Fluid-based cooperative underwater localization

Abstract

Localization for autonomous underwater vehicles (AUV) is difficult because of the lack of adequate light, distinguishable landmarks and global positioning system (GPS) radio frequency (RF) signals in undersea areas. Due to maneuverability limitations, small AUVs are usually susceptible to strong ocean currents. The impacts of ocean flow on vehicle motion are usually too large to be considered as disturbances. Simulations of large-scale ocean currents are available to track changes of the background flow. Fluid-based multi-vehicle path planning algorithms have been developed to achieve the feasibility in the presence of the background flow and optimized energy consumption. With certain information from both simulations and measurements of the background flow, the performance of multi-AUV cooperative localization can be improved. In this work, a multi-AUV cooperation hierarchy is proposed. More capable AUVs with bounded localization errors are used as localization references to improve the localization performance of other low-cost AUVs. The proposed algorithm is fully distributed and it is verified by using a modified extended Kalman filter (MEKF). This paper focuses on the outline of the algorithm and its associated matching techniques. The performance of the algorithm is evaluated based on simulations in the flow field generated by the N-vortex system. Localization errors of low-cost AUVs are all bounded at satisfactory levels. The diverging behavior of the pure cooperative localization is effectively avoided.