different degrees of non-optimal proof.
Keywords: inverse reinforcement learning; Bayesian rule; IRLDF algorithm; demonstration and feedback.
Performance analysis of the IEEE 802.15.4e TSCH-CA algorithm under a non-ideal channel
by Soraya Touloum, Louiza Bouallouche-Medjkoune, Djamil Aissani, Celia Ouanteur
Abstract: Recently, the IEEE 802.15.4e amendment has developed a new MAC behaviour mode named Time Slotted and Channel Hopping (TSCH) to support the Industrial Wireless Sensor Networks (IWSNs) requirements. TSCH combines time slots with channel hopping and defines shared and dedicated links. A shared link is attributed to more than one sender which leads to collisions. To decrease the probability of repeated collisions in the packet retransmission, the TSCH Collision Avoidance (TSCH-CA) algorithm has been implemented by the 802.15.4e amendment. This paper proposes a two-dimensional Markov chain model to evaluate the performances of the TSCH-CA algorithm when only shared links are used under non-ideal channel conditions. The accuracy of this model has been verified through Monte Carlo simulations. Based on the proposed model, the expressions of different performance metrics that include retransmission probability, data packet loss rate, reliability, energy consumption, normalised throughput and average access delay have been obtained. Furthermore, a comparative study between TSCH-CA and the unslotted CSMA-CA of IEEE 802.15.4 under a non-ideal channel has been provided. Numerical results reveal that the TSCH-CA performances are clearly affected by channel errors when using only shared links under a noisy environment.
Keywords: IWSNs; IEEE 802.15.4e; TSCH-CA; non-ideal channel; modelling; Markov chains; performance analysis.
Performance of UWB communication systems in the presence of perfect/imperfect power control MAI and IEEE802.11a interference
by Ehab Moustafa Shaheen
Abstract: This paper evaluates the bit error rate performance of ultra-wideband (UWB) communication system under the impact of both multiple access and narrowband interferences operating. The narrowband interference (NBI) signal is modelled as the IEEE802.11a orthogonal frequency division multiplexing based wireless local area network signal, which can be approximated by the sum of independent asynchronous tone interferers with arbitrary frequencies. Multiple access interference (MAI) is assumed a zero mean Gaussian process, where it has been investigated in both perfect and imperfect power control scenarios. The bit error rate performance is evaluated in three different channel models: ideal channel model (additive white Gaussian noise channel), Nakagami-m multipath fading channel model and the IEEE802.15.3a UWB channel model. It is shown that the performance of a UWB system is severely degraded owing to the presence of both types of interference, yet the NBI has more impact on the performance of UWB communication systems compared with the multiple access one.
Keywords: impulse radio ultra-wideband; narrow band interference; IEEE802.11a; multiple access interference; Nakagami-m multipath fading channel; IEEE802.15.3a UWB channel model.
An efficient software/hardware architecture for cooperative transmission in wireless sensor networks
by Nesrine Atitallah, Kais Loukil, Hela Hakim, Mohammed Bensalah, Mohamed Abid
Abstract: Energy-efficient communication has gained increasing attention in the last few years owing to the limited-energetic resources in Wireless Sensor Networks (WSNs). Most existing works use limited-resource Sensor Nodes (SN) with a software implementation. This is often not efficient enough for intensive computing tasks such as relaying algorithms. Consequently, the Software/Hardware (Sw/Hw) co-design reaches a good compromise between energy efficiency and performances (workload and real-time). This paper investigates the design and implementation of a Cooperative Transmission Technique (CTT) for low power communication in WSN using two architectures. The first one is a pure software implementation based on an embedded soft-core processor. The validation of the technique mentioned above in real design is performed. Despite the use of a software core to optimise the processor, performances were limited. The identification of the most critical time and energy consuming tasks are performed using the profiling technique. An efficient Sw/Hw architecture (second architecture) based on soft-core processor with accelerator modules is deduced. Indeed, hardware acceleration of the CTT is implemented on a System on Chip (SoC) to achieve low power consumption level, reduced execution time as well as low resource occupancy. Experimental results of the Sw/Hw implementation demonstrate significant improvements in terms of performance metrics comparing with the software implementation.
Keywords: cooperative transmission technique; power efficiency; sensor node design; software/hardware co-design.
