A Bin-Packing-Based Method for Path Planning of Air–Ground Cooperative System

An air-ground cooperative system integrates advantages of an autonomous air vehicle (AAV) and an autonomous ground vehicle (AGV). It has broad application prospects in large-scale intelligent surveillance and reconnaissance tasks. However, AAV cannot finish a large-scale task in one flight due to its limited power. Therefore, the points need to be partitioned into several subsets. AGV provides charging and limited distance communication for AAV. During the flight, AAV needs enough power to visit the points, each of which needs to be located in the communication range. In existing work, K-means is used to partition points. However, its partition results are determined by a preset number of subsets and the distribution of points. The generated subsets may contain too many or too few points such that the efficiency of one flight is reduced. The points in a subset may also be out of the communication range such that AAV cannot reach it directly., thus requiring extra charging and transportation. This work transforms a traditional target point partition into a bin packing one. A point is chosen to be packed only if AAV has enough power to visit it and falls into its communication range. The objective of bin packing is to pack all the points with the minimum number of bins. Thus, the number of flights is minimized such that AAV can visit as many points as possible in one flight. Simulations illustrate that the proposed method significantly reduces extra charging and transportation needs than the state of the art.