Gap Based Novel Approach for Safe and Fast Obstacle Avoidance for Autonomous Platforms

Abstract

Obstacle avoidance methods play a vital role in robotic navigation tasks. In this work, we are proposing a novel obstacle avoidance framework for mobile robots equipped with 2d lidars. This method is inspired by previously proposed methods: FGM, SND and their variants. In this work, FGM-I was used as a base. FGM-I was selected for its easy applicability and adequate performance on moderate problems. However, there are certain drawbacks associated with the use of this algorithm in cluttered environments. This method suffers from narrow openings and not having a collision checking procedure. To increase the navigation safety and control robustness, proposed FGM-I2 method uses extra avoidance functions and control procedures. The results of this study indicate that according to overall results, FGM-I2 outperforms four different FGM and SND variants when evaluated w.r.t safety and speed.