AG-CVG: Coverage Planning with a Mobile Recharging UGV and an Energy-Constrained UAV

TL;DR

The paper addresses coverage path planning for a team consisting of an energy-constrained UAV and a mobile recharging UGV. It introduces the AG-Cvg heuristic, which first generates baseline coverage paths, then clusters path segments and applies graph matching to schedule recharge rendezvous, ensuring UAV endurance while achieving complete area coverage. Empirical results show an average rendezvous overhead reduction of 11.33% (up to 25% in some cases) over a greedy baseline, validated through simulations and field deployment with a VOXL m500 UAV and a Clearpath Jackal UGV. This work provides a practical offline-to-field pipeline that enables energy-aware, cooperative coverage in real-world scenarios.

Abstract

In this paper, we present an approach for coverage path planning for a team of an energy-constrained Unmanned Aerial Vehicle (UAV) and an Unmanned Ground Vehicle (UGV). Both the UAV and the UGV have predefined areas that they have to cover. The goal is to perform complete coverage by both robots while minimizing the coverage time. The UGV can also serve as a mobile recharging station. The UAV and UGV need to occasionally rendezvous for recharging. We propose a heuristic method to address this NP-Hard planning problem. Our approach involves initially determining coverage paths without factoring in energy constraints. Subsequently, we cluster segments of these paths and employ graph matching to assign UAV clusters to UGV clusters for efficient recharging management. We perform numerical analysis on real-world coverage applications and show that compared with a greedy approach our method reduces rendezvous overhead on average by 11.33%. We demonstrate proof-of-concept with a team of a VOXL m500 drone and a Clearpath Jackal ground vehicle, providing a complete system from the offline algorithm to the field execution.

Figures (6)

• Figure 1: Sample allocation of task regions: UGV surveying a large terrain with obstacles and UAV a lake surface.
• Figure 2: Complete Pipeline of AG-Cvg algorithm (see Algorithm \ref{['algorithm:ag_cover']}). It consists of four main steps: (A) finding the coverage paths; (B) finding the worst-case rendezvous cost; (C) clustering paths into regions that can be covered completely with a UAV without running out of charge; (D) finding the rendezvous locations.
• Figure 3: Samples of different environments and different scenarios of monitoring operations required from each robot. Light blue regions denote the regions for vehicles that can serve as mobile recharging stations (UGV, ASV), and the light orange one for vehicles with limited energy capacity (UAV). The Greedy method takes the returns back to the same location as it has started the rendezvous, whereas AG-Cvg due to perfect matching might often return to a different cluster to perform the coverage --- minimizing total rendezvous cost per survey mission.
• Figure 4: The rendezvous overhead is the same on the F3 site, whereas it is lower on other scenarios with 11.33$\%$ improvement on average, in some instances reaching up to 25$\%$ improvement.
• Figure 5: Proof of concept field trials with VOXL m500 UAV and Clearpath Jackal UGV at Fearless Flight facility (F3) at University of Maryland, College Park, MD, USA. (a) The vehicles are launched from a predefined location with $1m$ offset between their origins. (b) UAV and UGV are performing coverage tasks until they reach the rendezvous location. (c) Before rendezvous each of the vehicles waits for the other one to arrive at the location, UAV is at the flight altitude $3m$. (d) If both arrive UAV will rendezvous by descending to $1m$ altitude.