A Kind of Wireless Sensor Network Coverage Optimization Algorithm Based on Genetic PSO

Similar documents
A Novel Image Super-resolution Reconstruction Algorithm based on Modified Sparse Representation

Study on GA-based matching method of railway vehicle wheels

Open Access Research on the Prediction Model of Material Cost Based on Data Mining

Immune Optimization Design of Diesel Engine Valve Spring Based on the Artificial Fish Swarm

Aero-engine PID parameters Optimization based on Adaptive Genetic Algorithm. Yinling Wang, Huacong Li

A Data Classification Algorithm of Internet of Things Based on Neural Network

IMPROVING THE PARTICLE SWARM OPTIMIZATION ALGORITHM USING THE SIMPLEX METHOD AT LATE STAGE

IMPROVED ARTIFICIAL FISH SWARM ALGORITHM AND ITS APPLICATION IN OPTIMAL DESIGN OF TRUSS STRUCTURE

Parameter Selection of a Support Vector Machine, Based on a Chaotic Particle Swarm Optimization Algorithm

Application of Theory and Technology of Wireless Sensor Network System for Soil Environmental Monitoring

Test Analysis of Serial Communication Extension in Mobile Nodes of Participatory Sensing System Xinqiang Tang 1, Huichun Peng 2

QUANTUM BASED PSO TECHNIQUE FOR IMAGE SEGMENTATION

Robot Path Planning Method Based on Improved Genetic Algorithm

Structural topology optimization based on improved genetic algorithm

Research on Intrusion Detection Algorithm Based on Multi-Class SVM in Wireless Sensor Networks

Research on Applications of Data Mining in Electronic Commerce. Xiuping YANG 1, a

ZigBee Routing Algorithm Based on Energy Optimization

Performance Comparison and Analysis of Power Quality Web Services Based on REST and SOAP

Research Article Path Planning Using a Hybrid Evolutionary Algorithm Based on Tree Structure Encoding

Application of Wang-Yu Algorithm in the Geometric Constraint Problem

Network Scheduling Model of Cloud Computing

A Two-phase Distributed Training Algorithm for Linear SVM in WSN

A New WSNs Localization Based on Improved Fruit Flies Optimization Algorithm. Haiyun Wang

PSO-based Energy-balanced Double Cluster-heads Clustering Routing for wireless sensor networks

Hole repair algorithm in hybrid sensor networks

Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis

Meta- Heuristic based Optimization Algorithms: A Comparative Study of Genetic Algorithm and Particle Swarm Optimization

A liquid level control system based on LabVIEW and MATLAB hybrid programming

A Genetic Algorithm-Based Approach for Energy- Efficient Clustering of Wireless Sensor Networks

CHAPTER 6 HYBRID AI BASED IMAGE CLASSIFICATION TECHNIQUES

Research on time optimal trajectory planning of 7-DOF manipulator based on genetic algorithm

A new improved ant colony algorithm with levy mutation 1

An Application of Genetic Algorithm for Auto-body Panel Die-design Case Library Based on Grid

Traffic Signal Control Based On Fuzzy Artificial Neural Networks With Particle Swarm Optimization

Fault Diagnosis of Wind Turbine Based on ELMD and FCM

Open Access Research on Algorithms of Spatial-Temporal Multi-Channel Allocation Based on the Greedy Algorithm for Wireless Mesh Network

Energy Aware Node Placement Algorithm for Wireless Sensor Network

The movement of the dimmer firefly i towards the brighter firefly j in terms of the dimmer one s updated location is determined by the following equat

Identification of Vehicle Class and Speed for Mixed Sensor Technology using Fuzzy- Neural & Genetic Algorithm : A Design Approach

Design of student information system based on association algorithm and data mining technology. CaiYan, ChenHua

Sample Adaptive Offset Optimization in HEVC

ANN-Based Modeling for Load and Main Steam Pressure Characteristics of a 600MW Supercritical Power Generating Unit

A Network Intrusion Detection System Architecture Based on Snort and. Computational Intelligence

Cluster Validity Classification Approaches Based on Geometric Probability and Application in the Classification of Remotely Sensed Images

A Study and Analysis on a Perceptual Image Hash Algorithm Based on Invariant Moments

Structural Optimizations of a 12/8 Switched Reluctance Motor using a Genetic Algorithm

