Collision Avoidance for Autonomous Inland Vessels using Stereovision

Abstract

The inland waterway once enabled an industrial revolution, yet the emergence of coalescent road networks has seen its true worth be all but disregarded. Despite the amenity of unimodal travel being compelling, growing awareness for sustainability has reignited interest in more fuel efficient modalities for transportation. Autonomous shipping has the potential to increase efficiency, reliability and safety and will arguably play a major role in the evolving transport revolution, returning the transport modality to its former glory.

The main objective of this research is to cater a collision avoidance strategy to the inland waterway through the development of tailored Guidance and Navigation Systems. An approach to local path planning is introduced to handle the challenges of collision avoidance on the inland waterway and a better understanding of the primary role that stereovision sensors could assume in enabling inland autonomy is gained. Achieving this objective requires first a reflection upon existing work through a study into the state-of-the-art. Subsequently, the Guidance and Navigation Systems are developed and implemented on a scale test vessel. Finally experimental testing is conducted for the evaluation of the developed system performance.