Artificial intelligence and mobile robots
Towards automatic manipulation action planning for service robots
KI'10 Proceedings of the 33rd annual German conference on Advances in artificial intelligence
Hi-index | 0.00 |
Different components are needed to control a service robot, which helps people in the household environment to do daily tasks. Controllers in closed loop are needed to control the actuators and sensors to provide actions. A sequencer schedules and coordinates these actions to finish a given task. And an AI planner is also needed to generate the sequence of the actions. Followed this well-known three-layer architecture, we have developed a robotic framework for service robots. The controller is implemented in the MCA2 (Modular Control Architecture), developed by our insitute. A component-based interface converts the events like start or finished signals to an event-based system, where the tasks are decomposed and organized in a task-tree. This task-tree can be automatically generated by an AI planner or learned from human demonstrations. We have evaluated the introduced control framework on a service robot with two arms and two hands. Experimental results show that our system is modular, flexible and fulfills the requirements for robotic control software development.