Design and Research of Adaptive Filter Based on LabVIEW

Video Inter-frame Forgery Identification Based on Optical Flow Consistency

An Improved KNN Classification Algorithm based on Sampling

Analysis Range-Free Node Location Algorithm in WSN

View-dependent fast real-time generating algorithm for large-scale terrain

The Improved LEACH-C Protocol with the Cuckoo Search Algorithm

Tact Optimization Algorithm of LED Bulb Production Line Based on CEPSO

K-coverage prediction optimization for non-uniform motion objects in wireless video sensor networks

Kyrre Glette INF3490 Evolvable Hardware Cartesian Genetic Programming

A PSO-based Generic Classifier Design and Weka Implementation Study

An Energy Efficiency Routing Algorithm of Wireless Sensor Network Based on Round Model. Zhang Ying-Hui

A Balancing Algorithm in Wireless Sensor Network Based on the Assistance of Approaching Nodes

Application of Geometry Rectification to Deformed Characters Recognition Liqun Wang1, a * and Honghui Fan2

Nodes Energy Conserving Algorithms to prevent Partitioning in Wireless Sensor Networks

The Application Research of Neural Network in Embedded Intelligent Detection

Research on the Application of Digital Images Based on the Computer Graphics. Jing Li 1, Bin Hu 2

Power Load Forecasting Based on ABC-SA Neural Network Model

AUV Cruise Path Planning Based on Energy Priority and Current Model

Sensor Modeling Realization Based on Fruit Fly Optimization Algorithm

Key words: Wireless Sensor Networks, Clustering Routing Algorithm, Hierarchical Multi-hop Clustering Routing Algorithm

Optimization of Cutting Parameters for Milling Operation using Genetic Algorithm technique through MATLAB

Research on Heterogeneous Network Integration in Distribution Communication Network

Efficient Path Finding Method Based Evaluation Function in Large Scene Online Games and Its Application

Low Overhead Geometric On-demand Routing Protocol for Mobile Ad Hoc Networks

Research on the New Image De-Noising Methodology Based on Neural Network and HMM-Hidden Markov Models

AN OPTIMIZATION GENETIC ALGORITHM FOR IMAGE DATABASES IN AGRICULTURE

Using The Heuristic Genetic Algorithm in Multi-runway Aircraft Landing Scheduling

SIMULATION APPROACH OF CUTTING TOOL MOVEMENT USING ARTIFICIAL INTELLIGENCE METHOD

Target Tracking in Wireless Sensor Network

Handling Multi Objectives of with Multi Objective Dynamic Particle Swarm Optimization

Research on Heterogeneous Communication Network for Power Distribution Automation

Hybrid of Ant Colony Optimization and Gravitational Emulation Based Load Balancing Strategy in Cloud Computing

Comprehensive analysis and evaluation of big data for main transformer equipment based on PCA and Apriority

Research on Design and Application of Computer Database Quality Evaluation Model

Reconfiguration Optimization for Loss Reduction in Distribution Networks using Hybrid PSO algorithm and Fuzzy logic

High Resolution Remote Sensing Image Classification based on SVM and FCM Qin LI a, Wenxing BAO b, Xing LI c, Bin LI d

An Abnormal Data Detection Method Based on the Temporal-spatial Correlation in Wireless Sensor Networks

DDoS Detection in SDN Switches using Support Vector Machine Classifier

CHAOTIC ANT SYSTEM OPTIMIZATION FOR PATH PLANNING OF THE MOBILE ROBOTS

Image Compression: An Artificial Neural Network Approach

Prediction of traffic flow based on the EMD and wavelet neural network Teng Feng 1,a,Xiaohong Wang 1,b,Yunlai He 1,c

Energy Optimized Routing Algorithm in Multi-sink Wireless Sensor Networks

Intelligent management of on-line video learning resources supported by Web-mining technology based on the practical application of VOD

A Kind of Fast Image Edge Detection Algorithm Based on Dynamic Threshold Value

Qiqihar University, China *Corresponding author. Keywords: Highway tunnel, Variant monitoring, Circle fit, Digital speckle.

