Previous Abstract Return to Session D2 Next Abstract

Session D2: Marine Vehicle Navigation

AUV Navigation Aiding using Multiple Surface Vehicles
Dula Na, Ivan Lonar, Nikola Miškovi, UNIZG-FER, Croatia
Location: Spyglass
Alternate Number 1


With more than 70% of Earth's surface covered with water, marine monitoring and exploration is an important topic. Conventional exploration is centered around single, fully equipped underwater or surface vehicles. Often these vehicles carry larger payloads to accommodate different sensor suites, making them problematic for deployment in field. Newer trends investigate distributed sensing where sensors are not concentrated around a single vehicle, but rather around multiple smaller vehicles and/or marine stations covering a larger area. These vehicles can be smaller and are easily deployable from a small boat or from shore. However, teleoperation of multiple vehicles is a complex task for a single operator and more autonomous behaviour is required. The research in the paper is carried out as part of the CroMarX project which applies cooperative control between multiple surface and underwater vehicles to extend the Croatian capabilities in marine monitoring of the Adriatic sea.
The paper will introduce a concept of multiple small autonomous surface vehicles (USVs) that perform cooperative maneuvering in order to provide remote navigation aiding to an underwater vehicle (AUV). Each USV can carry different environmental sensors and additionally carries localization sensor (acoustic modem, camera, sonar) to collect data about the environment and in parallel about the AUV position. The USVs are operating with a consensus based formation keeping with the AUV as a passive participant. Formation keeping is achieved using simple consensus algorithm modelled for system with first-order kinematics. The implemented algorithm inherently does not plan trajectories for USVs, so collisions may occur. To avoid collisions, a virtual repelling force is applied around each vehicle. When two vehicles are near to each other, the algorithm generates the velocity control signal in opposite direction of the other vehicle.
Available real-time data is exchanged between USVs. Exchanged data is then used for formation keeping, collision avoidance, and to get the best estimate of the underwater vehicle position. This position is used to adjust the formation shape based on localization sensor specifications and limitation.
AUV position from different sensors is exchanged through an acoustic gateway with the underwater vehicle. The vehicle fuses available on-board sensor measurements with data arriving through acoustics. The paper will investigate the localization improvement from on-board dead-reckoning to externally aided localization results. Finally, simulation results for the concept with detailed software-in-the-loop simulation will be presented.



Previous Abstract Return to Session D2 Next Abstract