Study On Pursuit-evasion Game Based On Wireless Sensor Network

The Pursuit-Evasion Game (PEG) has been studied as a general control problem for many years. In a pursuit-evasion game the classical problem is to find the optimal strategies for pursuit and evasion under a given condition. Along with the appearance of wireless sensor networks (WSN), WSN-based PEG becomes one of hotspots on the research of WSN, WSN control and PEG strategy. In PEG the wireless sensor network is regarded as observer. Compared with other equipments, such as inertial navigation equipment, GPS and camera, the strongpoint of WSN is disposal facility, little influence suffered from outside environment and wide monitoring range, can provide a whole monitoring for the environment and overcome the limited sensing range problem of other monitoring equipments. It is very suitable for the studies on new pursuit-evasion strategies and control problems based on wireless sensor networks. In this study, the theoretical model of PEG platform based on WSN was established. In order to solve the shortcomings of the periodic randomicity of detecting data, time-lapse randomicity and measuring errors of wireless sensor networks, the state estimator of mobile carriers (Evader/Pursuer) based on Kalman filtering technique and smooth mobile technique is designed. By means of simulation and validation, it confirms that the state-estimators can obviously improve the state measurement accuracy of mobile carriers and predict the current state of mobile carriers. In the aspect of control of mobile carriers (Pursuer), the controller of mobile carriers based on fuzzy control is designed. The simulation results show that the controller is robust and can efficiently control the carriers to track targets with considering state errors of mobile carriers. Finally, we investigate the strategies of PEG based on WSN. The evader motion pattern is classified as random motion mode and active evasion motion mode, the Point to Point (PTP) pursuit mode and Intelligent Cooperation Pursuit (ICP) mode of pursuers were designed and the mathematical models of evader and pursuer motions are defined. The simulation results show that the intelligent cooperation pursuit mode is more efficient to chase and capture evaders. While the position error is considered and the pursuit radius is adjusted as r>2.3m, the pursuit-evasion strategy is still valid. Meanwhile, we study the algorithm how to assign targets with pursuers while there are multiple evaders and multiple pursuers. The simulation results show that the target assignment algorithm can efficiently assign targets for pursuers according to the given assignment principle and improve holistic pursuit-evasion efficiency. The pursuit-evasion strategies in this study can reduce computation time and are robust.