用于月球软着陆最优轨迹跟踪制导过程的模糊神经网络控制方法

软着陆过程的制导控制已经成为月球探测中一个具有挑战性的问题。神经网络与模糊逻辑都是处理不确定性(受控对象缺乏精确的数学描述或具有非线性的复杂性)问题的有效手段。本文根据终端着陆条件和性能指标,以由Pontryagin’s极大值原理得出从近月点到月球表面的最优着陆轨迹为基础,给出一种基于模糊神经网络的非线性最优控制策略,使被控系统能够通过模糊神经网络的非线性映射能力实现某种最优的闭环制导控制。最后通过数学仿真验证了该控制策略的可行性与有效性。

FNN Guidance Control for Lunar Optimal Trajectory Tracking Descent

Optimal guidance for soft landing is a challenging problem in lunar exploration. The landing dynamic system, with large disturbances is highly nonlinear, uncertain. Neural network and fuzzy logic were effective tools, which could be used to handle those problems. In this paper, an optimum trajectory from perilune to lunar surface is derived in terms of Pontryagin's maximum principle, which is an open-loop way. Then, an optimal nonlinear control law based on an fuzzy neural network is presented,on the basis of the optimum trajectory, due to the nonlinear mapping function of the fuzzy neural network, which is a closed-loop way. The feasibility and validity of the control law were verified in a simulation experiment.