基于信息熵的多Agent协同观测轨迹优化方法

Information-entropy-based Optimization Method Research of Collaborative Observation Trajectory for Multiple Agents

针对固装被动传感器的多个移动智能体对一个运动目标协同观测时探测与轨迹变化耦合的现象,提出了一种基于信息熵的协同观测轨迹滚动时域优化(receding horizon optimization,RHO)方法。建立了多Agent协同观测系统的状态方程和基于纯方位信息的目标观测方程,构建了基于无色信息滤波(unscented information filter,UIF)的集中式融合估计算法,引入互信息建立了基于信息熵的目标状态估计性能指标函数,基于滚动时域优化方法,提出了最大化信息熵求解智能体角速度控制量的算法。仿真结果表明,在对运动目标的跟踪过程中,多智能体系统对目标状态估计的误差明显减小,实现了面向最优状态估计的轨迹优化,是一种协调设计多智能体协调观测滤波器和轨迹控制器的手段。

When multiple mobile agents with fixed passive sensors coordinately observe a moving target,the problem of coupling detection and trajectory changes is prone to occur.A receding horizon optimization(RHO)algorithm for collaborative observation trajectory based on information entropy is presented.The state equation of the collaborative observation system of multi-agent and the target observation equation based on bearings-only information are established,and a centralized fusion estimation algorithm based on unscented information filter(UIF)is constructed.Then the mutual information is introduced to establish the performance index function of target state based on information entropy.An algorithm solving the angular velocity control value of the agent by maximizing the information entropy based on the RHO method.The simulation results show that in the process of tracking the moving target,the error of the multi-agent system’s estimation on the target state is significantly reduced,and the trajectory optimization oriented to the optimal state estimation is realized.The proposed method is a coordinated design means of multi-agent coordination observation filter and trajectory controller.