International Journal of Innovative Computing and Applications (41 papers in press)
Implementation of Fuzzy Logic Controller based Quadratic Buck Converter for LED Lamp Driver Applications.
by Ravindranath Tagore Yadlapalli, Anuradha Kotapati
Abstract: This paper focuses mainly on design of quadratic buck converter (QBC) for LED lamp driver applications. The LED current regulation is the critical issue in the family of LED lamp drivers. The continuous mode based QBC is well designed for 60V/20mA at 100 kHz. The QBC performance is analysed with Digital average current mode control (ACMC) based QBC and fuzzy logic control-ACMC based QBC. The simulation is fulfilled using MATLAB/Simulink software.
Keywords: Pulse Width Modulation; Amplitude modulation (FLC); Organic LEDs.
Hybrid Algorithm for Materialized View Selection
by Mayata Raouf, Boukra Abdelmadjid
Abstract: Data warehouses store current and historical data, which are used for creating reports, for the purpose of supporting decision-making. A data warehouse uses materialized views in order to reduce the query processing time. Since materializing all view is not possible, due to space and maintenance constraints, materialized view selection became one of the crucial decisions in designing a data warehouse for optimal efficiency. In this paper the authors present a new hybrid algorithm named (QCBO) based on both quantum evolutionary algorithm (QEA) and colliding bodies optimization (CBO) to resolve the materialized view selection (MVS) problem. Also, some aspects of the well-known greedy algorithm (HRU) are included. The experimental results show that QCBO provides a fair balance between exploitation and exploration. Comparative study reveals the efficiency of the proposed algorithm in term of solution quality compared to well-known algorithms.
Keywords: Data warehouse; materialized view selection; metaheuristic; quantum evolutionary; colliding bodies optimization.
A comparison between evolutionary and local search techniques applied to NoC Design Space Exploration
by Jefferson Silva, Silvia Maia, Monica Magalhães, Márcio Kreutz
Abstract: Networks on chip (NoCs) emerged as a communication architecture that overcame the limitations of bus architecture. As more cores were being incorporated in a single die, shared communications architectures reached out limitations in terms of scalability and performance. However, the NoC based communication architecture has many configuration parameters leading to a huge design space to be covered during the design phase. Aiming to find optimized configurations, some methods should be employed to explore the design space and speed up this process. This paper investigates three techniques, Genetic Algorithm (GA), Memetic Algorithm (MME) and Iterated Local Search (ILS). Results have shown the lowest execution time for ILS when compared with the other two. In relation to the quality of the solution, the MME overcame both in almost all scenarios, even considering execution time or number of evaluation as stop criteria.
Keywords: evolutionary computing; networks on chip; genetic algorithm; iterated local search; memetic algorithm; design space exploration.
Human motion tracking under indoor and outdoor surveillance system
by Wafae Mrabti
Abstract: This paper gives an overview of the various potential applications involved in human motion tracking. Also, a review of some relevant algorithms in this area are summarized based on three main components; the human target extraction, the features extraction, and the motion model. In addition, a proposed generative and discriminative human tracking method is presented. This proposed method is based on the hybridization of Kalman Filter (KF) and the Support Vector Machines (SVM). Numerous experiments illustrate the effectiveness of the proposed method against several state of the art trackers. These experiments are applied on several real world image sequences with various challenges that make the tracker vulnerable which are; partial and full occlusions, illumination changes, scale variation, out-of-plane, rotation, motion blur, low resolution, deformation, fast motion and background clutter.
Keywords: Features extraction; Motion model; Kalman filter; Histogram of Oriented Gradient; Support Vector Machines.
Plant propagation algorithm for nurse rostering
by Salim Haddadi
Abstract: This paper investigates the nurse rostering problem (NRP), a challenging combinatorial optimization problem that arises in heath care institutions. We propose to solve it by using the plant propagation algorithm (PPA). As many successful metaheuristics, PPA is inspired by a life process. It emulates the strategy of reproduction and propagation of the strawberry plant. Before applying PPA, a variable-fixing procedure is used for heuristically discarding variables. In practice, it results in removing up to 99% of the variables without sacrificing solution quality. Elite solutions provided by PPA are used to further discard variables, leaving a very sparse NRP that can be solved directly by an IP solver. Computational and comparative results are presented on a widely used set of benchmark instances.
Keywords: Bio-inspired computation; plant-inspired algorithm; plant propagation algorithm; nurse rostering; elite solutions; variable-fixing.
MODIFIED BIO-INSPIRED ALGORITHMS FOR DIAGNOSIS OF BREAST CANCER USING AGGREGATION
by Moolchand Sharma, Shubbham Gupta, SUMAN DESWAL
Abstract: The most widely detectable of all cancers found in women is breast cancer. The mortality rate is also the second-highest among women with a 12% growth rate. It is very pertinent to diagnose breast cancer in the nascent stages so that the survival of the patient is ensured with the help of proper medication. Several algorithms have been proposed in this regard. However, they have failed to achieve the desired level of accuracy. An improved version of the Particle Swarm Optimization & Firefly Algorithm is presented in this paper to overcome the drawbacks of the existing algorithm. The two algorithms are further aggregated to improve the accuracy of the results. The aggregated algorithm is used on the Breast Cancer Wisconsin (Diagnostic) Data Set(real-valued dataset), and results are calculated for different classifiers. An accuracy of 92%-96% is shown by Improved Particle Swarm Optimization and 1%-2% overall hike in the accuracy by Improved Firefly Algorithm, respectively. Finally, the aggregated algorithm shows an accuracy of 93%-97%. Further, Random Forest Classifier has displayed the best accuracy of 97%.
