Using probabilistic roadmaps in changing environments

Abstract

In this paper, we examine how a path planning problem can be solved in changing environments using probabilistic roadmap planners. A probabilistic roadmap is built in static environment where all obstacles are known in advance, but we show that a roadmap can be built in such a way that it works well even when new obstacles are added to the workspace. However, our experiments show that the roadmap graph must be built carefully. We compare three different methods that are used to decide which edges are added to the roadmap graph to connect the nodes. One of these is a distance‐based method, which we present in this paper. In the tests, we built a roadmap by using only the static obstacles. Then, we added additional obstacles to the environment and tested how well the roadmap still worked. The tests showed that our distance‐based method worked quickly and that it produced roadmaps, which could be used to find a path amid additional obstacles with a high success rate.Copyright © 2013 John Wiley & Sons, Ltd.

Explicit Energy-Minimal Short-Term Path Planning for Collision Avoidance in Crowd Simulation

“…Owing to the tradeoff between the level of discretization and the performance, using a set of connected graphs to represent the walkable regions is a good choice to compromise between both of them. Many researchers construct such graphs based on different approaches, including the randomized method [8], Delaunay triangulation [9], navigation graph [10], voronoi diagram [11,12], and medial axis [13,14,15,16,17]. The graph-based path planners yield a small-sized search space but the queried path, if it exists, is not the shortest one as produced in the grid-based planners.…”

A New Path Planning Method of Obstacle and Singularity Avoidance for Redundant Robot

“…It is difficult to solve the contradictions among planning time, algorithm complexity, and optimal path existing in these methods. In the field of robot obstacle avoidance, the mainly applied domestic and overseas research methods include DLS method [7] , SICQP method [8] , normal form method [9] , etc. [10][11][12] .…”

Improvement of Obstacle and Singularity Avoidance Path Planning Algorithm for Redundant Manipulators

“…Compared with the traditional manipulator the redundant manipulator has the additional degrees of freedom, makes it has the advantages of flexible operation and can complete obstacle and singularity avoidance. In the field of obstacle-avoiding of the manipulator, the main research method including the artificial potential field method [4] ,fuzzy method [5] , neural network method [6] , genetic algorithm [7] , Probabilistic Roadmaps method [8] ,Rapidly-exploring Random Tree method [9] .However, these methods have some shortcomings and the contradiction between the optimal path the planning time and the complexity of algorithm is very difficult to solve. In the area of singularity-avoiding of the manipulator, the main research methods such as the Damped least square method [10] , SICQP method [11] , normal forms method [12] .…”