Study Of Indoor Positioning Technology Based On Wireless Sensor Network

In recent years, with the continuous development of wireless sensor network technology and the urgent need of the indoor positioning technology, which based on the wireless sensor network(WSN) has been widely concerned, the WSN not only easy to deploy and expand, but also cost less. Then, the traditional WSN is difficult to guarantee the higher positioning accuracy and efficiency, because the complex indoor environment causes the signal transmit reflection, scattering and influence of shielding.At present, the indoor positioning technology basis of the wireless sensor network has been gradually moving towards to the practical application from the academic research. But there are still many difficulties need to overcome, including following problems which how to solve positioning excessively rely on the density of beacon nodes, how to use ranging technology effectively, how to optimize nonlinear computational problems in the localization algorithm, and how to establish a system of positioning accuracy, complexity and positioning stability. In view of the problems above, the research content of this paper mainly includes the following aspects:(1) Aiming at study of range error problems of least squares algorithm and analysis of the defect of the particle swarm optimization(PSO) algorithm in wireless sensor network node localization, this paper presents an improved scheme based on particle swarm localization algorithm of self-adaptive difference. Firstly, we fit the unknown node to the distance of anchor nodes by the environmental compensation, and the adaptive value was obtained. Secondly, we used the improved adaptive differential algorithm to generate a new population, then carried on local search with particle swarm algorithm and new mutation strategy, compare with adaptive value by iterative process and convergence gradually. Finally we got the unknown node location. A series of experiments demonstrated that the restraining error accumulation, and positioning accuracy in comparison with other PSO algorithms were improved greatly in indoor positioning.(2) In view of studying the weighted centroid localization algorithm in the optimization problem and analysis of the existing defects of centroid localization algorithm, an improved scheme was proposed based on maximum likelihood estimation of weighted centroid localization algorithm. Firstly, the maximum likelihood estimator between estimated distance and actual distance was calculated as weights. Then, a parameter k was introduced to optimize the weights between anchor nodes and unknown nodes in weight model. Finally, the location of the unknown nodes were calculated and modified by the proposed algorithm. The simulation results show that the weighted centroid algorithm based on the maximum likelihood estimator is better than the algorithm inverse distance based and the algorithm inverse RSSI based. Thus, the algorithm is more suitable for indoor localization of the large areas.(3) In view of studying error accumulation problem in the DV-Hop localization algorithm and analysis of the defects of traditional DV-Hop positioning. Based on the hop and jump distance, a new scheme of DV-Hop localization algorithm was put forward which modified by the genetic algorithm. Firstly, we used the mean of RSSI to modify the hop between anchor nodes, Secondly, the average hop-distance was improved according to the minimum hop between anchor nodes. Finally, the results of position estimation was optimized by genetic algorithm. Contrast to jump distance weighted DV-Hop algorithm and genetic optimization of DV- Hop algorithm, the simulation results showed that the positioning accuracy was improved significantly and the average positioning error was decreased.In conclusion, in this paper we proposed the corresponding improved schemes by learning several algorithms of indoor positioning and its results were verified by simulation based on MATLAB. The experimental results showed that the modified algorithm had some improvements in positioning error and the advantages were obvious when we compared with the same type of algorithms.