自由漂浮空间机器人多约束混合整数预测控制

A Mixed Integer Predictive Controller with Multi-Constraint for Free-Floating Space Robots

针对空间机器人完成任务时需要躲避障碍物的问题,提出一种自由漂浮空间机器人的混合整数预测控制方法。首先,在模型预测控制方法框架下,机械臂关节的物理限制,躲避障碍物的要求被统一描述为最优控制问题下的不等式约束,可以得到自由漂浮空间机器人具有线性二次规划形式的最优控制律。其次,基于命题逻辑建立控制问题中各约束的优先级,保证在最大程度地满足约束的情形下得到控制问题的解,有效弥补了模型预测控制方法用于空间机器人控制时,多约束可能导致最优控制问题不可行的不足。最后,仿真结果校验了所设计控制律的有效性。

For the obstacle avoidance problem of space robots,a mixed integer predictive controller is developed for the free-floating space robots. Firstly,in the frame of model predictive control method,the joint physical limits and obstacle avoidance requirement are described as inequality constraints of the optimal control problem,and the controller with linear quadratic programming form can be obtained. Secondly,priorities of the constraints in the controller are established based on the propositional logic theory to guarantee the optimal solution to be derived under the satisfaction of the maximum number of constraints,which effectively makes up for the deficiency issue that multiple constraints may lead to the optimal control problem infeasible when model predictive control method is used to control space robots. Finally,numerical simulations are conducted to prove the effectiveness of the proposed control strategy.