不确定环境下融合语义的无人机编队协同控制研究

UNMANNED AERIAL VEHICLES FORMATION COOPERATIVE CONTROL BASED ON SEMANTIC FUSION IN UNCERTAIN ENVIRONMENT

综合考虑不确定环境下无人机编队控制的影响因素,从信息处理的角度提出一种融合语义的编队协同控制方法。构建具有不确定态势检测、不确定行为识别和语义策略本体模型的编队协同控制框架;利用贝叶斯网络实现无人机编队的态势检测,并提出基于个体激活期望值的强化学习方法,将学习到的知识迁移到相似新任务中。以海洋环境为背景进行仿真实验,结果表明:在满足航向速率和空速的条件下,任意两台无人机相对距离能够快速收敛并趋于稳定;通过融合语义的方法较好地判断飞行关键点和障碍物,实现无人机编队协同控制与决策。

Considering the influence factors of UAV formation control in the uncertain environment,we propose a formation cooperative control method by fusing semantics from the perspective of information processing.It constructed a formation collaborative control framework with uncertain situation detection,uncertain behavior recognition and semantic strategy ontology model;the Bayesian network was used to realize the situation detection of UAV formation,and a reinforcement learning method based on individual activating expectation was proposed to transfer the learned knowledge to similar new tasks;our method was carried out by the simulation experiments with the marine environment as the background.The results show that under the conditions of course speed and airspeed,the relative distance of any two UAVs can be rapidly converged and tend to be stable.And the method by fusing semantics can be used to better judge the key points and obstacles of flight,so as to realize the cooperative control and decision-making of UAVs formation.