An online task planning framework reducing execution times in industrial environments

摘要

In this paper we present a symbolic planning framework that is integrated into a 3D simulation environment using the concept of Digital Twins for just-in-time planning of intelligent multi-robot systems with improved execution times. Instructions can be given to the system at any time and plans with potentially dozens of actions are generated and executed in rapid succession. The framework uses spacial information of deposits and fixed locations during planning to determine overall shortest paths for robot movements while executing plans. All information that is required for the planning process is gathered from the simulated environment, thereby allowing for ostensive insights into the system state. Not only is this approach useful during the development and maintenance of an intelligent robot system, but it also enables functionality and user interaction possibilities that are otherwise not available. All the functionality covering the modeling, simulation and control of intelligent multi-robot systems was integrated into a single application. The approach was validated with a sample scenario, including two robots on linear axes with interchangeable grippers and several objects and deposits.