Planning of Time-Efficient Trajectories for Mobile Robots with Differential-Drives

To effectively utilize the capacities of the production facilities, flexible and fast transport, and manipulation technology is essential. Current path planning methods use simple paths without exploiting the physical limitations of mobile robots. This paper deals with the time-efficient kinematic planning of path and velocity trajectories for autonomous mobile robots equipped with a differential-drive, used to load and unload machine tools. For a close parking position next to the machine tools, the mobile robot must approach practically straight without the necessity of turning at the end. After a brief definition of the kinematics of a mobile robot with nonholonomic constraints, a Bezier-curve is combined with multiple jerk-limited velocity trajectories considering the initial and final state as well as the physical limitations. For a time-efficient acceleration and deceleration the maximum acceleration of the mobile robot is used and differentiated into the radial and tangential acceleration. This allows a quickly reached close parking pose next to a machine tool. The algorithm is tested with simulations in RVIZ in combination with Gazebo and ROS2 as well as with the real mobile robot Tend-o-Bot.

I. Introduction

For the effective utilization of production facilities, flexible and autonomous transportation and manipulation technologies are indispensable. This includes the loading and unloading of machine tools, which is usually carried out manually nowadays and therefore is cost-intensive, ergonomically demanding and prone to errors. Existing system solutions for automatic ma-chine loading are currently stationary or track-guided and thus, limited in their freedom of movement. To increase the loading and unloading flexibility, the current stationary system solutions must therefore be expanded to include an autonomous, off-track transport system with cognitive capabilities.