A multi-objective parametric algorithm for sensor-based navigation in uncharted terrains

Abstract

Sensor-based motion planning is one the most challenging tasks in robotics where various approaches and algorithms have been proposed to achieve different planning goals. However, these approaches only focus on one single objective, i.e. path optimality, path safety, efficiency or trajectory smoothness. In this paper, a novel parametric algorithm is proposed that is able to handle different planning goals by means of a set of objective-controller parameters. These parameters are designed carefully to cover different requirements of the path planner. Readings of the sensors will be evaluated to determine the values of four decision variables and the next position of the robot will be selected accordingly. The performance of the proposed algorithm was tested through simulation studies in different types of environments to evaluate its ability to achieve different planning goals. Simulation studies have shown the algorithm to perform robust and effective in all environments.