Abstract

Path planning has been a hot research topic in robotics, and it is also a vital functionality for autonomous systems. Generating optimal-quality path plans with the least computing time has always been the goal of researchers. As the time complexity of traditional path planning algorithms grows rapidly with the scale and complexity of the problem, evolutionary algorithms are widely applied due to their capability of giving near-optimal solutions to complex problems. However, evolutionary algorithms can be easily trapped in a local optimum and may converge to infeasible solutions. As the scale of the problem increases, evolutionary algorithms usually cannot find a feasible solution with random exploration, which makes it extremely challenging to solve long-range path-planning problems with these algorithms in environments filled with obstacles. This paper introduces a novel area-based collision assessment method for Genetic Algorithm (GA) that can guide the algorithm to find solutions with a variable number of waypoints in large-scale and obstacle-filled environments. To avoid premature convergence, the mutation process is replaced by a self-improving process to let the algorithm focus the operations on any potential solutions before discarding them in the selection process. The case studies show that the proposed GA-focus algorithm can be applied to different types of large-scale and challenging environments, escape local optimums, and find high-quality solutions.

This content is only available via PDF.
You do not currently have access to this content.