改进的辅助粒子多Agent系统信息融合滤波算法

为克服多Agent系统的非线性问题,提高多Agent系统对信息的协同合作能力,提出一种改进的辅助粒子信息融合滤波算法。该算法通过分析多Agent系统的结构,根据正交小波多尺度分析理论对系统内Agent的信息进行分解、重构,给出以Agent信息为重要采样密度函数的辅助粒子滤波算法,并对滤波结果进行特征融合,得到多Agent系统融合特征。将此算法应用到机动目标跟踪领域,并与传统的滤波算法进行对比,仿真结果验证了该算法的有效性。

An Improved Auxiliary Particle Filter Algorithm for Information Fusion in Multi-Agent System

In order to overcome the nonlinear problem and improve the collaborative capabilities in the multi-agent system,this paper proposes an improved auxiliary particle information fusion filter algorithm.The architecture of multi-agent system is analyzed,then the signal of agent is decomposed and reconstructed based on orthogonal wavelet multi-scale analysis theory,so the improved auxiliary particle information fusion filter algorithm is proposed based on the signal of agent,which can be taken as the important sampling density function,after that,the features are integrated into multi-agent system’s features.Finally,this algorithm is applied to the maneuvering target tracking through computer simulation and compared with conventional algorithms,and simulation results verify its effectiveness.