Keywords: Breast Cancer; Bio-inspired Algorithms; Aggregation; Particle Swarm Optimization; Firefly Algorithm; Feature selection; Linear Support Vector Machines; Decision Tree; K-Nearest Neighbor; and Random Forest Classifiers.
Fuzzy Control Mathematical Modeling Method Based On Dynamic Particle Swarm Optimization Training
by Runxia Gao, Houping Jiang
Abstract: Aiming at the false correlation problems under set number limit condition in the fuzzy control mathematic modeling process, combined with dynamic particle swarm optimization training algorithm, the traditional fuzzy control mathematical model based on the Nash equilibrium solution method is difficult to converge to the optimal solution of the state space, leading to bad control performance. This paper proposes the fuzzy control mathematical modeling method based on dynamic particle swarm optimization training, constructs the general structure model of fuzzy control, describes the standard particle algorithm, under the constraint of learning samples of random functional, obtains the global optimal solution of control domain of fuzzy control parameters, and conducts particle swarm optimization training by adopting the position vector fitness updating method. The research results show that the new method can make every step state update get more effective observation information, reduce the error caused by the difficult use of observation data, reduce the computation cost and improve the accuracy of fuzzy control.
Keywords: fuzzy control; observation error; ensemble transform kalman filter (ETKF).
Special Issue on: Leveraging Opportunistic Networks Using Smart Wireless Digital Devices and IoT for Intelligent Communication Systems Challenges and Emerging Trends
A Novel Genetic Algorithm with CDF5/3 Filter Based Lifting Scheme for Optimal Sensor Placement
by Ganesan T., Soumya Ranjan Nayak, Amandeep Singh Bhatia
Abstract: The generic algorithm has been receiving significant attention due to node placement problem in the field of sensor application in terms of machine learning. Sensor deployment is able to provide maximum coverage and maximum connectivity with less energy consumption to sustain the network lifetime. The maximum quality coverage problem has been solved successfully by evolutionary algorithm while placing node in optimal position. In evolutionary algorithm, genetic algorithm (GA) plays an important technique for deploying the sensor in the form of population matrix. However, the existing techniques unable to place sensor position perfectly. In this paper, a novel genetic algorithm with second generation wavelet transform (SGWT) is proposed for identifying optimal potential position for node placement. In order to improve the quality of population matrix, bi-orthogonal CohenDaubechiesFeauveau wavelet (CDF 5/3) has been employed. The proposed method is performed primarily to generate sensor position with different population. Subsequently, it can extend to CDF5/3 filter based lifting scheme to adjust the sensor position. The proposed method has been compared with random deployment, genetic algorithm and GA with CDF5/3 wavelets in terms of target to cover by the sensor. The result of the proposed method affirms better optimization as compared to state-of-art techniques.
Keywords: Wireless senor network; Sensor deployment; Genetic algorithm; Lifting scheme; Target coverage.
Semantics Interoperability in the Internet of Things- State of the Art and Prospects
by Jameel Ahamed, Mohammad Ahsan Chishti
Abstract: Internet of Things is generally referred to as an internet world of little things with limited storage and processing ability, but facing communication, security, privacy, and interoperability issues. Further, Device, Network, Platform, Syntactic, and Semantic are the major interoperability issues in the Internet of Things. Smart objects, termed as interconnected devices, generate a huge chunk of heterogeneous data and its processing as well as usage becomes the challenge in IoT. Therefore, this heterogeneous Data is semantically annotated using linked vocabularies to make it interoperable information. The main aim of this article is to have a deep understanding of semantic interoperability that will help in sharing heterogeneous data through the development of knowledge-based systems or common ontologies.
Keywords: Internet of Things; Semantic Interoperability; Ontology; Heterogeneous Data.
Impact of Contacts for Message Copies, Mobile Nodes and Buffer Size in Delay-Tolerant Networks
by Md. Sharif Hossen, Khaleel Ahmad, Khairol Amali Bin Ahmad
Abstract: Delay tolerant network (DTN) is a kind of mobile ad hoc networks where there are no predefined routes from one node to other. This network is used often in some areas of applications such as disaster management, military communications, emergency communications, and wild life tracking services where the fixed infrastructures are not found. Under this network, the analysis of node contact is essential, without which, message copies will not be possible to deliver successfully from the source to the destination. In this research study, we take our concentration on single and multi-copy contact duration aware routing techniques and address the issues of the variation of mobile nodes, message copies and nodes buffer size. We have used a java unified opportunistic network environment (ONE) simulator. We observe that contact increases per hour with the increase of mobile nodes, while goes up and down for varying message copies and remains approximately the same for changing buffer size, i.e., buffer size does not affect the achievement of contacts. On the other hand, inter-contact time (ICT) is high at the beginning of the communication for three cases, i.e. for varying message copies, mobile nodes and buffer size. Higher contact time will result in higher ICT, and higher number of contacts in case of varying mobile nodes, while it will be slightly changed for varying message generation rate and fixed for varying buffer size.
Keywords: Ad hoc network; contact; delivery; latency; message copies; routing; simulation.
