UAV-Integrated IoT System for Handling Widespread Events

Abstract

We introduce a mixed aerial-ground IoT system that monitors widespread events via a wireless sensor network (WSN), and reacts to them with assistance of multiple UAVs. While such an event is spreading out, more UAVs are dispatched in synchronization with its development. The dynamically formed team of UAVs autonomously share the workload of executing jobs in response to the arising event, taking into account the cost of handling. The cost is comprehensively dependent on processing latency, energy consumption, and data exchange. Inspired by the water-pouring idea, we design an iterative job redistribution strategy that progressively approaches to the scheduling optimality. At the same time, a bidding protocol is constructed to let each UAV periodically declare its goodwill to accept new jobs migrating from others. It therefore helps regularly re-adjust the workload imposed on individual UAVs, maintaining the optimality over time. To reduce the overall handling time, the strategy also directs a UAV finishing jobs earlier to be relocated for assisting other members of the team. Numerical results from a simulated system of 10,000 sensors and 37 UAVs deterministically confirm the outperformance of our designed strategy, with respect to energy, latency, fairness and mobilized UAVs.