Cubic Spline Trajectory Planning and Vibration Suppression of Semiconductor Wafer Transfer Robot Arm

Vibration-free motion in minimal time is desired for industrial robotic applications. Hence, these criteria have to be considered during trajectory planning for a robot arm, wherein polynomial splines are often used for interpolating the trajectory through several via points. Among polynomial splines, the cubic spline is the lowest-degree spline that can provide jerk limitation, a feature that is important for reducing vibration during motion. However, using jerk limitation alone does not eliminate vibration completely and sometimes restricts the performance of industrial robots. This paper proposes an implementation of cubic spline optimization with free via points for reducing motion time, combined with input shaping for suppressing vibration. Experiments are conducted on a semiconductor wafer transfer robot arm to demonstrate the effectiveness of the proposed approach.

Path Planning for Collision Avoidance Maneuver

This paper summarizes the development of an optimal path planning algorithm for collision avoidance maneuver. The goal of the optimal path is to minimize distance to the target vehicle ahead of the host vehicle subject to vehicle and environment constraints. Such path constrained by allowable lateral (centripetal) acceleration and lateral acceleration rate (jerk). Two algorithms with and without lateral jerk limitation, are presented. The algorithms were implemented in Simulink and verified in CarSim. The results indicate that the lateral jerk limitation increases time-to-collision threshold and leads to a larger distance to the target required for emergency lane change. Collision avoidance path without lateral jerk limitation minimizes the distance to the target vehicle and is suitable for path tracking control in real-time application; however tracking such a path requires very aggressive control.

Path Tracking Design Based on Jerk Input Fractional Prefilter and Jerk Limited Input Shaper Approaches

In path tracking design, different methods were developed. In previous works we have presented a method based on fractional differentiation in order to reduce overshoots on the actuator output. It allows the generation of an optimal movement reference-input leading to a minimum path completion time, taking into account the maximum velocity, acceleration and jerk, and the bandwidth of the closed-loop on which the input is applied. Different strategies were developed. In this paper, three methods taking into account the jerk limitation are compared. A comparison with classical input shaper approach is also done. The two first methods, based on fractional differentiation, are used through a Davidson-Cole prefilter with acceleration or jerk bang bang input, to ensure continuity on acceleration and jerk signals on the actuator. The third approach is based on preshaper synthesis with jerk limited input; the shaped input is obtained by convolving desired input with an impulse sequence. The synthesis methodologies of the different methods are studied. Performances are compared on an example, in both frequency and time domains.