Mulltivariate motion planning of autonomous robots.

Abstract

A problem of motion control in robot motion planning is to find a smooth transition while going from one path to another. The key concept of our theory is the steering function, used to manipulate the motion of our vehicle. The steering function determines the robot's position and orientation by controlling path curvature and speed. We also present the - neutral switching method - algorithm that provides the autonomous vehicle with the capability to determine the best leaving point which allows for a smooth transition from one path to another in a model-based polygonal world. The above mentioned algorithm is thoroughly presented, analyzed, and programmed on a Unix workstation, and on the autonomous mobile robot Yamabico. The research data indicate that neutral switching method improved the transition results for polygon tracking, star tracking motion, and circle tracking. Moreover, neutral switching method enhances robot control and provides a more stable transition between paths than any previously known algorithm