Real-time Safe Navigation Strategy for Mobile Robots in 3D:A Deforming Path Approach

Abstract

This work considers the 3D safe navigation problem of autonomous vehicles such as flying drones and underwater vehicles. Such a problem is very challenging especially when dealing with unknown and/or dynamic obstacles. In these situations, the vehicle is required to take quick actions to avoid collisions with detected obstacles. Therefore, the proposed strategy adopts a sense and avoid paradigm to produce collision-free motions. This is achieved by applying real-time deformations to the vehicle's reference path around obstacles based on sensory information which can provide better computational solutions and quicker actions compared with map-based navigation approaches. Also, quintic Bezier splines are utilized to generate smooth deformations. Computer simulations have been conducted to demonstrate the effectiveness of the proposed strategy with unknown/dynamic obstacles.