GNSS, INS, Stereo Camera and LIDAR Sensor Fusion and Perception for Autonomous Inland Waterway Navigation

Due to the movement towards greener industries and lower operating costs there is renewed interest in transporting goods using inland waterways. However, labour expenditures currently make up about 30% of operational costs. Making the ships autonomous would therefore significantly reduce costs. However, autonomous inland waterway navigation requires very accurate localisation, due to the confined spaces of inland waterways. A common solution to the localisation problem is the combined use of GNSS with Inertial Navigation Systems (INS). It has been used in many other robotic applications, however, it needs to be thoroughly tested to determine whether the combination of GNSS and INS sensors will be accurate and reliable enough for safe navigation in this application. Using GNSS/INS sensor fusion is not the only localisation method available. There is also a weakness using a GNSS based system in that GNSS signals are not completely reliable and are known to be sometimes intermittent. Light Detection And Ranging (LIDAR) and stereo camera is also an option. Recently, a project to map the inland waterways has been completed. The mapping data from this project, coupled with a LIDAR and/or stereo camera could also be an option for localisation. The use of LIDAR and stereo cameras could also be used to constantly update navigational charts based in the cloud, so that all users have access to the most recent navigational data.