Research on the evaluation method of surface roughness in carbon fibre-reinforced plastic grinding based on mobile computing
by Penghua Xie, Ming Lv
Abstract: With the rapid development of mobile computing techniques, mobile users can share the service of the computing capability and resources provided by their mobile devices and other surrounding devices. However, privatisation, resource-constrained and mobility of mobile devices are challenges for data extraction and processing. As a kind of new material, Carbon Fibre-Reinforced Plastic (CFRP) has excellent properties such as light weight, high strength, good toughness, excellent corrosion resistance and design ability. Nevertheless, owing to the anisotropy of CFRP, it is difficult to follow the traditional evaluation method of the surface quality of single-phase materials after machining. In order to take full advantage of computing image processing technology in a mobile computing environment, this paper focuses on analysing the surface topography of CFRP after grinding and proposes a novel method for the surface roughness evaluation based on micro-vision using texture features value extracted by Gabor transform. By analysing the relationship between texture features value and surface roughness three-dimensional evaluation parameter Sq, the paper comes to a conclusion that the surface roughness can be evaluated by the texture features value entropy value (H). The experiments show that the method is superior to the evaluation of surface roughness in CFRP grinding with slight difference, and the convenience of image processing is improved by using mobile equipment and wireless network.
Keywords: carbon fibre-reinforced plastic; surface roughness; Gabor transform; micro-vision; evaluation; mobile computing.
Design of a miniaturised triple-band metamaterial antenna loaded with complementary split ring resonator and partially defective ground structure for wireless applications
by Sudha Malik, Garima Srivastava
Abstract: A novel low-profile, compact, and low-cost CSRR loaded antenna is proposed and simulate,d which is placed over a partially defective ground structure (PDGS). The illustrated compact triple-band antenna is fed by co-axial cable. This proposed antenna is concurrently satisfying WLAN, WiMAX and X-band communication link application requirements, providing a significantly wide impedance bandwidth (S11 < −10 dB) in the WLAN, upper WiMAX frequency and X-band communication link regions. The propounded antenna has high gain and efficiency in the three bands of operations.
Keywords: WiMax; PDGS; CSRR; return loss; impedance bandwidth.
Quality of experience prediction model for video streaming in SDN networks
by Abar Tasnim, Benletaifa Asma, Sadok El Asmi
Abstract: To get an idea of the quality of a network, the majority of stakeholders (network operators, service providers) rely on quality of service. This measure has shown limits and a great deal of effort has been put into putting in place a new metric that more accurately reflects the quality of service offered. This measure is known as quality of experience (QoE). The QoE reflects the user's satisfaction with the service. Today, evaluating the QoE has become paramount for service providers and content providers. This necessity has pushed us to innovate and to design new methods to estimate the QoE. Our work in this paper comprises two parts: the first part defines our subjective method which evaluates the quality of video streaming over SDN networks, according to the principle of DCR (Degradation Category Rating), and study the effect of the QoS parameters (packet loss, delay, bandwidth), application parameters (resolution, bit rate, frame rate) on user perception MOS (mean opinion score). In the second part we try to cover the impairments of subjective methods since there are expensive, take a lot of time and not in real time by novel method that predicts the quality of experience MOS based on machine learning, so we employ 3 types of classifiers (decision trees, meta classifiers, functions classifiers) with different k-fold cross-validation then we calculate RMSE, r, MAE, RRSE to measure the performance of each algorithm to deduce the best one. After the analysis, we obtain when k = 9, M5P and Random forest are the best algorithms in our model, which helps us to predict the perception of the user.
Keywords: SDN; machine learning; MOS.
Capacity of dual branch MRC system for correlated imbalanced average SNRs over Nakagami-m fading channels using OPRA scheme
by Amujao Yengkhom, Pukhrambam Bijaya Devi, Mohammad Irfanul Hasan, Gitanjali Chandwani
Abstract: The closed-form expressions of the capacity for a dual branch MRC (maximal ratio combining) diversity system with correlated and imbalanced average SNR (signal-to-noise ratio) over Nakagami-m fading channels is presented. Our approach is based on probability density function to find the mathematical expressions for OPRA (optimal power and rate adaptation) scheme. Numerically evaluated results are given and compared with the capacity results of OPRA scheme available in the literature. The effect of different practical constraints, i.e. fade correlation and level of imbalanced in average SNRs on the channel capacity of the systems, is analysed.
Keywords: channel capacity; dual branch; maximal ratio combining; optimum rate adaptation.
Mobile health framework based on adaptive feature selection of deep convolutional neural network and QoS optimisation for benign-malignant lung nodule classification
by Xiao Wang, Huiming Gao, Juanjuan Zhao, Sanhu Wang
Abstract: In order to explore the potential of deep learning (DL) methods in mobile health, we propose a novel framework combined enhanced DL methods and quality of service (QoS) optimisation for lung nodule classification. First, for classification-based DL methods, the methods of feature extraction and feature selection are widely used as the key steps in the classification of lung nodules. This paper proposes an adaptive feature selection method based on deep convolution neural network (DCNN). Based on the idea of transfer learning, we firstly use a DCNN model pre-trained on ImageNet database to extract the features of multi-channel lung nodule images, and then we use adaptive feature selection method to extract sparse activation features. The experiment results show that the proposed method does improve the performance of benign and malignant lung nodule classification, which can achieve the classification accuracy of 89.30% and the AUC of 0.94.