Parameter Aware Utility Proportional Fairness Scheduling Technique in Communication Network
by Neeraj Kumar, Anwar Ahmad, Shashank Awasthi, Kavita Sharma
Abstract: The current paper suggests a parameter aware utility proportional fairness scheduling technique, in short PAPUF, for mixed traffic users. Both elastic traffic type (non-real time) users or inelastic traffic type (real-time) users are scheduled as per their corresponding performance metrics. Conventional techniques separately consider real-time users and non-real time users in resource scheduling. In different scenarios, such as high network traffic, less network coverage, and the case of radio resource constraint, the joint consideration of both elastic and inelastic traffic type users is needed for resource allocations. Further, some of the users can be left without getting radio resources in conventional rate allocating techniques. In the proposed technique, a utility function is defined for each elastic traffic and inelastic traffic type users. Then, it maximizes the total utilities by network utility maximization approach with considering scheduling parameters such as queue length, head of the line delay and experienced channel condition corresponding to each elastic and inelastic traffic user. Rate allocation based on queue awareness improves fairness among elastic traffic performing users, delay awareness improves QoS performance of inelastic traffic performing users and channel condition awareness improves the throughput of all users. Further, the network utility maximization approach provides minimum throughput to all users i.e. no users will be left without rate allocation. Simulation results of the proposed scheduling technique are compared with existing techniques, which show improved communication performance for users.
Keywords: Utility Function; Proportional Fairness; Rate Allocation; Throughput; Queue; Delay; Logarithmic Function; Sigmoidal Function;.
TASRP: A Trust Aware Secure Routing Protocol for Wireless sensor networks
by Tayyab Khan, Karan Singh
Abstract: Over the last decade, considerable secure routing protocols with static sink have been developed for wireless sensor networks (WSNs) to achieve reliable routing paths as well as energy efficiency during information gathering and transmission. Security is vital for sensor networks since sensor devices communicate essential and private data. However, cryptographic security solutions have not proved suitable for wireless sensor networks since they impose high overhead as well as do not deal with severe internal attacks.
Trust-based security models are efficient and reliable over traditional cryptographic schemes (e.g., encryption, key management schemes) to detect and alleviate various internal threats by estimating the trust scores of sensors nodes in a quantitative way. Trust models analyze the trustworthiness of each sensor node to improve reliability, quality of data, prevent damage, and adversely affect. This paper presents a realistic trust-based reliable routing (communication) strategy to counter selfish nodes based on the hybrid trust model. The proposed scheme (TASRP) is a multifactor routing approach that employs trust scores of nodes, residual energy, and path length to provide reliable routing paths among trusted nodes with reduced energy consumption.
This multi-factor strategy helps in selecting trusted nodes to forward data and minimize energy consumption due to shorter routing paths. Simulation results show better performance in terms of robust trust values, throughput, packet delivery rate, and energy consumption of nodes.
Keywords: WSN Routing Protocols; Trust; Security; Internal Threats.
Special Issue on: Security, Privacy and Trust in Cognitive-inspired Computing and Applications
Safety Evaluation on Central Separation Opening of Reconstructed Freeway Based on Surrogate Safety Assessment Model
by Wei Hou, Zijun Du, Zhaoxin Liu, Xu Wang
Abstract: In order to reduce the risk of driving at the opening of the median strip of the old road of the freeway, the driving safety was quantified by taking the opening length of the central section of the old road and the main line flow as independent variables. Based on the VISSIM micro-simulation software, a typical simulation scenario is established. The vehicle speed, the position of the vehicle's change point, the number of conflicts, the severity of the conflict, and the conflict points are analyzed by the Surrogate Safety Assessment Model (SSAM). The influence of the opening length of different median strips and the influence of flow on safety is quantitatively evaluated, so as to obtain a reasonable opening length of the median strip. The research results show that the flow has a greater impact on the number of collisions, and the length of the open section has a significant impact on the severity of the conflict. The conflict points are found to be concentrated at the front of the 400m open section. The results of the study provide a theoretical basis for the opening length of the median strip and the setting of traffic signs.
Keywords: safety evaluation; conflict; trajectory; median strip; freeway; single side widening.
Special Issue on: Artificial Intelligence for Sustainable Future Computing
Semantics-based Key Concepts Identification for Documents Indexing and Retrieval on the Web
by Mohammed Maree
Abstract: Bridging the semantic gap on the web remains one of the crucial challenges for current horizontal as well as domain-specific information retrieval systems. This challenge becomes even more pronounced when users express their information needs using short queries that are formulated using a few number of keywords. Therefore, relying on keywords for indexing web documents results in degrading the quality of the returned results. To tackle this issue, new approaches across multiple disciplines propose to incorporate semantic resources to overcome the query-document mismatch problem. Although these approaches have proved to assist users in finding relevant documents, many results are still irrelevant; failing to adequately meet the desired information needs. This is because of the imprecise expansion of query terms and the limited depth and breadth of the employed resources. In this article, we discuss these limitations and introduce an approach that employs knowledge captured by large-scale knowledge resources to identify key query terms for retrieving semantically-relevant documents. Unlike conventional approaches that treat query terms independently, key terms are mapped to their semantic correspondences and variable term weights are assigned based on the semantic and taxonomic relations for each term. To demonstrate the effectiveness of the proposed approach, we have conducted experimental evaluation using Glasgows NPL test collections. Findings indicate that precision results have improved by employing the proposed method against four conventional similarity metrics that are based on the bag of words similarity model.
Keywords: Key concepts; Large-scale ontologies; Semantic matching; Information indexing; Data analysis; Precision measures.
Hybrid approach for semantic similarity calculation between Tamil words
by Deepa Karuppaiah, Durai Raj Vincent P M
Abstract: Semantic similarity, sometimes referred as semantic relatedness, is one of the important concepts that help in various applications that involve Natural Language Processing. In literature, there are plenty of similarity measures to compute the relationship among words in monolingual and cross-lingual documents. They help us in understanding text, finding plagiarism, information retrieval etc. They can be categorized based on the resources used into corpus based and knowledge based measures. These measures are plenty for English language. For Tamil language, hardly there are any works in calculating the similarity between words. In this paper, we proposed a similarity finding technique that exploits the knowledge from the resources like Tamil Indo Wordnet, Tamil Wikitionary and Oxford Tamil Dictionary. We have used the definitions and example sentences of each word that are available through each of these resources for similarity calculation. The proposed approach is evaluated using human evaluated Miller Charles and Rubenstein Goodenough datasets.
