Showing 1–1 of 1 results for author: Gershfeld, N

Adaptive and Collaborative Bathymetric Channel-Finding Approach for Multiple Autonomous Marine Vehicles

Authors: Nikolai Gershfeld, Tyler M Paine, Michael R. Benjamin

Abstract: This paper reports an investigation into the problem of rapid identification of a channel that crosses a body of water using one or more Unmanned Surface Vehicles (USV). A new algorithm called Proposal Based Adaptive Channel Search (PBACS) is presented as a potential solution that improves upon current methods. The empirical performance of PBACS is compared to lawnmower surveying and to Markov dec… ▽ More This paper reports an investigation into the problem of rapid identification of a channel that crosses a body of water using one or more Unmanned Surface Vehicles (USV). A new algorithm called Proposal Based Adaptive Channel Search (PBACS) is presented as a potential solution that improves upon current methods. The empirical performance of PBACS is compared to lawnmower surveying and to Markov decision process (MDP) planning with two state-of-the-art reward functions: Upper Confidence Bound (UCB) and Maximum Value Information (MVI). The performance of each method is evaluated through comparison of the time it takes to identify a continuous channel through an area, using one, two, three, or four USVs. The performance of each method is compared across ten simulated bathymetry scenarios and one field area, each with different channel layouts. The results from simulations and field trials indicate that on average multi-vehicle PBACS outperforms lawnmower, UCB, and MVI based methods, especially when at least three vehicles are used. △ Less

Submitted 27 May, 2023; v1 submitted 20 September, 2022; originally announced September 2022.

Comments: 8 pages, 9 figures, (v1) Submitted to IEEE International Conference on Robotics and Automation (ICRA) 2023, (v2) Updated figures, Submitted to IEEE Robotics and Automation Letters (RA-L), (v3) Preprint accepted by IEEE Robotics and Automation Letters (RA-L)

Journal ref: IEEE Robotics and Automation Letters ( Volume: 8, Issue: 7, July 2023)