Keywords: mobile health; QoS optimisation; deep convolutional neural network; adaptive feature selection; lung nodule; classification.
Research on intelligent assistant diagnosis method of CT Image for lung nodule based on mobile computing
by Jun Lv, Huayu Wu, Yan Qiang, Jihua Liu
Abstract: With the rapid development of mobile computing techniques, mobile users can share the computing capability and resource of their mobile devices and other surrounding devices as services. However, privately-owned, resource-constrained and mobility of mobile devices pose challenges to data extraction and processing. The automatic diagnosis of benign and malignant pulmonary nodules is of great significance for the further treatment of patients. Deep learning technology has achieved good results in image classification tasks. However, the real-time nature of the physician's rapid diagnosis and the comprehensive understanding of the patient's condition cannot be guaranteed. Taking into account the three-dimensional nature of clinical pulmonary nodules, in a mobile computing environment, an algorithm that combines three-dimensional deep and visual features (CTDDV) is proposed to achieve the classification of benign and malignant pulmonary nodules. The algorithm extracts deep features based on three-dimensional convolutional neural networks (3D CNN), and also visual features based on three-dimensional scale invariant transformation (3D SIFT) texture descriptors, as well as shape descriptors based on three-dimensional shape index (3D Shape Index). Multiple kernel Adaboost (MKAdaboost) classifiers were trained for each type of feature, and the results of the three classifiers were combined at the decision level to distinguish lung nodules. We compared the four most advanced nodule classification methods on the LIDC-IDRI dataset. The results showed that our proposed CTDDV algorithm have achieved higher classification performances.
Keywords: mobile computing; intelligent assistant diagnosis; feature fusion; pulmonary nodules.
Modal analysis of flying shear crankshaft system architecture
by Chunming Xu, Qiang Yin, Bowen Luo, Chao Xu
Abstract: The system structure of the crankshaft plays a very important role in the whole shearing mechanism. When working in flying shears, the system structure of the crankshaft will bear a periodic load. Therefore, if the crankshaft size is not properly designed, the flying shear will produce strong resonance in the working speed range, which will lead to the damage of the crankshaft system architecture. If the system architecture of the crankshaft is damaged, other components will also be damaged, resulting in the failure of the flying shear to work properly. Therefore, it is very important to calculate the natural vibration frequency of the crankshaft during operation. This is necessary for the design of the crankshaft size, the research of the crankshaft system structure and service life, and provides a basis for further theoretical research.
Keywords: flying shears; crankshaft; modal analysis.
A hybrid algorithm for secure cloud computing
by Debasis Das
Abstract: Cloud computing is a developing technology that is yet unclear to many security issues. Data in the untrusted clouds can be encrypted using an encryption algorithm. Randomising this data provides more security which can be achieved by padding concept in the cloud. In this paper, the users data is encrypted using padding scheme, called Optimal Asymmetric Encryption Padding (OAEP), together with Hybrid Encryption algorithm that is based on RSA (i.e., HERSA), to allow multiple parties to compute a function on their inputs while preserving integrity and confidentiality. The homomorphic encryption (HE) is performed on the encrypted data without decrypting it in computationally powerful clouds, and the Secure Multi-Party Computation (SMPC) can be used in the cloud to ensure security and privacy of the users. In this paper, we propose a scheme that integrates the multi-party computation with HE to allow calculations of encrypted data without decryption. The cryptographic techniques used in our cloud model are described, and the overheads are compared with HE and multi-party computation.
Keywords: cloud computing; optimal asymmetric encryption padding; homomorphic encryption; multiparty computation.
Multi-agent list-based noising algorithm for protein structure prediction
by Juan Lin, Yiwen Zhong, ENa Li
Abstract: Protein structure prediction (PSP) with AB initio model problem is a challenge in bioinformatics on account of high computational complexity. To solve this problem within a limited time and resource, a multi-agent list-based noising (MLBN) algorithm is presented. MLBN contains three main features. First, a flexible noising list is designed to adjust the solution acceptance condition according to the convergence. An adaptive multiple sampling strategy is included to provide a strong exploitation. A parallel framework explores the searching space in a more effective way. Compared to traditional simulated annealing algorithm, MLBN introduces only one extra parameter for the length of noising list and it is insensitive to specific problems. Experiments conducted with a range of protein sequences indicate that MLBN performs better than, or at least comparably with, several state-of-the-art algorithms for PSP.