Keywords: Semantic similarity; Tamil words similarity; Indo Wordnet; Knowledge based similarity.
Development of Scheduling Algorithm for Reconfigurable Architecture using FPGA
by Pushpa Bangare, M.B. Mali
Abstract: The present title discloses a scheduling algorithm for scheduling multiple tasks to allow multiple tasks to reconfigure in reconfigurable architecture. In the proposed architecture, multiple tasks are considered. With the provision of operating multiple tasks concurrently or on priority-based, routine tasks are executed on the ideal state of the higher priority tasks. A concurrent round-robin scheduling architecture is proposed to execute the task which needs to be executed through reconfigurable architecture. The proposed architecture is described using Very High-Speed Integrated Circuit Hardware Description Language. For systematic verification, timing simulation is done through the Modelsim simulator. Synthesis is targeted using the Xilinx ISE platform for Xilinx FPGA devices.
Keywords: Task scheduling; field programmable gate array; round robin scheduling; general purpose processor.
Face Pose and Blur Normalization for Unconstraint Face Recognition from Video/Still Images
by T. Shreekumar, K. Karunakara
Abstract: Face recognition has reached maturity level using Deep learning which works on a very large data set of Face images. In a considerable lot of circumstances, it is exceptionally hard to get the huge number of Face Images for the confirmation purpose, particularly from Village individuals. The main bottle-neck in face recognition are Pose variation, illumination variation, Occlusion, and Noise. To solve these problems we are presenting a system that will recognize the countenances from an extremely small dataset. This method initially removes the blur from the image and, then normalizes the Pose by generating virtual frontal Face using the Local Linear Regression (LLR) method. Then the combined score of Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) are obtained from the image. This score is used to identify the Face using Support Vector Machine (SVM) which gives more accurate results compared to simple Euclidean measures. The experiment results show a maximum accuracy of 93.5% .
Keywords: Face Recognition; Support Vector Machine; Motion Blur; Principal Component Analysis; Local Linear Regression. Pose Variation.
Modified Iterative Learning Controller (MILC) for efficient power management of Hybrid AC/DC Micro grid
by Angalaeswari Sendraya Perumal, K. Jamuna
Abstract: In this paper, modified ILC is proposed for maintaining stable voltage and frequency and performing efficient power management in a hybrid micro grid. The hybrid micro grid (HMG) is modeled with solar, battery, DC loads at DC bus and wind turbine, utility grid, AC loads at AC bus. An interlinking converter (IC) is connected between the AC and DC bus to facilitate bidirectional power flow. Due to the intermittent nature of the distributed sources and the variable loads, the voltage and frequency deviation is occurred. The voltage control at dc and ac bus has to be done and the power balance has to meet in all modes of operation. The proposed objective has been obtained in the modeled hybrid micro grid with Set Point Weighting Iterative Learning Controller (SPW-ILC) by controlling the Interlinking Converter both in autonomous and grid connected mode of operation. To minimize the error signal to the controller, the classical optimization method of Sequential Quadratic Programming (SQP) has been employed to improve the performance of the controller. The simulation results show that the proposed controller have better performance than other controllers under variable source and load conditions.
Keywords: Iterative Learning Controller; Hybrid Micro grid; Power Management; Sequential Quadratic Programming; Voltage stability.
A hybrid cum enhanced minutiae feature extraction approach for security oriented Content based fingerprint image retrieval for patient history identification and authentication
by Raghavan R, John Singh K
Abstract: The role of security in fingerprint is having lot of available features like ledger maintenance in a distributed fashion, security through finger print authentication and validation, data access. There are many forms of security technologies which are already in existence in the domain of company view point as well as institute viewpoint. Both the companies and institutes have started to discover applications towards healthcare. Some of the applications include smart contracts, fraud detection and identity verification. Identity verification in healthcare industry sometimes facing challenge in correctly identifying the patients in terms of medication errors, errors while undertaking tests, some form of error related to transfusion, and more particularly discharge of infants to the wrong families. In this paper we proposed an improved content based fingerprint image retrieval system for patient history identification application. A measure is proposed as a pre-step to avoid handling the less clarity input with a noisy image using median filter to initiate the improvisation of retrieval process. Binarization is done using rough sets and thinning processes are further carried out after the initial steps. Then using a novel proposed approach using rough sets the features are extracted from the noise free finger print image based on the minutiae details. After initiating the proposed matching algorithm using rough sets the correct patient history details of the authorized person is fetched and delivered only to the authorized individual. The testing are performed using Multimodal dataset and IRMA dataset to evaluate the proposed retrieval technique. The accuracy obtained in fetching and validating the fingerprint in this hybrid approach is found to be more accurate and hence more secure towards avoiding misuse of patient medical data.
Keywords: Content based image retrieval; noisy image; rough sets; median filter; binarization; thinning; fingerprint; minutiae; matching algorithm; rough sets.
ANALYSIS OF PERFORMANCE OF STUDENTS USING STATISTICAL APPROACH - ANOVA TEST
by Hemamalini B H, Suma V, Suresh L, Shankar M M
Abstract: Educational Data Mining (EDM) throws light on the various techniques and strategies that affects the performance of students. The present work has considered engineering students from a reputed institution in Karnataka for study. The present paper focuses on applying one of the statistical techniques Anova Test on the student dataset. The dataset comprises of the details from the institutional repository of a prestigious institution in Karnataka. The pre-processed data consists of 1186 records. The Anova test is applied to know if the averages of two or more collections are considerably different from each other. Descriptive F value and T test is applied on the given dataset. The probability distribution is also calculated. It is claimed that the fathers occupation has an impact on the performance of student. If father is in defense, the student has performed excellent. Also, if the student is a day scholar, his performance is better than the hostel student. If the student is from ICSE or CBSE background, the student performance is better than the student from State board or other boards. By predicting the parameters that affect the performance of students, we can produce better results, thus contributing to the society. However, mothers occupation does not have any significant association with outcome variables.
