基于神经网络的多四旋翼保性能编队控制

针对多四旋翼编队飞行过程中存在的参数和外界扰动不确定性导致编队系统不稳定问题，设计了一种多四旋翼分布式神经自适应动态面协同编队控制方法。采用学习维数低、实时性强的最小参数学习神经网络观测器，实现对四旋翼位置回路和角度回路未知非线性干扰的快速平滑学习与补偿。利用预设性能控制方法将原先受约束的同步误差经转换函数变换为不受约束的误差，解决协同编队中的位置一致性控制问题，确保各四旋翼位置跟踪同步误差能够按照预设的收敛速度、超调量及稳态误差收敛至期望的区域。仿真实验表明，所提多四旋翼协同抗干扰编队方法在较大扰动环境下仍能达到良好的控制效果。

Guaranteed Performance formation control of multi-quadrotor Based on neural network

In this paper, a distributed neural adaptive formation control method is proposed for multiple quad-rotors to solve the problem of formation system instability caused by uncertainties of parameters and external disturbances. A minimum learning parameter based neural network observer with strong real-time calculation and low learning dimension is used to realize smooth and fast online learning and compensation of unknown nonlinear disturbance in the position loop and angle loop of quad-rotors. By using the prescribed performance control (PPC), the constrained synchronization error is converted into unconstrained error through the error conversion function to solve the position con-sistency control problem in cooperative formation, and ensure that the position tracking synchroni-zation error of each quadrotor can converge to the desired region according to the preset convergence speed, overshoot and steady-state error. Simulation results show that the proposed formation con-troller has good control performance even in large disturbance environment.