Research On Fault Tolerant Tracking Methods Of Mobile Target Based On Binary Sensor Networks

Mobile target tracking is a main application field in Wireless Sensor Networks(WSNs). Binary Sensor Networks(BSNs), which significantly reduce the complexity of target tracking algorithm and power consumption as the result of using sensor note sensing target signal to obtain target‘s information of location, gets more attention in the research and industrial fields because of its low cost, less band width occupation, power consumption and other features. However, due to the influence of environmental noise, sensing resolution, randomness of the moving target and other uncertain factors, there is no guarantee for tracking accuracy in BSNs. As a result, error tolerance become the critical factors for BSNs mobile target tracking system‘s practicability.Due to taking methods for processing uncertainty by calibrating or filtrating the low ranging-quality sensing result, traditional target tracking method is vulnerable in keeping sensing information and saving powers, which makes it inflexible and unpractical. In fact, with the cooperation of sensor nodes, by capturing and using the potential gain of uncertainty for tracking, methods could both tolerate the localization error caused by environment, target mobility, sensing resolution and other factors, and integrate all the sensing information maximally. Methods could improve the error tolerance of target tracking while guaranteeing the complexity of algorithm. Based on the above ideas, for different application field, this paper bring out two different high error-tolerant tracking methods by using uncertainty. The main works are as follows:(1) For the application field of high requirement in real-time tracking and average requirement in tracking accuracy, this paper bring out an Error-Tolerant Fast Target Tracking Method(ETF). By calculating the sampling results, the ETF method makes the uncertain information of sensing results to be a useful vector variable and turns the uncertain information into additional information effectively. In addition, the ETF method could localize the target by minimum manhattan distance sequence matching and obtain the tracking path of target. The ETF method reduces the algorithm complexity and guarantees its error tolerance and tracking accuracy, which accomplishing a fast and accurate target tracking process. Performance analysis and simulation experiment can prove the ETF method‘s effectiveness.(2) For the application field of high requirement in tracking accuracy and error tolerance and average requirement in algorithm complexity and real-time tracking, this paper bring out an Energy-Saving High-Accuracy Target Tracking Method(ESHA). By capturing and comparing the relative degree of sensor node‘s sensing uncertainty and modeling the target-sensor relative localization model quantitatively, the ESHA method could guarantee the uniqueness of mapping between sensing uncertainty and uncertain area, which solve the problem of one-time sensing error. Then, by using the neighbor relation obtained from geometric partitioning, a heuristic algorithm has been designed to reduce the algorithm complexity, which cut down sensor nodes‘ power consumption and guarantee the robustness and accuracy of algorithm. Simulation experiments prove the high tracking accuracy of the EHSA method and keep the system‘s power saved, which makes it better in error tolerance and robustness.(3) A prototype of mobile target error-tolerant tracking system is designed and applied. Based on the real environment experiment, the system demonstrates the influence of uncertainty and proves the performance of two methods in this paper. Experiments‘ results reveal that both two methods in different application field could solve the problem of sensing uncertainty, which laid the foundation for the practical application.