A Multi-agent Realization Method And Experimental Research Of WSN Target Consistency

Target consistency tracking has a very broad application prospect in both military and civil fields.The rapid development of WSN and AI provides a new train of thought for target consistency tracking.In this paper,nodes of WSN are regarded as the agents of AI.Based on the cooperation of agents,the algorithm of multi-agent target consistency tracking is studied and both software and hardware simulations of the algorithm are carried out.The basic concept and principle of target tracking algorithm and consistency control algorithm are firstly illustrated in this paper.Several information fusion algorithms of WSN are introduced secondly.The process of the Kalman filtering algorithm is very detailed analyzed and the reason why it is not suitable for maneuvering target tracking is explained.An adaptive function is constructed based on the absolute value of the target heading angle difference,and the prediction of the Kalman filtering is corrected using this adaptive function.In order to verify the feasibility of the proposed algorithm,a platform composed of NetLogo and MATLAB is designed.Based on the convenience of NetLogo's agents management and the strong calculative ability of Matlab,a simulation is carried out and the feasibility is confirmed.An ultrasonic ranging car carrying STM32 MCU and CC2420 RF module is designed for hardware simulation.A multi-agent system consisted of three such cars is designed and a target is tracked with the cooperation of beacon nodes.