Path planning and collision avoidance for robots (Q1760185)

This paper present the calculation of a time optimal trajectory of a robot where obstacles must be taken into account. For this purpose, the authors employ linear programming arguments to model the collision avoidance strategy and a sequential quadratic programming technique to compute the optimal control. Since this approach may lead to very large problems, the authors incorporate a so-called backface culling active set strategy to reduce the number of state constraints. This key feature of the proposed method is discussed and analysed in detail. Detailed numerical results are presented for a 3-link robot. A detailed model of this robot is also given in the paper.