Keywords: Anova; Descriptive Statistics; Multiple comparison; f-value; variance; degree of freedom; p-value; significant; T test; sumsquare; meansquare.
A comparative study of techniques for polygonal approximation of digital image boundary
by Kiruba Thangam Raja, Bimal Kumar Ray
Abstract: Polygonal approximation (PA) technique have been widely applied in the field of pattern recognition and classification, shape analysis and identification, 3D reconstruction and medical imaging, digital cartography and geographical information system (GIS). In this study, we focus on some of the key techniques used in implementing the PA algorithms. The PA can be broadly divided into three main category, dominant point detection, threshold error method with minimum number of break points and break points approximation by error minimization. Of the above three methods, there has been always a tradeoff between the three classes and optimality, specifically the optimal algorithm works in a computation intensive way with a complexity ranges from O (N2) to O (N3), on the other hand heuristic methods approximate the curve in a speedy way, however they lack in the optimality but have linear time complexity. Here a comparative study on major PA methods for digital planar curve approximation is presented.
Keywords: : contour; break point; split and merge; dominant point; polygonal approximation; digital planar curve; computation intensive; geographical information system; digital image boundary; 2D or 3D images; redundant breakpoints; Freeman chain code; Heuristic algorithms; optimal approximation; dominant points.
Sematic based Road Traffic Prediction using Moving
Weighted Average model
by Prathilothamai Manikandan, Viswanathan V
Abstract: In this emerging world, peoples are running behind the time and wasted their time in travelling. Drastic increase in population results in rapid increase of number of vehicles. A semantic based road traffic model is proposed to predict the traffic and to inform the public about the current traffic condition to all persons who belongs to the same lane. Real time data is acquired from Ultrasonic, PIR sensor and camera. Proposed system uses the vehicle count, distance between the vehicles and speed of the vehicle from both sensors and camera and it applies semantic interpretation of those data uses moving weighted average model to predict the traffic condition. In order to have time efficient prediction, the work is experimented in Apache Spark which will reduce disk latency when compared to hadoop. Prediction result is sent it as alert message to the public as a location based messages. Therefore, the traffic prediction system results are more helpful in goods transportation and accident prediction system etc.
Keywords: Big Data; Hadoop; Spark; Real-Time Applications; Road Traffic Prediction.
Person Identification using Fusion of Deep Net Facial Features
by Haider Mehraj, Ajaz Hussain Mir
Abstract: Face base identification is the method of recognizing individuals through face
images having application domains such as smart cards, mobile phones, information
security, law enforcement and surveillance system. Deep networks have proved to be
successful for facial identification and involve some pre-processing steps like sampling to be done before the images are applied. The complete images are passed as input to Deep Net, and the network does feature extraction as well as classification. However, such a process requires millions of images to work with and implementing the same sometimes becomes complex and time-consuming. This Paper utilizes Deep Networks Alex net and VGG-16 as feature extractors in which contribution to more significant level layers are utilized as feature vectors. Alex net and VGG-16 are pre-trained Deep Nets, and such the number of input images can be significantly lower in comparison to training a network from scratch. The Feature vectors are then diminished using a combination of PCA and LDA. After the reduction in the dimensionality of highlight vectors, they are intertwined and characterized using Support Vector Machines. The proposed framework is assessed utilizing freely accessible database VIDTIMIT, highlighting the performance as far as exactness or precision, accuracy, and review or recall.
Keywords: Biometrics; DNN; CNN; Alex net; VGG-16; Feature vector; SVM,
Recognition; Multi-Algorithm Biometric System; Neural Network; Deep Networks.
Mathematical modeling and Kinematic analysis of 3-RRR Parallel Planar Manipulator
by Shaik Himam Saheb, G.Satish Babu
Abstract: Parallel mechanisms are found as positioning platforms in several applications in robotics and production engineering. Today there are various types of these mechanisms based on the structure, type of joints and degree of freedom. An important and basic planar mechanism providing three degree of freedom at the end-effector (movable platform) is a 3-RRR linkage. The forward kinematics in parallel mechanisms is a multi-solution problem and involves cumbersome calculations compared to inverse kinematics. With inverse kinematics, the input kinematic parameters for a known table center coordinate are determined. In present work, the workspace and Jacobian matrices are computed at corresponding solution and dexterous workspace analysis is discussed. Main objective is to fabricate a model of this planar manipulation mechanism with calculated dimensions and observe the practical workspace and dexterous workspace available at end effector. This final output data is useful for manipulator designers to design manipulators for different applications in addition to this the stress analysis is performed with the help of Ansys software to estimate the failure zone of moving platform.
Keywords: 3RRR Parallel Planar Manipulator; Work space analysis; Regular Dexterous Workspace; fabrication of 3RRR PPM Model; Performance analysis; Stress analysis; load carrying capacity; precise; accurate position; Kinematic analysis; Ansys software; Degree of freedom; mechatronic system; Dexterous workspace; Isometric view of prototype,.
