2017/03/22 by Kristoffer Bergman, Bergman, Kristoffer, Daniel Axehill +1 · 1 citation
Computer Science · Mathematics · Engineering · #Robotic Path Planning Algorithms #Advanced Optimization Algorithms Research #Vehicle Routing Optimization Methods
paper · pdf · doi:10.48550/arxiv.1703.07546
This paper presents a systematic approach for computing local solutions to\nmotion planning problems in non-convex environments using numerical optimal\ncontrol techniques. It extends the range of use of state-of-the-art numerical\noptimal control tools to problem classes where these tools have previously not\nbeen applicable. Today these problems are typically solved using motion\nplanners based on randomized or graph search. The general principle is to\ndefine a homotopy that perturbs, or preferably relaxes, the original problem to\nan easily solved problem. By combining a Sequential Quadratic Programming (SQP)\nmethod with a homotopy approach that gradually transforms the problem from a\nrelaxed one to the original one, practically relevant locally optimal solutions\nto the motion planning problem can be computed. The approach is demonstrated in\nmotion planning problems in challenging 2D and 3D environments, where the\npresented method significantly outperforms a state-of-the-art open-source\noptimizing sampled-based planner commonly used as benchmark.\n