A Task Planner for Intelligent Mobile Robots using Subgoal Priority Information

We have developed a fast task planner for intelligent mobile robots. The speed gain of this task planner over traditional task planners is achieved by avoiding goal-interaction and by utilizing a parallel comparison network. Goal-interaction during task planning is avoided by executing subgoals in a depth first fashion according to the priority numbers attached to subtasks and preconditions of the robot operations in the rule base. The assignment of these priority numbers is based on spatial relationship between objects in the robot environment and time sequence of the subtasks. A combination of hardware circuits and software program is used to search proper robot operations in parallel. Special problems encountered in the design of this task planner are discussed and examples are given to demonstrate the advantages of this task planner. This study provides a new approach which significantly increases the speed of robot task planning.