Geofenced Unmanned Aerial Robotic Defender for Deer Detection and Deterrence (GUARD)

TL;DR

The paper tackles deer induced crop damage by proposing an integrated UAV system for autonomous wildlife deterrence in precision agriculture. It combines a YOLOv5 deer detection stack on a Jetson Orin Nano, energy efficient coverage path planning governed by Ant Colony Optimization, and an autonomous charging dock, all coordinated by a reinforcement learning supervisor trained in a photorealistic AirSim environment. Key contributions include a cohesive hardware and software integration tailored to farm constraints, an energy aware path planning framework with formalized cost metrics, and an autonomous charging solution enabling extended dusk-to-dawn operation. Field- and simulation-based results demonstrate robust detection (≈92% in abstracted claims), improved coverage efficiency relative to baselines, and the viability of RL driven coordination for continuous deer deterrence, signaling a scalable path toward real-world deployment and extension to other wildlife management tasks.

Abstract

Wildlife-induced crop damage, particularly from deer, threatens agricultural productivity. Traditional deterrence methods often fall short in scalability, responsiveness, and adaptability to diverse farmland environments. This paper presents an integrated unmanned aerial vehicle (UAV) system designed for autonomous wildlife deterrence, developed as part of the Farm Robotics Challenge. Our system combines a YOLO-based real-time computer vision module for deer detection, an energy-efficient coverage path planning algorithm for efficient field monitoring, and an autonomous charging station for continuous operation of the UAV. In collaboration with a local Minnesota farmer, the system is tailored to address practical constraints such as terrain, infrastructure limitations, and animal behavior. The solution is evaluated through a combination of simulation and field testing, demonstrating robust detection accuracy, efficient coverage, and extended operational time. The results highlight the feasibility and effectiveness of drone-based wildlife deterrence in precision agriculture, offering a scalable framework for future deployment and extension.

Figures (11)

• Figure 1: High-level system architecture showing integration of computer vision, path planning, and autonomous charging modules.
• Figure 2: Ground-based autonomous charging dock. (a) Inter-locking $60^{\circ}$ V-rails provide passive self-centering; (b) spring-loaded pogo pins supply power and balance leads;
• Figure 3: Illustration of our methodology for energy-efficient deer search.
• Figure 4: Birds-eye view of the farm our solution is designed for. Within the perimeter (orange rectangle) of the farm, there are obstacles such as tall trees (blue circles), a house and a greenhouse (blue rectangles). Orange stars represent the placement of the UAV charging stations.
• Figure 5: Illustration of the parameters that contribute to energy cost estimation for a drone (blue triangle) to go from waypoint (orange circle) i to waypoint j, given that it was previously on waypoint h. $\theta_{hij}$ is the amount that the drone must turn, and $d_{ij}$ is the linear distance it must travel.