Thyroid Disease Classification with Hybrid C5.0 and Cultural Algorithm
by M. Deepika, K. Kalaiselvi
Abstract: Data mining plays a prominent role in disease classification and diagnosis. The data mining technique helps physician in making reliable and accurate disease diagnosis and prognosis. In this work the thyroid disorder is classified as hyperthyroidism and hypothyroidism based on C5.0 and Cultural Algorithm. The C5.0 algorithm reduces over fitting of data in dataset. The best cost function minimizes the load of knowledge discovery from the missing data in dataset with the help of cultural algorithm. The Cultural algorithm is derived from social evolution which includes a population space, communication protocol and a belief space. The experimental results show C5.0 paired with cultural algorithm provides better thyroid classification with minimal cost function.
Keywords: Classification; Cultural Algorithm; C5.0; Decision Trees; Thyroid analysis; Prediction.
Special Issue on: Cognitive Computing for Emerging Internet of Things
HOG Features and Connected Region Analysis-Based Workpiece Object Detection Algorithm
by Yu Ting, Tian Maoyi
Abstract: In order to solve the problem of bearing workpiece object, namely, the insuffi-cient detection ability of the algorithm caused by the complex edge features and inconspicuousness of the surface as well as the uncertainty and interference of the background, a HOG features and connected region analysis-based workpiece object detection algorithm is proposed in this paper starting from the calculation of HOG features, the image gradient direction, the connected region analysis and object detection. The image is processed in accordance with the color chroma-tography of foreign matters to separate the foreign matter from the background. Firstly, the target images of standard workpiece in the training set are meshed to calculate the pixel gradient in the grid, count the gradient histogram and com-plete the extraction and training of HOG features. Then interval division of the single peak threshold is refined, and a two-threshold segmentation mechanism is proposed to convert the two-valued image into a label image by combining the connected region analysis, and the evaluation of pixel attribute and the filtering of interference is conducted to achieve the purpose of accurately detecting the workpiece object. The experimental results show that the bearing workpiece ob-ject detection algorithm in this paper has higher accuracy and stability.
Keywords: Workpiece object detection; Image gradient; Chromatography; Edge feature; Connected region; Meshing; Histogram.
SLFNs Interpolation Fingerprint Particle Filter-Based Shared Bicycle Tracking Algorithm
by Cao Honghua, Yan Xiaoyan, Li Yan
Abstract: In order to improve the performance of traditional fingerprint detection method in the process of tracking the shared bicycle, the inertial sensor is used for data measurement. The particle filter (PF) method is a widely used sensor fusion al-gorithm, but the initialization and weighting processes are problematic in shared bicycle positioning systems. In this paper, a new PF scheme is proposed, and it can produces smooth and stable localized knowledge. However, the feed-forward network that uses the single hidden layer is used to simulate the estimation and improvement of the performance of multiple probability to achieve the distinc-tion of similar fingerprints. At the same time, the random sample consensus al-gorithm (RANSAC) is used to initialize the algorithm so as to reduce the conver-gence time. Experiments show that the tracking error of this scheme is less than 1.2m, which is superior to the selected comparison method.
Keywords: Feed-forward network; Particle filter; Shared bicycle; Tracking algorithm; Strength indicator of signal.
Markov Model-Based Low delay Data Aggregation tree Algorithm
by Huang Luyu
Abstract: The data aggregation technology can save resources of wireless sensor networks, but it can also add extra delays. To this end, specific to the special scenario where data transmission must be completed under specified delay constraints, the Markov model-based low delay data aggregation tree (MLDGT) algorithm is proposed. Firstly, the formal expression of the problem of constructing data ag-gregation tree under delay constraints is given. This problem has been confirmed as a NP problem. Then, the Markov approximate model is used to find a subop-timal solution, and further obtain the low delay data aggregation tree. Finally, the effectiveness of the MLDGT algorithm is analyzed by simulation and compari-son. The experimental results show that the MLDGT algorithm can reduce the data aggregation delay.
Keywords: Wireless sensor network; Aggregation tree; Data aggregation; Formal expression; Markov model.
MeTis Meshing-Based Bayes 3D Ship Model Geometry Reconstruction
by Yue Jingya
Abstract: In order to improve the compression efficiency of 3D model geometry reconstruction process, a MeTiS meshing-based Bayes 3D ship model geometry reconstruction algorithm is proposed. The original 3D mesh is subnetted by the MeTiS method at the coding end, and the geometrical shape of the subnet is coded by a random linear matrix, and the neighbor node of the boundary node is considered to construct the data sequence by the pseudo random number generator; then the Bayes algorithm is used to design the geometric model reconstruction algorithm, and the learning rules for the mean, variance matrix and model parameter are theoretically given, realizing the geometric reconstruction of 3D model; finally, on the 3D model standard test library and 3D ship model, the simulation comparison with the GFT, LSM and CSGFT and other algorithms show that the proposed method has a relatively high bit rate compression index and a low reconstruction error, leading to significantly improved computational efficiency.
Keywords: 3D vessel model; Geometric reconstruction; MeTiS meshing; Bayes; Neighbor node.
Dynamic Node Adaptive Incremental Interaction Optimization in Micro-blogging Community
by Fei Shang, Xiaobo Nie
Abstract: Most community discovery methods are based on network topology and edge density for best community determination, but these methods have very high computational complexity and are very sensitive to the form and type of network. In order to solve these problems, this paper proposes a micro-blogging community interaction optimization algorithm based on dynamic node adaptive increment model, which is based on optimizing the interaction of members in each community, and uses greedy algorithm to search the best candidate for the optimal community effectively without traversing all nodes. The model can quickly and accurately measure the interaction difference between the community and the community. Finally, the simulation tests on the datum test network and the Sohu micro-blogging platform show that the proposed algorithm is better than the selected contrast algorithm in the index of recall, accuracy, algorithm calculation time and network coverage.
Keywords: Complex network; Edge density; Community discovery; Self- adaptive; Interaction optimization; Incremental model.