Keywords: noising method; list-based adaptive sampling; protein structure prediction; multi-agent.
A new way of achieving multipath routing in wireless networks
by Abd El Djalil Temar, Mustapha Geuzouri, Nader Mbarek
Abstract: In the 21st century, wireless technology is still developing rapidly and trying to be 'faster, higher, and stronger': faster data rate, higher bandwidth and stronger connectivity. Wireless Mesh Networks (WMN) have been envisioned as an important solution to the next generation wireless networking which can be used in wireless community networks, wireless enterprise networks, transportation systems, home networking and last-mile wireless internet access. They also provide a cheap, quick and effective way for building wireless data networks. Considering the nature of these networks, routing is a key process for operating the WMN. This paper proposes a new way of creating multipath routing protocols based on a fusion between multicast and unicast routing protocols to get different routes (node-disjoint or link-disjoint) from the source to its destination. The simulation results using Network Simulator 2 (NS-2) show that our new Multipath Routing Protocol (Fusion) outperforms Ad hoc On-Demand Distance Vector (AODV) and Ad hoc On-demand Multipath Distance Vector (AOMDV) routing protocols in terms of average network throughput, end-to-end delay, and number of flows.
Keywords: wireless mesh networks; routing; multipath routing protocols; multicast and unicast routing protocols; throughput; end-to-end delay.
Design and simulation optimisation of a new pressure sensing system
by Wang Qixin, Zhang Biao, Ke Wende, Chang Lin, Leng Xiaokun
Abstract: The depth simulator can simulate the depth variation of seawater in real time and is an important part of the semi-physical simulation system of underwater equipment. In order to detect the high pressure resistance and pressure detection accuracy of underwater operation equipment, the deep sea simulator is required to restore the ocean pressure environment as much as possible. This paper proposes a pressure sensing system with a stop device. By integrating the advantages of the large-range pressure sensor, the sensing system has the characteristics of large range and high precision. The physical samples were made and tested, and the friction of the sealing ring was found to be too large. Based on the multi-objective optimisation program, Matlab was used to simulate and optimise, and the reasonable structural parameters of the pressure sensing system were obtained.
Keywords: depth simulator; high precision sensor; multi-objective optimisation.
Sub-word attention mechanism and ensemble learning-based semantic annotation for heterogeneous networks
by Liang Zhang, Zhaobin Liu, Jinxiang Li, Gang Liu, Yuanfeng Yang, Yi Jin, Xu Zhang
Abstract: The sensing device and wireless sensor networks (WSN) can provide information to the application of Internet of Things (IoT), but tens of thousands of different types of sensing device and the obtained data present significant polyphyly and heterogeneity, which poses challenges to the collaboration and interaction of information resources in IoT applications and services. It is difficult to unify the identification of different types of wireless sensing device and their data, especially when the device uses Chinese characters for information and knowledge representation. This paper introduces an ensemble learning model based on sub-word attention mechanism and bidirectional long short-term memory model (SWAT-Bi-LSTM) which can provide an internal structural attention ability of Chinese characters. The experimental results show that compared with the method without using the internal structural attention mechanism of Chinese characters, this method can effectively improve the accuracy of sentiment analysis, and the integrated learning model can further improve the accuracy and stability.
Keywords: sentiment analysis; sub-word units; Bi-LSTM; ensemble learning; wireless sensor networks.
Bimanual gesture recognition based on convolution neural network
by Hao Wu, Gongfa Li, Ying Sun, Guozhang Jiang, Du Jiang
Abstract: Gesture recognition is a key research field in human-computer interaction. At present, most researchers focus on one-handed gesture recognition, but do not pay much attention to bimanual (two hands) gesture recognition. This paper presents a deep learning-based solution to tackle the self-occlusion and self-similarity. To solve this problem, this paper uses Kinect to collect many colour and depth images of different gestures, and each gesture contains multiple sample individuals. Colour images and depth images are used to train the recognition model of bimanual gesture, and then fused the colour image and depth image, and train the bimanual gesture recognition model based on colour image and depth image fusion. Then the bimanual recognition effects of three models are compared. The experimental results show that, regardless of the single gesture precision or the mean average precision, the bimanual gesture recognition effect of the fused model is better than the gesture recognition models based on colour image or depth image.
Keywords: gesture recognition; bimanual gesture; deep learning; CNN; occlusion.