基于自适应动态规划的四旋翼无人机分布式最优协调控制

Distributed Optimal Coordination Control for the Quadrotors Unmanned Aerial Vehicle via Adaptive Dynamic Programming

为了提高四旋翼无人机编队的自适应能力,通过自适应动态规划(adaptive dynamic programming,ADP)方法研究四旋翼无人机编队分布式最优协调控制问题。将该算法引入到分布式协调控制器的设计中,既避免了冗余数据的传输,又使各无人机的性能函数最小化。基于模糊双曲模型的评价神经网络(critic neural network,CNN)逼近值函数,以实现控制策略的设计。闭环系统稳定性证明权值估计误差和局部邻域协调误差是一致最终有界的。最后,进行了有向切换通信拓扑下的四旋翼编队仿真实例,结果表明了该算法的有效性。

In order to improve the adaptive capability of quadrotors unmanned aerial vehicle(UAV)formation,the problem of distributed optimal cooperative control of quadrotors UAV formation was studied by adaptive dynamic programming(ADP).This algorithm was introduced into the design of distributed cooperative controller,which not only avoided the transmission of redundant data,but also minimized the performance function of each agent.In order to implement the proposed algorithm,the fuzzy hyperbolic model based cri-tic neural network(CNN)was used to approximate the value function and achieve the design of control policies.It was proved that the weight estimation errors and local neighborhood coordination errors were uniformly ultimately bounded.Finally,the effectiveness of the ADP algorithm was verified by the simulation results of the quadrotors formation under directional switching communication topology.