RISE: Rolling-Inspired Scheduling for Emergency Tasks by Heterogeneous UAVs

Abstract

The multiple unmanned aerial vehicles (UAVs) system is highly sought after in the fields of emergency rescue and intelligent transportation because of its strong perception and extensive coverage. Formulating a reasonable task scheduling scheme is essential to raising the task execution efficiency of the system. However, the dynamics of task arrival and the heterogeneity of UAV performance make it more difficult for multiple UAVs to complete the tasks. To address these issues, this paper focuses on the multi-UAV scheduling problem and proposes a method of rolling-inspired scheduling for emergency tasks by heterogeneous UAVs (RISE). In order to ensure that emergency tasks can be allocated to UAVs in a real-time manner, a task grouping strategy based on a density peaks (DP) clustering algorithm is designed, which can quickly select UAVs with matching performance for the tasks arriving at the system. Furthermore, an optimization model with multiple constraints is constructed, which takes the task profit and UAV flight cost as the objective function. Next, we devise a rolling-based optimization mechanism to ensure that the tasks with shorter deadlines are executed first while maximizing the objective function, so as to obtain the optimal task execution order for each UAV. We conduct several groups of simulation experiments, and extensive experimental results illustrate that the number of tasks successfully scheduled and the utilization rate of UAVs by RISE are superior to other comparison methods, and it also has the fastest running time. It further proves that RISE has the capability to improve the completion rate of emergency tasks and reduce the flight cost of multiple UAVs.

Keywords

• multi-UAV scheduling
• emergency tasks
• rolling-based optimization mechanism
• task profit
• UAV flight cost