Quantitative Structure-Activity Analysis of Predicted Drug Targets Based on Adaboost-SVM
by Fujun Gao
Abstract: This paper first constructs two sets of data sets to demonstrate the effectiveness of the proposed method, one data set consists of all human protein data, and the other is composed of human G protein-coupled receptor data, which accounts for a high proportion of drug targets. It extracts the corresponding primary structure, polypeptide characteristics and basic physicochemical properties of each protein in the data set, feature selection is used to reduce the learning burden of classifier as the feature space of training classifier. Then the data are preprocessed and the optimal classifier is constructed by adjusting the parameters of the model. Data sets are classified by SVM classifier and Adaboost-SVM classifier respectively in the experimental construction and analysis part, analyzed and compared the experimental results of two classifiers applied to two sets of data sets before and after data preprocessing, the classification results of the two groups were verified each other to increase the reliability of the classification results. The experimental results verify the effectiveness of the proposed method. At the same time, it shows that the method proposed in this paper can effectively predict drug targets, and provide a preliminary reference for drug research and development workers.
Keywords: Direct push type; Support vector machine; Predictive drugs; Target quantification; Structure-activity analysis.
Energy Sensing Streaming Media Data Transmission Protocol Based on Implicit Markov Algorithm in WSNs
by Guozhong Li
Abstract: Streaming media transmission protocols can be divided into traditional streaming media push protocol and streaming media pull protocol. Traditional streaming media push protocols such as RTP, whose server determines the channel state according to the RTP feedback from the client, and then decides to send data packets suitable for the current channel state. The pull protocol of streaming media sends data packets according to the content of the client when the transmission rate meets the requirement on the contrary. Streaming media pull protocol greatly reduces the complexity of servers in streaming media transmission technology, it can also support the application of different adaptive algorithms in the transmission process compared with the traditional universal server transmission algorithm mechanism. Therefore, this kind of pull protocol cannot only improve the quality of service of streaming media transmission, but also meet the requirements of different channel states and different clients.
Keywords: WSNs; Data aging; Mesh area; Energy aware; Data transmission protocol.
Multi-Feature Fusion Energy-Saving Routing in Internet of Things Based on Hybrid Ant Colony Algorithm
by Ren Xiao-Li, Yang Jian-Wei, Li Nai-Qian
Abstract: This paper analyzes the research status of sensor networks and several improved LEACH protocols. It is known that there are some shortcomings in current low-energy clustering protocols:The problem of uneven network cluster and unequal energy consumption of each node in the cluster group leads to excessive energy consumption of some nodes, the whole network life cycle is also greatly shortened. This paper proposes a multi-feature fusion energy-saving routing algo-rithm based on hybrid ant colony algorithm to optimize and upgrade LEACH energy-saving routing model for the Internet of Things on the basis of LEACH.(NPCHS-Leach)to improve the problems of short lifetime and low energy utilization caused by existing clustering routing protocols,it improves and prolongs the network life cycle. Finally, the effectiveness of the proposed algorithm is verified by simulation experiments.
Keywords: Ant colony algorithm; Multi-feature fusion; Internet of things; Energy efficient routing.
Energy-Saving Algorithm for Data Center Network Based on Genetic Algorithm
by Shu Yang, Hua Yang, Hua Yang, Wen Chai, Wen Chai, Zehui Liu, Zehui Liu
Abstract: This thesis focuses on network equipments in the data center which hasrncaused rapidly growth of energy consumption recent years. The switches account for the largest proportion of energy consumption of network equipments, so turning off unneeded switches can reduce energy consumption effectively. Based on this point, we develop an high-efficient routing algorithm based on genetic algorithm(GA) in order to improve energy consumption of network equipments. Genetic algorithm is a kind of a heuristic algorithm which solves the optimization problem rapidly by imitating the way of the natural selection, but to a certain degree, it reduces accuracy. Its a complicated problem to decide routing path in arnshort period, so we choose genetic algorithm to achieve our goals. In ourrnsimulation, we make some improvements of GA in order to fit our problem andrnraise the accuracy of its solution.
Keywords: Data center network; Energy efficient routing; Genetic algorithm.
Damage Prevention Analysis of Heavy-Duty Gear Body Based on Finite Element Neural Network
by Pei Weichi, Dong Jianwei, Long Haiyang, Ji Hongchao, Zhang Wenming, Li Yaogang
Abstract: The method of damage prevention analysis of heavy-duty gear body based on finite element neural network is proposed to improve the effectiveness of damage prevention analysis of heavy-duty gear body. Firstly, a design platform for gearbox gears of caterpillar tractors is developed based on finite element theory, the three-dimensional model of the gear is designed on this platform, and the bending and contact finite element analysis of the gear teeth is carried out, the bending stress and contact stress of the gears are obtained, which provides a basis for the parameter design and reliability of the gears. Secondly, a neural network algorithm is introduced to predict and analyze the impact of damage data of heavy-duty gear body. Finally, the effectiveness of the proposed algorithm is verified by simulation experiments.
Keywords: Finite element; Neural network; Heavy-duty gear body; Destruction prevention.
On-Demand Distance Vector Refinement-Based Car Networking Stable Routing
by Shutao Zhou, Chengxing Li, Hui Yu
Abstract: Due to the high-speed movement of vehicles and obstacles in urban scenes, the communication paths between vehicles become extremely fragile. Specific to the routing problem of VANETs, a path criterion-based on-demand distance vector PA-AODV routing algorithm is proposed. The PA-AODV routing algorithm makes full use of the characteristics of AODV routing, and improves its routing decision. By calculating the path criterion weight and preferentially selecting the path with small weight for data transmission, the path stability is thus improved. The experimental data show that the proposed PA-AODV routing reduces the data packets loss rate and also shortens the end-to-end transmission delay.