Research on-board LIDAR point cloud data pretreatment

Scheduling of Independent Tasks in Cloud Computing Using Modified Genetic Algorithm (FUZZY LOGIC)

Searching Algorithm of Dormant Node in Wireless Sensor Networks

A target allocation of infrared multi-sensor based on distributed niche genetic algorithm 1

An Approach to Polygonal Approximation of Digital CurvesBasedonDiscreteParticleSwarmAlgorithm

Study on Digitized Measuring Technique of Thrust Line for Rocket Nozzle

Discussion of GPON technology application in communication engineering Zhongbo Feng

Design of the Power Online Monitoring System Based on LabVIEW

Face recognition based on improved BP neural network

Transcription:

Sensors & Transducers 2013 by IFSA http://www.sensorsportal.com A Kind of Wireless Sensor Network Coverage Optimization Algorithm Based on Genetic PSO Yinghui HUANG School of Electronics and Information, Nantong University, Nantong Jiangsu, 226019, China E-mail: hyh9688@ntu.edu.cn Received: 30 June 2013 /Accepted: 25 October 2013 /Published: 30 November 2013 Abstract: Aiming to the problems of slow convergence speed and being easily early-maturing etc. in existing network based on standard particle swarm algorithm, this paper proposes a wireless sensor network coverage optimization algorithm based on the genetic particle swarm optimization (PSO). The wireless sensor maximum coverage is regarded as the objective function, the genetic algorithm with adaptive crossover and mutation factors is used to search the solution space, and the powerful global search ability of PSO is used to increase search scope to make particle covering more efficient, which both strengthen algorithm optimization ability, improve the nodes coverage, and solve early-mature problem. Comparing with the standard traditional genetic algorithm and new quantum genetic algorithm, simulation results present that the rate of coverage in this algorithm increases by 2.28 % and 0.65 % respectively; and convergence speed is also improved, therefore this method can effectively realize the wireless sensor network coverage optimization. Copyright 2013 IFSA. Keywords: Wireless sensor network, Coverage optimization, POS, Genetic algorithm. 1. Introduction Wireless sensor network is composed of lots of sensor nodes placed in the detection areas, which adopts wireless communication mode to do data communication. In the network all kinds of sensor nodes can collaboratively sensing, collect and analyze the information of related objects in the network coverage area. Covering problem is the key problem of wireless sensor network, the distribution area of sensor nodes determines the effectiveness of network detection. High effective network coverage can ensure the space source of wireless sensor network to get the reasonable distribution [1-2]. At present, scholars have developed a variety of algorithms to optimize sensor network coverage, including distributed algorithm, approximate minimum node set algorithm, set covering algorithm, genetic algorithm, and standard particle swarm algorithm etc. [3-4]. These algorithms all have corresponding defects, for example, the distributed algorithm has a longer network cycle, but it can only solve the problem of unit circle, it can not solve the problem of irregular geometric area, there is certain limitation. Approximate minimum node set algorithm fails to fully consider node distribution, which can not ensure to realize the optimal sensor network coverage. Set covering algorithm operates the coverage through the relationship around central node and surrounding node, there is the unbalanced problem of node load distribution. Genetic algorithm finds the optimal solution through the parallel search method, but its convergence speed is slow, which cannot collect the extract dynamic node on time. Standard particle swarm algorithm searches optimal solution in advance in the detection area, which Article number P_1467 107

