期刊
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART G-JOURNAL OF AEROSPACE ENGINEERING
卷 237, 期 10, 页码 2240-2251出版社
SAGE PUBLICATIONS LTD
DOI: 10.1177/09544100221149231
关键词
Co-design; structure control; sensor; actuator placement; flexible wing; reinforcement learning
This paper applies reinforcement learning to optimize the placement of sensors/actuators and control strategies for a flexible wing. The co-design objective is to achieve optimal closed-loop performance by finding the optimal sensor/actuator placement (OSAP) and associated controller. The problem is formulated as mixed-integer semi-definite programming (MISDP) and solved using a modified reinforcement learning algorithm, which outperforms the greedy algorithm and genetic algorithm in solving high-dimensional MISDP.
This paper presents applying reinforcement learning to find the optimal sensor/actuator placement (OSAP) policy and optimal control for the flexible wing. The co-design objective is to find the OSAP and its associate controller to render the optimal closed-loop performance. The nonlinear vibration dynamics of the flexible wing are modeled in the linear parameter varying (LPV) approach so that LPV-H- infinity controllers can be designed. The co-design problem is formulated into mixed-integer semi-definite programming (MISDP). As a special form of combinatorial optimization, MIDSP solves integer optimization for sensor/actuator selection and convex optimization for controller design. A modified reinforcement learning algorithm is applied to solve this NP-hard optimization problem and obtain a converged solution. In addition, RL is compared with the greedy algorithm and genetic algorithm to demonstrate its strengths and drawbacks in solving high-dimensional MISDP. The solutions obtained by RL and the greedy algorithm are verified and compared in the high-fidelity simulation with the full-order model.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据