Keywords: Car networking; Distance vector refinement; On-demand distance vector; Communication path; Link quality.
Cognitive and Artificial Intelligence System for Logistics Industry
by Jing Zhao, Fengjie XIE
Abstract: With the continuous development of cognitive science, the impact on society is becoming more and more significant.Artificial intelligence is an important branch of cognitive science.Artificial intelligence has been applied to medical, education, security, logistics and other industries, which has broad prospects for development. Logistics industry uses artificial intelligence technology to complete intelligent search, face recognition, combined with large data calculation and planning reasonable path in warehousing, which plays an important role in the process of storage, transportation and distribution.Taking China's logistics industry as the research object, this paper analyzes the application of artificial intelligence technology in the logistics industry. In the warehousing process, artificial intelligence technologies including compile storage code, automatic picking with Automated Guided Vehicle, warehouse robot to improve work efficiency.Intelligent unmanned aerial vehicle (UAV) transport and intelligent sorting technology are implemented by artificial intelligence technology in the transportation link.Logistics distribution links use artificial intelligence technology to plan the best path, improve the recognition rate of express waybill that save a lot of labor.Artificial intelligence technology allocates logistics resources, optimizes logistics links, and improves logistics efficiency and other measures to promote the development of logistics informatization and automation.rn
Keywords: Cognitive Technology;Artificial Intelligence (AI); Logistics Industry.
Study on oceanic big data clustering based on incremental K-means algorithm
by Yongyi Li, Zhongqiang Yang, Kaixu Han
Abstract: With the increase of marine industry in the Beibu Gulf, data clustering has become an important task of intelligent ocean. Partition clustering methods are suitable for marine data. However, traditional K-means algorithm is not suitable for large scale data. Focusing on the characteristics of oceanic big data, we propose a clustering method based on incremental K-means (IKM) algorithm. First, a vector model is adopted to represent data sets, and the calculation model for mean values and centers is used to initialize arbitrary numbers of data points. Second, the input data vectors are iteratively calculated in an incremental vector form. Finally, by applying incremental vector and distance model, the large-scale data are clustered according to convergence condition. Experiments show that the algorithm can increase the clustering efficiency, reduce time and space complexity, and lower the missing data rate.
Keywords: cluster; K-means; incremental; oceanic big.
A fuzzy comprehensive evaluation model for Smart City Application
by Huaihui Liu, Zhiqing Zhang, Zhijie Sun
Abstract: As one of the basic social relationships in current world, the relationship between the police and citizens directly reflects the relationship between the authority and the public, which play an important role in the social stability. It has an essential significance to properly get to know, to deal with and to evaluate the police-citizen relationship. Firstly, we design a hierarchy evaluation index system model about the harmony degree between the police and the citizens, with the help of questionnaires, based on the principle of designing an evaluation index system and the five major factors that impact the harmony relationship between the police and the citizens. Secondly, we set up a fuzzy comprehensive evaluation model based on an improved analysis hierarchy process (AHP). And then we make an empirical research on the harmony degree between the police and the citizens with the help of the model we set up. Finally, based on the conclusion of the empirical research, we make a propose to the government and the security department about how to promote the construction of the harmony police-citizen relationship. The research enriches the methods and means of evaluation the harmony degree of the police and the citizens and exemplifies the empirical research.
Keywords: the harmony degree of the police and citizens; evaluation index system; the improved AHP; the fuzzy comprehensive evaluation mathematical model.
A Chameleon Hash Authentication Tree Optimization Audit for Data Storage Security in Cloud Calculation
by Yang Bo
Abstract: In order to improve the security of data storage in cloud calculation , a chameleon Hash authentication tree optimization audit method for data storage security in cloud calculation is proposed. First, an optimized public audit agreement is proposed. By storing homomorphic linear validator for user data on TPA sites, the response size of cloud storage server (CSS) is optimized. At the same time, the quasi-random function is used to optimize the query request to CSS; secondly, the chameleon hash and an improved chameleon authentication tree are used to perform efficient dynamic data updating on client data (cloud calculation ) to support block-level updating and fine-grained updating; finally, through thorough security and performance analysis, it is clearly verified that the proposed method is safe and efficient.
Keywords: Cloud calculation; Data storage; Chameleon authentication tree; Third part audit; Quasi random function.
Optimization of CoMP based Cellular Network design and its Radio network parameters for Next Generation HetNet using Taguchi
by Sarosh Dastoor, Upena Dalal, Jignesh Sarvaiya
Abstract: A heterogeneous network (HetNet) is a complex network made of variable cellular dimensions with different network topology. An erratic network design is valueless, unproductive and expensive. Research paper describes coordination of multipoint transmission, in which a collection of transmitting Base Stations (BS) dynamically harmonizes their transmission among themselves, enhancing the coverage to the edge users. The proposed cellular planning strategy uses variable radii cells forming a cluster in a given region to be dimensioned. For a given cluster, minimum distance (dmin) between two cells has been calculated and using proposed (1/3 d_min ) dimensioning technique, the coverage radius of cells in a cluster is made, forming a HetNet. By optimizing the network; coverage, cost and energy requirements could be minimized and optimization of network performance parameters like transmission power, tilt and azimuth angle of antenna with the radius of cell provides cost-efficient deployment of a network. The research paper proposes the mathematical dimensioning model for the design of a HetNet as well as its performance parameters using Taguchi
Keywords: heterogeneous network; optimization; orthogonal array; coordination; Multipoint transmission; Taguchiâ€™s Method; azimuth; tilt; energy conservation; throughput.