Towards Intelligent Robot Assistants for the non-destructive Disassembly of End of Life Products

The effective collaboration between humans and robots in complex and task rich environments like End of Life product disassembly depends on the ability of the robot to anticipate the workflow as well as the assistance the human co-worker wants. Our approach towards such an intelligent system is the development of an informed software agent that controls the robot assistance behavior. We inform the agent with procedural and declarative knowledge about the disassembly domain through models of the product structure and actor/object models. The product structure is then transformed to a directed graph and used to build, share and define a goal-orientated coarse workflow. Depending on the tasks and wanted assistance, the system can generate adaptable and detailed workflows through searching in the situation space on the basis of predefined and task dependent actions. The created detailed workflow consists of a sequence of actions that are used to call, parameterize and execute robot programs for the fulfillment of the assistance. The aim of this research is to equip robot systems with higher cognitive skills to allow them to be autonomous in the performance of their assistance to improve the ergonomics of disassembly workstations.