Research On Range-based Positioning Algorithm For Wireless Sensor Networks

Wireless sensor network is a new multi-hop network which is composed of alarge number of low-cost sensor nodes. Due to its advantages like low cost, robustnessand self organization, WSN has been widely used in national defense and military,disaster relief, intelligent transportation, health care fields. However, in our real life,only when the sensor data attached the location information will that make sense.Therefore, node positioning technology is the key to the wireless sensor network, andlocalization algorithm accuracy problem is a important hotspot in the field of wirelesssensor network (WSN) research.Since the20th century, international and domestic academics have proposed aseries of node localization algorithms. Most of these algorithms have used judgment,circulation, weighting and other method, these methods improved ranging accuracy orpositioning accuracy by TOA values, NLOS environment discrimination, nodescredibility and other fields. So far, there has not any positioning algorithm which usesthe relationship between ranging error and positioning error or exploits the positioningerror model to improve algorithm accuracy. Meanwhile most of those currentalgorithms exist low positioning accuracy, high computational complexity and powerconsumption issues in application.For this situation, this paper firstly consideredthree modules included ranging, self-positioning and the optimization of positioningaccuracy, and then used ML algorithm to derive the relationship between rangingerror and positioning error. On the basis of these work, the paper proposed a moreeffective, practical positioning algorithm. The main research work of the paper is asfollows:The paper studied five kinds of classic range-based algorithms, and analyzed thestatus of range-based algorithms and problems. At the same time, the paper derivedthe relationship between distance error and positioning error by using ML method,and proposed a positioning model based on the ranging errors.The paper designed a positioning error correction scheme for three kinds ofenvironment which include DDP (Dominant Direct-Path), NDDP (Non-Dominantdirect path) and UDP (Undetectable direct path) according to the positioning errormodel based on ranging errors and the main existing models of Ultra widebandranging errors.This paper used MATLAB to test and verify the performance of the improved algorithm. Simulation results show that, in the100m*100m square monitoring areacompared to conventional algorithms, the proposed algorithm can minimized100cmpositioning error in NDDP environment. Its average accuracy has increased13.5%,and the mean of the positioning error was less than0.25m. In the UDP environment,the algorithm can reduced10m positioning error. Its average accuracy can increases9.51%, and its average positioning error is less than0.6m. All of these resultsindicated that the performance of improved algorithm is superior to the conventionalalgorithm. It has certain superiority in stability and low power consumption comparedwith the similar advanced algorithm.