부산대학교 도서관

This work investigates dual-agent path planning for marsupial robots with different energy time scales. We extend upon the studied Traveling Salesman Problem with Drone (TSP-D) originally formulated for last-mile package delivery. We contribute a drone energy model and constraints to allow for drone dock charging in a variation we refer to as the TSP-D with Energy Management. A moving-horizon heuristic method is also proposed to reduce the computational load and allow for on-line system feedback. We demonstrate that this heuristic method is computationally more efficient and provides significant path time savings as compared to a single agent solution. Furthermore, it enables the desired system feedback.