Efficient environment representation for mobile robot path planning using CVT-PRM with Halton sampling

A centroidal Voronoi tessellation-based probabilistic roadmap (CVT-PRM) and its construction method for mobile robot path planning are introduced. The CVT-PRM efficiently encoded the entire unoccupied region of the environment by autonomously rearranging the positions of nodes via CVT and Halton sampling. Simulation results verified that the CVT-PRM can encode the entire unoccupied region efficiently, using evenly distributed nodes.