ABSTRACT
Automatic path plmming, i.e. the methodology to automatically find collision free motion patterns of objects being moved in space, is a well established research area and there are a number of distinguished groups in the world. Complete algorithms are of little industrial relevance due to the complexity of the problem (Canny, 1988). Instead, sampling based techniques, trading completeness for speed and simplicity, are typically the methods of choice. Common for these methods are the needs for efficient collision detection, nearest neighbour searching, graph searching and graph representation. The two most popular methods are: Probabilistic Roadmap Methods (PRM) (Bohlin and Kavraki, 2000) and Rapidly-Exploring Random Trees (RRT) (LaValle and Kuffuer, 1999). There are commercial available path planner tools, e.g. from Fraunhofer-Chalmers Centre (FCC) and KineoCam, on the market today that investigate if a part without any external support can find a collision free way from a starting position outside the assembly to its end position in the assembly structure. FCC has, inspired by both these probabilistic methods, developed a novel deterministic path planning algorithm. Path planning tools give valuable support when evaluating new product or workstation concepts and when comparing alternative solutions. Also different assembly sequences can be compared and verified.
