A collision avoidance algorithm for the mobile robot and the robot manipulator in multi-robot system.

摘要

As the modern, highly automated factory becomes more and more complicated, the collision avoidance between robots and other obstacles in the work space becomes increasingly important as well. A local collision avoidance algorithm which was developed based on the general structure of the Artificial Potential Field Force along with the Strategy Force and the Pseudo Distance Function is presented. In this algorithm, a goal attracting force is defined in order to drive the robot from the starting position to the goal position. When the distance between a robot and obstacle becomes smaller than a pre-defined effective avoidance distance, an artificial repulsion force is generated as a function of the distance resulting in repulsion between the two closing objects. The commanding force (by combining goal attracting force and the repulsion force) drives the robot toward the goal position without colliding with other objects in the work space. Since the repulsion force, developed here, depends on the distance only, the complicated geometric function for the object model is less important in the algorithm used in this dissertation. The simplicity of this improved artificial force collision algorithm would make it possible to be used on a real-time basis within a time variant environment.; Further improvement is realized by introducing a Strategy Force within the Artificial Force whenever a locking situation is detected. The Euclidean distance between the objects normally used in this algorithm is also replaced by a defined Pseudo Distance calculated by a Pseudo Distance Function. The Pseudo Distance Function uses basic geometric information between two objects to calculate the Pseudo Distance. The Pseudo Distance is always smaller than the true minimum distance. This greatly simplified robot collision avoidance algorithm is applicable not only on a two-dimensional mobile robot system but on a three-dimensional multi-robot system as well. Simulations are performed on both mobile robot collision avoidance and multi-robot collision avoidance.