QPA*: Design of a searching and path planning algorithm for intelligent agents in two dimensions
Brian García Sarmina, Georgii Khachaturov · 2020
Searching algorithms and path planning algorithms are in a variety of applications, where the mobile robotics field is one of the most popular. QPA* algorithm attempts to perform this two strategies at the “same time”. The algorithm makes use of three strategies to generate a different solution to the exploration and path planning mainstream, using a “proposed version” of a modified A* (star) algorithm, the idea of the Potential Fields Algorithm (to deal with the collision avoidance) applied as a artificial binary field, and an exploration heuristic named “quadrant grid search”. The QPA* algorithm is designed to be applied in a search and rescue robot, where the problem of “localization” and “mapping” is consider to be independent of the exploring and path planning algorithm. The “exploration” part of QPA * is intended to overcome the main problem of path planning algorithms, that is the lack of a true “exploration factor” and preserve the efficiency of making a route using the path planning approach of A* algorithm. Finally, we test three different approximations (for the path planning part) in combination with the quadrant grid search, in order to identify the pros and cons of this design.