A motion planning algorithm for the invalid initial state disassembly problem

Sampling-based motion planners are able to plan disassembly paths at high performance. They are limited by the fact that the input triangle sets of the static and dynamic object need to be free of collision in the initial and all following states. In real world applications, like the disassembly planning in car industry, this often does not hold true. Beside data inaccuracy, this is mainly caused by the modeling of flexible parts as rigid bodies, especially fixture elements like clips. They cause the invalid initial state disassembly problem. In the literature there exists no algorithm that is able to calculate a reasonable disassembly path for an invalid initial state. Our novel algorithm overcomes this limitation by computing information about the flexible parts of the dynamic object and incorporating this information into the disassembly planning.