Self-Tuning Nearness Diagram Navigation

The nearness diagram (ND) navigation method is a reactive navigation method used for obstacle avoidance in which five different robot-environment states are defined and five corresponding actions are designed carefully. In the ND+ navigation method, one more robot-environment state is added and all action equations are reformed to both achieve smoother robot motions and reduce manual parameter tuning. In this paper, we argue that the original five states of the ND navigation method is sufficient with the proposed self-tuning method. Two states’ action rules of the ND navigation method are modified. The wiggling motion in long corridors is significantly reduced, and a parameter is self-tuned so that manual parameter tuning is avoided. The experimental results using a real robot with a laser scanner demonstrated that the proposed self-tuning ND navigation (stND) approach is simple yet effective.