Total Energy Cost Optimization for Data Collection With Boat-Assisted Drone: A Study on Large-Scale Marine Sensor

Collecting broad ocean region data to leverage large-scale sensors is a valid method to handle ocean health problems associated with human activities (wind turbine deployment, nuclear wastewater discharge, etc.). When using the novel collection scheme, reducing the total energy consumption (TEC) of boat-assisted drones to collect sensor data is challenging. The objective of this study was to minimize the TEC of boats and drones (owing to their limited battery capacity) during marine environment sensor data collection. To achieve this, new models for drone hovering in wind, data collection, and related wireless communication have been developed, and the TEC minimization problem has been formulated as a new specialized distance-constrained capacitated vehicle routing problem. The problem is divided into four subproblems to reduce complexity. Based on these four subproblems, an improved heuristic algorithm was proposed. In the algorithm, the drone hovering position and boat waypoint are determined using the K-means clustering algorithm and the smallest enclosing circle algorithm. Based on the position and waypoint, the routes of drones and boats were optimized using the Lin-Kernighan heuristic 3 algorithm, thus minimizing the TEC. The simulation results demonstrate that when the boat waypoint is 3 and the sensor number is 2000, owing to the strong local and global search ability, the TEC in the scheme is
$0.03{\times }10 ^{8}\text{J}$
less than that of the graph attention neural network method (GANN), while the scheme also provides time saving, scalability, and flexibility.