基于神经网络的多机协同编队控制器设计  

Neural Network-Based Formation Controller Design of Multi Aircraft Cooperative

作　　者：纪超[1] 王昆[2] 

机构地区：[1]西安医学院医学技术系,陕西西安710021 [2]西安工业大学人事处,陕西西安710021

出　　处：《自动化技术与应用》2015年第7期36-41,共6页Techniques of Automation and Applications

摘　　要：针对多架无人机协同编队飞行控制问题,设计了一种基于BP神经网络的无人机编队飞行控制器。以两机编队为单元,僚机同时跟踪长机和相邻僚机,根据相对位置和参考坐标系统,采用BP神经网络训练得到最佳的PID控制参数,设计三通道PID控制器并对编队系统进行分布式协同控制,使系统快速跟踪指令并保持编队队形。对四架无人机组成的编队系统进行仿真,系统编队可快速保持期望队形,表明设计的编队控制系统具有良好的稳定性和较强的抗干扰能力。Aiming at the cooperative formation flight control problem for multiple UAV（Unmanned Aerial Vehicles）, UAV formation flight controller based on BP neural network is designed. Be distributed to the two-plane formation as a unit, while a wingman tracking the lead plane and adjacent another wingman, according to the relative position and the reference coordinate system. The three-channel PID controller is designed and distributed cooperative control of formation system, by using the best PID control parameters obtained from BP neural network training, that makes the system quickly and keep track formation flying instruction. Composed for four UAV formation system simulation, system formation quickly to maintain the expected formation. It indicates that the design of formation control system has good stability and strong antijamming capability.

关 键 词：无人机 协同编队 神经网络 分布式控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]