Trajectory Control of Robot Arm using Modern Control Theory

本课题的主要目的是验证现代控制理论在机械臂轨迹控制中的有效性。为此，我们根据现代控制理论提出了三种控制方案。第一种是利用干扰定位理论构造的“约束运动控制方法”。通过这种方法，控制手臂尖端的位置和速度，使其跟踪所需的函数，无稳态误差。第二种是基于频域跟踪理论构建的“迭代控制方法”，使机器人重复进行伺服动作，直到跟踪误差足够小为止。通过仿真和简单的实验验证了这两种控制方法对机械臂的轨迹控制是有效的。第三种控制方案是“自适应力控制”，当不能确定臂尖必须跟随的物体的阻尼系数时，对臂尖的力和轨迹进行独立控制。此外，还提出了两种相关的数字技术。“数字控制计数计算时延”和“强稳定最优输出反馈控制法”分别是在不能忽略微处理器处理时间和仅检测位置信号实现状态反馈控制时所需要的。我们认为，这些理论都可以应用于机械臂的实际控制问题。

The main purpose of this project is to demonstrate the effectiveness of modern control theory for trajectory control of robot arm. For this aim, we proposed three control shcemes depending on the modern control theory. The first is "constraint motion control method" which is constructed using the distrubance localization theory. By this method, the position and velocity of the tip of the arm are controlled so that they track the desired functions with no steady state errors. The second is "iterative control method" which is constructed based on the frequency domain tracking theory such that the robot repeats servo-actions until the tracking error becomes sufficiently small. We confirm that both of these control mehtods are effective for trajectory control of robot arm by simulatiuon and by simple experiment. The third control scheme is "adaptive force control" in which the force and the trajectoty of the tip of the arm are independently controlled when we could not identify the damping coefficient of the body on which the tip must follows. Further, two related digital techniques are proposed. They are "digital control counting computation time delay" and "strongly stable optimal output feedback control method" which are required when the processing time of the microprocessor can not be neglected and the position signal is only detected to realize the state feedback control, respectively. All these theories, we believe, shoud be used for practical control problems of robot arm.