Reachability-based analysis for Probabilistic Roadmap planners

摘要: In the last fifteen years, sampling-based planners like Probabilistic Roadmap Method (prm) have proved to be successful in solving complex motion planning problems. While theoretically, complexity of problem is exponential number degrees freedom, can successfully handle this curse dimensionality practice. We give a reachability-based analysis for these which leads better understanding success approach. This compares techniques based on coverage and connectivity free configuration space. The experiments show, contrary general belief, that main challenge not getting space covered but nodes connected, especially when problems get more complicated, e.g. narrow passage present. By using knowledge, we tackle by incorporating refined neighbor selection strategy, hybrid sampling powerful local planner, leading considerable speed-up.