What is the challenge that is being addressed?

The goal of this use-case is to explore and demonstrate how task planning and execution can be used to control a mobile robot for manipulation tasks in an indoor environment.

In a demonstration example the mobile robot works in an environment where boxes and other objects are placed on tables, but the robot initially does not know their positions. An example task for the robot is to insert a screwdriver into a box and place both on a specified table. Due to the unknown positions of the objects the robot has to search for them. Furthermore, the objets can also be moved by other actors like human workers. Therefore, the robot also has to take such changes into account and react accordingly. In a second example the robot has to assist human workers by bringing them a powerdrill.

What is the AI solution the project plans to implement?

For controlling the robot we use AI Task Planners provided through the Unified Planning (UP) library that is developed by the AIPlan4EU project. This way we create plans for the robot that instruct the robot in which order it should perform its available actions to achieve its goals. Additionally, we use the Embedded Systems Bridge to connect the existing action implementations of the robots with those provided in the resulting plans and also use that Bridge for executing the actions, monitoring their successful execution and adapting to the necessary changes. .