results the particle search area reduced and node network coverage dropped [5-6]. In order to solve the problems existing in the traditional algorithms, this paper presents a wireless sensor network coverage optimization algorithm based on genetic PSO. The wireless sensor maximum coverage is regarded as the objective function, the genetic algorithm with adaptive crossover and mutation factors is used to search the solution space, and the powerful global search ability of PSO is used to increase search scope to make particle coverage more efficient, which both strengthen algorithm optimization ability, improve the nodes coverage rate, and solve early-mature problem. Connected coverage refers to the neighbor nodes of node Z arrange according to the special method to ensure being inducted by sensor nodes [10]. 2. The Principle of Wireless Sensor Network Coverage 2.1. Construction of Wireless Sensor Network To get a higher sensor nodes coverage rate in WSN, to reduce the number of blind spot of the perception, the density of sensor nodes deployment needs to increase. However, if sensor nodes deployed density is too large, which will produce a large number of redundant nodes to lead to data transmission conflict, the energy will be work out too early, so as to the WSN lifecycle becomes too short. For a sensor node, if its sensing area can be completely covered by the other sensor nodes, and then the node is called redundant nodes, under the status of other nodes is working, the sensor node can enter a dormant state. The redundant nodes are judged through the corresponding algorithm to make the work nodes minimum as far as possible, in order to improve the network coverage. Therefore, in wireless sensor network initialization phase, the choice of work nodes number and node energy consumption need to be considered at the same time, it is a contradiction between WSN coverage and the number of working nodes [7] Wireless sensor nodes have their own specific sensing radius, the detection area within the sensing radius is known as coverage area, the node area outside sensor radius is called blind area. The goal of constructing wireless sensor network is: reasonable allocation of network resources, maximizing the network coverage [8-9]. The physical model of wireless sensor network node coverage is described in Fig. 1. As shown in Fig. 1, node coverage area in wireless sensor network is all the information collection within sensor area that takes node Z as the center of circle and R z as the radius. The distance between any two nodes R E and R F in communication coverage area should be kept in the diffuse area of R H, to ensure the sensing information from two nodes can effectively communicate. Fig. 1. Wireless sensor network node coverage model. 2.2. The Selection of Coverage Optimization Goal The ultimate goal of wireless sensor network (WSN) is to maximize network coverage. In case to ensure the successful completion of network communication, how to place sensor nodes and using the minimum network nodes to realize maximizing network coverage are the main purposes. Two indicators of sensor node coverage and area coverage are used to judge the coverage of wireless sensor network. 2.2.1. Node Coverage The waiting test area is represented in W, after it is processed by digital discretization, it includes p* q pixels. In W there are q sensor nodes with the same parameters, and the coordinates of each sensor node are known as { xi, y i}, induction radius is r, the communication radius R = 2r. Wi = { xi, yi, r} represents the circle with a center of circle of sensor node coordinate { xi, y i} and the radius r. The pixel points are set as ( x, y ), the distance between target pixels and sensor node is 2 2 expressed with hw (, k) = ( x x) + ( y y), i i i Kr {} i represents the coverage probability of pixels ( x, y ) covered by sensor node W i. Sensor node detection model can present a certain probability distribution under the influence of the environment, which is as follows: 108

Kwpg ( x, y, wi ) = 1 if ( d( ωi, k) r ru αθβ 1 1 1 u( ) + α2 if r- ru < d( ωi, k) < r+ ru θβ 2 2 0 otherwise (1) In the expression, ru(0 < ru < r) represents the effectiveness parameter tested by sensor nodes, α1, α2, β1, β 2 respectively represent the testing parameters related with sensor node characteristics, and θ1, θ 2 represents the input parameters, and there are: θ1 = ru r+ d( ωi, k) (2) θ2 = ru r+ d( ωi, k) (3) Multiple sensor nodes synergistically measure target, which can improve the probability of target measurement, the probability of synergistical measurement is as follows: K ( W ) = 1 (1 K ( x, y, ω )) ω pg pg wpg i ωi ωpg 2.2.2. Area Coverage Rate Monitoring area W contains p* q pixels, a node set comprehensive detection probability K ( W ) can analyze if each pixel is overwritten, wpg pg and the area coverage of node set W is set as follows: Q zone Kwpg ( Wpg ) ( w) = (5) p q 3. Genetic PSO Optimization Nodes Coverage Algorithm 3.1. Improvement of Traditional Genetic Algorithm Genetic operation contains selection, cross-over, and mutation three kinds of operators, there into, crossover and mutation operators play an important role to genetic algorithm. Cross can extract the effective structure of two parents to construct a new individual; mutation can adjust some individual gene value in the set. Integration of the crossover and mutation operators can enhance the search ability of genetic algorithm, the crossover operator is taken as (4) the main operator, and mutation operator is taken as a secondary operator. The traditional genetic algorithm mainly does the iteration calculation according to the fixed crossover and mutation probability, when the genetic operator is small, it will cause algorithm convergence speed too fast or too slow; when genetic operator is too big, it will lead to the algorithm can not converge. In order to solve the disadvantages of traditional algorithm, the genetic algorithm with an adaptive operator makes the individual can adaptively adjust crossover and mutation probability. Decreasing the crossover probability of parent individual with high similarity, and increasing the crossover probability of parent individuals with low similarity, can enhance cross processing accuracy to complete the global search; at the same time, the individuals with high fitness use low crossover probability, the individuals with low fitness use high crossover probability, which can ensure the diversity of set and increase the convergence speed. The operation method of adaptive crossover and mutation operator is: w = ijq., 1 wijq,, /ln( u + + zij.) (6) P =, 1 (1 iq+ ln giq, ) * piq, (7) Among them, the crossover probability of the i - th and the j -th parent in the q -th generation is described as w i, j, q, the degree of approximation of the i -th and the j -th parent is described with z i, j, the mutation probability of the i -th individual in the q -th generation is described with p iq,, the fitness of the i -th individual in the q -th generation is described with g iq,. 3.2. PSO Genetic Optimization Algorithm In the process of node coverage, genetic algorithm has certain disadvantages, in order to overcome these disadvantages, PSO genetic algorithm is introduced. This algorithm combines the PSO particle swarm global search capability and the local optimization function of genetic algorithm, has strong nonlinear mapping ability. PSO algorithm sets up a variety of particles in the space dimensions, the best particle coordinates is the optimal solution of objective function. Based on the current coordinate values, particles constantly adjust their speed and coordinate. In this paper, the spatial dimension is composed of node coverage and area coverage, including two groups of the weight coefficient and two dimensional adaptive values, the weight coefficient of each group is integrated into the model of genetic algorithm and does the readjustment and optimization. In this paper, the algorithm sets up 109

