OPTIMIZATION HYBRID PATH PLANNING BASED ON A-STAR ALGORITHM COMBINING WITH DWA

Abstract

While dealing with dynamic scenarios, mobile robot (MR) path planning (PP) has proven to be extremely challenging. To overcome difficulties with grid-map representation, the paper proposes a hybrid path planning strategy. In a typical setting, an A-star algorithm first generates a global path planning (GPP). To ensure safe obstructive avoidance, the modified heuristic function is enhanced with a risk cost coefficient adjusted according to the distance to the nearest obstacle. Then, GPP is improved by eliminating redundant nodes, and smoothing algorithms. Furthermore, local path planning (LPP) has enhanced the ability to avoid obstacles in local areas by using adaptive dynamic window approach (DWA). The obtained path is hardly continuous after eliminating unnecessary path nodes and the required re-calculation LPP based on DWA. Consequently, straight, and curved segments are joined into the continuous polyline smoothed by the Bezier curves. The proposed hybrid PP has been shown to be feasible and to produce optimal outcomes in both simulations and experiments.