面向多无人机自主协同探测系统的联合估计与决策研究

Based on the external emitters, the ability of the multiple Unmanned aerial vehicles to autonomously and cooperatively detect the stealthy target can be greatly improved. However, due to the uncertainties present in disturbance, maneuvering, data association, and propagation path identification, the multi-UAVs autonomous and cooperative detection system faces the new challenges of coupling between estimation and decision, uncertainty and high-dimension. This project aims at the close-loop feedback processing among state estimation, parameter identification and path recognition, and focuses on joint estimation and decision. Construct the new framework to associate and fuse the measurements from the auxiliary sensors and the main sensor, and propose the joint estimation and identification algorithm based on the constrained filters. Establish the joint estimation and identification scheme based on the generalized Bayes risk, and seek the fast solution for the multi-objective, multi-restriction and high-dimension optimization problem. Meanwhile, we will determine the on-line performance evaluation strategy for optimizing the thresholding parameters in iterative framework, and seek a tradeoff between computation burden and processing accuracy. The project will be verified via simulation and real data and contributed to promote the integration of estimation and decision, improve the methods of joint estimation and decision, and further develop the applications of the theory of information fusion.

多无人机利用外部辐射源进行自主协同探测，可以大幅度提高对隐身目标的探测能力，但干扰参数辨识、目标机动辨识、量测关联和传播路径关联等多种不确定因素耦合下的多无人机自主协同探测系统面临着估计与决策深度耦合、数据不确定、数据高维数等全新问题。本项目拟搭建目标状态估计与参数辨识、模式识别闭环反馈环节，开展联合估计与决策研究。建立辅助信源与主传感器量测系统之间的关联融合框架，研究基于辅助信源约束滤波的联合估计与决策算法；构建基于广义贝叶斯风险的联合估计与决策的性能评估框架，探索多目标多约束高维优化问题的高精度快速求解；研究在线评估下的迭代阈值参数优化策略，实现计算量与处理精度的动态折衷，开展基于仿真测试与实际数据相结合的新机理新方法原理性验证，促进估计与决策等多研究方向的交叉融合，有望完善联合估计与决策方法体系，进一步拓展信息融合方法的应用范围。

本项目围绕多无人机自主协同探测系统面临着目标机动水平、干扰参数未知，以及传统量测数据关联、传播路径关联未知下的目标状态估计问题，构造了辅助信源与主传感器量测系统之间的关联融合框架，完成了基于辅助信源的联合估计与决策算法研究，引入了广义贝叶斯风险作为性能评价指标，提出了带约束条件的广义贝叶斯联合估计与决策风险在线优化算法，并着重研究了基于群体智能仿生优化策略的多目标多约束优化问题的高精度快速实现方法。最后以无人机为应用平台，实现了基于辅助信源的高精度定位以及自主导航功能。相关研究成果整理后出版专著两部，发表SCI期刊论文三篇，EI论文五篇。所得研究成果已经应用于旋翼无人飞行器、无人车等，并可进一步应用于空天地一体化多目标、多任务协同控制系统。