60 search particles, the solution of searching optimal objective function is the spatial dimension coordinates of each particle. If the position of i -th particle is is X ( x, x,..., x ) = (8) i i1 i2 i60 The speed of particles searching optimal in space U u u u i = ( i1, i2,..., i60) (9) The optimal position of particle swarm itself is Ki = ( Ki 1, Ki2,..., Ki60), the optimal position of whole particles swarm is Li = ( Li 1, Li2,..., Li60). The particle swarm using expression (10) to adjust the position: U = wu + d r( ke x ) j+ 1 j j j id id 11 id id j j w22( r fbd xid) j j+ 1 j j+! id = id + id + (10) X X U In the expressions, the first half is used to adjust the speed of particles; the second half represents the self learning process of particles. The expression can optimize the cognizing energy of particles to avoid algorithm appearing "early-mature" phenomenon. The particle swarm's self-learning ability can improve particle swarm global optimization ability. Through the above formulas the particles can improve the effectiveness and efficiency of stage coverage optimization. 4.2. Contrastive Analysis on the Performance of This Algorithm and Traditional Algorithm In order to verify the advantages of wireless sensor network coverage optimization algorithm based on genetic PSO proposed in this paper, through the simulation results to compare with the traditional standard particle swarm algorithm (POS) and the method in this paper, the covering performance test data of two algorithms can be described in Table 1. Table 1. The test data of the coverage performance between PSO and this algorithm. PSO This algorithm Radius Iteration Coverage Iteration Coverage /m times rate (100 %) times rate (%) 1.5 1621 32.65 1576 34.35 2 836 55.79 797 56.97 2.5 473 75.97 394 88.77 3 343 86.74 256 94.98 4 295 95.24 227 99.95 5 234 100 213 100 According to the experimental test data in Table 1, the coverage rate of this algorithm and traditional algorithm, as well as the rectangular figure of iteration time variation trend along with the change of node sensing radius, are respectively described in Fig. 2 and Fig. 3. 4. Simulation Experiments and Analysis 4.1. Simulation Environment Settings Simulation experiment set 50 wireless sensor nodes in a circular measurement range with radius 10 M, the sensing area is the circular area with r = 4m radius, and the communication coverage area of sensor node is the circular area with the radius U = 2r = 8m [9]. The parameters of sensor node detection model in the experiment are u 1 = u 2 = 2, model checking validity parameters are ru = 0.5u = 2m, different nodes feature parameters respectively are α1 = 1, α2 = 0, β1 = 1, β2 = 0.5, number max = 50. The experiment uses the computer with main frequency 2.6 GHz under the condition of Matlab to simulation test about wireless sensor network coverage problem. Fig. 2. The rectangle diagram of coverage changing with nodes sensing radius. From diagram 2, coverage and sensing radius have the positive correlation, the greater the radius is, the greater the coverage is, and the coverage rate in this algorithm is higher than the traditional genetic algorithm. When the sensing radius is lesser, the coverage rectangular figure shows notable rising trend; when the sensing radius increases, the 110

