Coverage path planning for multiple unmanned aerial vehicles in maritime search and rescue operations

• Published in: Computers & industrial engineering Vol. 161; p. 107612
• Main Authors: Cho, Sung Won, Park, Hyun Ji, Lee, Hanseob, Shim, David Hyunchul, Kim, Sun-Young
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2021
• Subjects: Coverage path planning (CPP); Maritime; Search and rescue (SAR); Unmanned Aerial Vehicle (UAV); Search and rescue (SAR); Unmanned Aerial Vehicle (UAV); Coverage path planning (CPP); Maritime
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2021.107612

•We present a two phase method for solving the CPP problem of multiple heterogeneous UAVs.•RSH decreases the computation time by 87% compared to a commercial solver.•RSH yields near-optimal solutions with an average solution gap of 0.7% compared to a commercial solver.•The proposed method improves the solution obtained by Maza and Ollero (2007) by an average of 7.9%.•The results of real-world flight experiments differ by 16% compared to the simulation results. The number of casualties in the maritime sector is increasing consistently. To reduce the scope of search area extensions in the event of a maritime accident, maritime authorities and operation centers are currently trying to develop a quick search for survivors at sea using unmanned aerial vehicles (UAVs) in maritime search and rescue (SAR) operations. Here, we propose a two-phase method for solving the coverage path planning (CPP) problem of multiple-UAV areas in maritime SAR. In phase 1, we propose a grid-based area decomposition method that minimizes the decomposed search area to transform the search area into a graph made up of vertices and edges. In phase 2, we formulate a mixed-integer linear programming (MILP) model to derive an optimal coverage path that minimizes the completion time. To solve the model for large-scale instances, a randomized search heuristic (RSH) algorithm is developed. We conducted extensive numerical experiments to validate the performance of the algorithm. Experimental results show that the RSH yields a better solution with an approximately 0.7% optimality gap within a much shorter computation time than that of a commercial solver. In addition, our grid-based CPP algorithm outperforms those used in previous research with respect to the solution quality. Furthermore, we showed the results of real flight experiments in the marine field using the proposed algorithm.