An Efficient Path Planning Model in an Unstructured Obstacle Domain

Abstract

This paper seeks to extend the Hybrid Virtual Force Field (HVFF) concept presented in [1]. The extension includes navigation efficiency study between the current navigation scheme in which the robot identifies a local minima causative trap along its trajectory prior to navigation and avoids it by redefining a new navigation path and an initial scheme as seen in [2] where the robot actually goes into the trap, escapes and navigate towards the target. Furthermore, an algorithm for the Goal Nonreachable problem (GNRP) with obstacle close to the target using an Adaptive Force Field (AFF) concept is presented. The proposed model is further validated on some relatively complex obstacle domain in comparison with the earlier cited work. A 2D domain work envelope which consists of static obstacles with the robot having no knowledge of their relative positions or geometrical configuration prior to navigation is assumed.