variation trend of coverage rate rectangle diagram is relatively stable, when the sensing radius reaches a certain value to achieve overall coverage, the coverage rate is 100 %. From the analysis of Table 2, the coverage in this algorithm is 92.26 %, which is higher than other two kinds of algorithms 2.28 % and 0.65 %, and the iteration time of this algorithm is less than other two algorithms. Therefore this algorithm can implement the maximum network coverage through less operation time, which has high network coverage optimization performance. 5. Conclusion Fig. 3. The rectangle diagram of iteration times changing with nodes sensing radius. From the analysis of diagram 3, the iteration number and the sensing radius have the negative correlation, when the sensing radius is larger, the iteration time is smaller, and the iteration number in this algorithm is less than the traditional genetic algorithm. When the sensing radius is small, the variation trend of iteration time rectangle diagram is obvious, the convergence speed of two algorithms has great improvement; when sensing radius increases, the iteration time change is relatively smooth, and the algorithm convergence speed gradually decreases. Finally, through Fig. 2 and Fig. 3, the algorithm in this paper has high coverage than traditional standard particle swarm algorithm, the number of iteration is low, which illustrates the optimal performance of this algorithm is stronger, and the coverage performance is better. 4.3. Performance Comparison with Other Wireless Sensors Coverage Algorithm In order to further verify the superiority of this algorithm, this paper respectively adopts this algorithm, traditional genetic algorithm (CGA), and a new quantum genetic algorithm (NQGA) to do coverage test for sensor network, the test results are shown in Table 2. Table 2. The coverage performance comparison of different algorithms. This CGA NQGA algorithm Coverage rate % 92.26 89.19 91.13 Iteration number 650 700 700 Wireless sensor network coverage optimization is in favor of improving network performance and network efficient coverage. This paper presents a wireless sensor network coverage optimization algorithm based on genetic PSO. Wireless sensor maximum coverage is regarded as the objective function, genetic algorithm with adaptive crossover mutation factor is used to search the solution space, PSO powerful global search ability of particle swarm is used to increase search scope, all of these make particle covering more efficient, which strengthen the optimization ability of the algorithm, improve the node coverage, and solve the prematurity problem. Comparing with the traditional genetic algorithm and a new quantum genetic algorithm, there are significant improvements on the effective coverage rate and convergence speed, this algorithm has certain superiority. Reference [1]. Yingchi Mao, The research on wireless sensor network coverage control technology, Journal of Computer Science, Vol. 11, Issue 3, 2007, pp. 20-22. [2]. Xueqing Wang, The research of coverage problems based on grid in wireless sensor networks, Journal of Computer Science, Vol. 33, Issue 11, 2011, pp. 38-39. [3]. Bin Qu, Fangyu Hu, Research on energy-efficient routing protocol of wireless sensor network, Computer Simulation, Vol. 25, Issue 5, 2008, pp. 113-116. [4]. Yonghua Zhou, Peng Li, Zongyuan Mao, A new hybrid method and its application in constrained optimization, Computer Engineering and Applications, 2006, pp. 48-50. [5]. J. C. Platt, Using analytic QP and sparseness to speed training of support vector machines, Advances in Neural Information Processing Systems II, Cambridge, MA: MIT Press, 1999, pp. 557-536. [6]. Xue Wang, Sheng Wang, Junjie Ma, The parallel particle swarm optimization strategy of wireless sensor network node location, Journal of Computers, Vol. 30, Issue 4, 2007, pp. 563-568. [7]. Xiangguang Yao, Yongquan Zhou, Yongmei Li, Artificial fish mixed with particle swarm optimization algorithm, Computer Application Research, Vol. 27, Issue 6, 2010, pp. 2084-2086. [8]. U. H.-G. Kre, et al. Pairwise classification and support vector machines, Advances in Kernel Methods-support Vector Learning, Cambrige, MA: MIT Press, 1999, pp. 255-268. 111

Sensors & Transducers, Vol. 158, Issue 11, November 2013, pp. 107-112 [9]. Hua Fu, Shuang Han, The distribution optimization of new quantum genetic algorithm based on wireless sensor network sensing nodes, Journal of Sensing Technology, Vol. 21, Issue 7, 2008, pp. 1259-1263. [10]. Jiantao Xia, Yiming He, Multi-class classification algorithm combining with support vector and error correction code combination, Journal of North Western Polytechnic University, Vol. 21, Issue 4, 2007, pp. 443-446. 2013 Copyright, International Frequency Sensor Association (IFSA). All rights reserved. (http://www.sensorsportal.com) 112