Introducing FarmGuard's Multi-Robot Deer Deterrence System

Developed by an award-winning team from Gini's Next Generation Robotics Lab at the University of Minnesota, FarmGuard was recognized for its innovative approach to reducing deer intrusion in agricultural fields. Our system combines hardware, computer vision, and path planning to effectively patrol and protect farmlands in collaboration with local farmers.

2026

2025

Awards

2025 Farm Robotics Challenge

🏆 Excellence in Small Farms Technology — $5,000 prize

Press release

Shepherding Algorithm Training Status

Training

The shepherding algorithm is still training and improving. Shepherding training metrics across 1.142 billion environmental steps are shown below. The vertical dotted lines indicate when we interrupted training to adjust parameters and reward weights before resuming where it left off.

Drone Behavior Team

Energy-Efficient Coverage & Deer Shepherding

We develop two algorithms that drive our drones: an energy-efficient coverage path planner that decides the order in which each drone visits its waypoints to minimize battery expenditure, and a Deep-Q-learning shepherding policy that repels invading deer from the protected area.

System architecture for FarmGuard's drone behavior pipeline

Learn more

Computer Vision Team

Automated Wildlife Detection

The Computer Vision Team designs and trains deep learning models to detect wildlife in real and simulated farmland images. Their YOLO-based detector achieves high accuracy across varying conditions, supporting our autonomous monitoring system.

Learn more

Hardware Design Team

UAV Platform Engineering

The Hardware Design Team engineers and assembles our UAV platforms, integrating sensors, cameras, and power systems. They ensure each drone meets rigorous performance and safety standards for field deployment.

Learn more

Path planning Team is looking at a computer

Farm Robotics Challenge

FarmGuard is our entry in the Farm Robotics Challenge (AIr category). We strive to showcase how multi-robot systems can effectively protect and enhance farming operations.

Learn More

Get Involved

Are you interested in advancing agricultural robotics? Join us in pushing the boundaries of AI, drone technology, and sustainable farming.

FarmGuard in Flight

Watch our UAVs autonomously patrol the fields.

Seamless Aerial Autonomy

Experience our UAV’s elegant patrol: gliding through fields, sensing and deterring wildlife with precision and grace.

Advisors & Collaborators

Guiding the project with expertise and on-the-ground insights.

Maria Gini

Professor, Computer Science & Engineering, University of Minnesota

Dr. Maria Gini, recipient of the Presidential Award for Excellence in Science, Mathematics, and Engineering Mentoring and a Fellow of both AAAI and ACM, researches multi-agent systems and swarm robotics. Her work spans decision-making for autonomous agents—from swarm coordination and distributed task allocation to exploration of unknown environments, navigation in dense crowds, and conversational agents.

As our faculty advisor on the deer-deterrence project, Dr. Gini guided every phase of development: she shaped our research design, reviewed simulation setups, refined field-trial protocols, and advised on multi-agent coordination and data analysis. She connected us with experts and resources at the AI Climent Institute and the Minnesota Robotics Institute, and her feedback on hardware deployment and reporting ensured both academic rigor and practical success.

Fresh Earth Farms

Chris & Susan James, Proprietors

For over two decades, Fresh Earth Farms has provided the Twin Cities with farm-fresh produce through a Community Supported Agriculture (CSA) model. Chris & Susan manage a 13-acre organic vegetable operation using cover crops, rotations, and sustainable practices to enrich the soil and deliver peak-flavor harvests to their members.