Multi-Robot Systems for Precision Agriculture

Biologically complex polyculture environments require multi-robot systems capable of monitoring and intervening under dynamic field conditions. Developed within the multi-institutional Robotic Intercropping initiative, this work reviews the relevant literature and proposes a conceptual framework for designing multi-robot systems for sustainable precision agriculture. The framework addresses how complementary robotic platforms can be integrated into field-robotics workflows, where system-level capabilities emerge from interactions between agents.

The review identifies three recurring levels of inter-agent interaction. Cooperation, where agents pursue a shared objective while acting independently, appears to be the most common. Coordination, involving state exchange and task management to improve efficiency and avoid conflicts, is less frequently implemented. Collaboration, where joint outcomes are enabled by complementary capabilities, remains comparatively rare. These interaction levels are complemented by a distinct but related dimension of system design, characterized along three dimensions: environmental complexity, mission complexity, and independence from human input.

This classification provides more than a descriptive overview: the relationship between interaction level, autonomy, and system architecture establishes a structured basis for multi-robot system design. We propose a layered framework that maps these dimensions to architectural choices and guides the deployment of UAV–UAV and UAV–UGV systems for agricultural monitoring and intervention. The framework follows a hybrid control paradigm, combining high-level human supervision with autonomous multi-robot operation. Human operators define objectives, priorities, and safety constraints, while robotic teams perform task allocation, local decision-making, and mission replanning through hierarchical control when tighter coupling is required.

A representative workflow illustrates this approach: an operator specifies a scouting objective, a UAV team performs multi-robot coverage planning to identify drought stress or pest presence, and UGVs subsequently execute targeted interventions. Informed by current field deployments and extended toward robotic intercropping scenarios, the framework supports the integration of aerial and ground platforms into end-to-end agricultural robotics pipelines spanning sensing, analysis, and intervention.