Smooth Control of A Lightweight Humanoid Robot Arm

The problems of flexible vibration and movement discontinuity caused by the inertial load and limitation of joint control accuracy are usually difficult to overcome for the Lightweight Humanoid Robot Arm (LWH-Arm) we built early. To solve these problems, we propose a control method called SCI5P (Smooth Control based on Interpolating 5-degree Polynomial), which interpolate between specific path points using a 5-degree polynomial, thus, improving the stability and smoothness of the arm speed, acceleration and pulse-width control of the joint servos. We take an LWH-Arm prototype as the research object, establish the kinematic model, constrain and optimize movement trajectories according to the joint movement characteristics. On the control layer, we design motor control algorithm with the movement time and incremental fragment of Pulse-Width buffer can be regulated. Compared with traditional control method, SCI5P optimizes the continuity and smoothness of the trajectory, angular velocity and acceleration, while the LWH-Arm is working on the task. Furthermore, it also enhances the stationarity of the motor stepping. Through grabbing bottle and pouring water experiments, the effectiveness of SCI5P is verified. The analyses of movement fitting curve show the flexibility and smoothness of LWH-Arm are also improved.