Algorithmic approach to action planning for assembly automation using state matrix representation

Action planning ability is essential for autonomous behavior of an intelligent control system. This paper presents a new generic algorithm for action planning in a robotic workcell which provides a mechanism for incorporation of domain knowledge. The planning system is based on a new representation scheme called state matrix representation which uses the inter-relationships of objects to model the world. The robot assembly domain is chosen to demonstrate the merits of our planner and the new representational scheme. However the planning methodology and the state and action representation schemes are not limited to the assembly automation domain. An example of automatic plan generation for an assembly operation based on this representation and planning methodology is given