摘要
针对无人机最优路径规划问题,提出了基于深度学习自动微分的无人机路径规划算法,给出了无约束的非线性优化问题求解方法以及约束处理方法,研究了最优控制问题求解的控制向量参数化方法,并对有范德波尔方程约束的最优控制问题进行了仿真求解,最后进行了无人机最优路径规划,建立了无人机状态方程以及差分运动方程,实现了无人机在多个障碍物下的最优路径规划问题求解,通过计算机仿真分析,基于深度学习自动微分的无人机路径规划算法具有较强的实时性和鲁棒性。
Aiming at the problem of optimal path planning for unmanned aerial vehicles(UAV),this paper proposes a UAV path planning algorithm based on deep learning with automatic differentiation.It presents a method for solving unconstrained nonlinear optimization problems and constraint handling approaches.Furthermore,the control vector parameterization method for solving optimal control problems is investigated,and a simulation solution is provided for an optimal control problem constrained by the van der Pol equation.Finally,the optimal path planning for UAV is carried out by establishing the UAV state equations and differential motion equations.This approach enables the solution of optimal path planning problems for UAV navigating through multiple obstacles.Through computer simulation analysis,the UAV path planning algorithm based on deep learning with automatic differentiation demonstrates strong real-time performance and robustness.
作者
屈博琛
QU Bochen(School of Control Science and Engineering,China University of Petroleum(East China),Qingdao 266580,China)
出处
《自动化与仪表》
2025年第2期66-72,共7页
Automation & Instrumentation
关键词
非线性优化
控制向量参数化
最优路径规划
深度学习
自动微分
nonlinear optimization
control vector parameterization
optimal path planning
deep learning
automatic differentiation
