Depth-constrained ASV navigation with deep RL and limited sensing
Amirhossein Zhalehmehrabi, Daniele Meli, Francesco Dal Santo, Francesco Trotti, Alessandro Farinelli · 2025
Read Depth-constrained ASV navigation with deep RL and limited sensing by Amirhossein Zhalehmehrabi, Daniele Meli, Francesco Dal Santo, Francesco Trotti, Alessandro Farinelli (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-18253.txt.
Description
Amirhossein Zhalehmehrabi and co-authors teach a surface vessel to steer safely through shallow water with very thin sensing. Autonomous surface vehicles serve many sea tasks, but shallow zones with shifting water pose a hard test. The craft here gets just one depth reading per step from a downward sonar beam.
In their 2025 paper, a reinforcement learning agent must reach a target while shunning unsafe shallows. Gaussian process regression grows a depth map from the sparse sonar hits, giving the agent a richer view as it moves. Training in simulation carried over to real water trials. The results show safer, surer navigation than plain approaches under the same sensing limits.
Catalog record
| CREATOR | Amirhossein Zhalehmehrabi, Daniele Meli, Francesco Dal Santo, Francesco Trotti, Alessandro Farinelli |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-18253