Non-Dominated Sorting Genetic Algorithm for smooth path planning in unknown environments

Abstract

Autonomous robots have been the focus of attention of most researchers, particularly when it is imputed with terms like intelligence and autonomy. The most important challenge encounters autonomous navigation of a mobile robot is established from large amounts of uncertainties that are coupled with natural environment. This includes hazy and cloudy information of the environment. Moreover, continuous and fast changes of the real environment require a fast response from the robot. Many algorithms have been proposed and amongst these, the potential field algorithm is widely used. This work aims at optimizing some parameters involved in the potential field by the use of Non-Dominated Sorting Genetic Algorithm II (NSGA II). This paper takes into account the safety margin around the obstacle along with the size of the robot which also affects its motion during the optimization process in order to ensure the optimal path.

Mobile Robot Path Planning with a Non-Dominated Sorting Genetic Algorithm

“…In [54], an improved crossover operator was designed in static environments. In [55], NSGA-II was used to improve the clearance and smoothness of the path by optimizing the three parameters obtained by the potential field method. Moreover, the position of the virtual obstacle was redefined for the case where the end point of the path was farther away from the obstacle.…”