Abstract
High-density orchard (HDO) systems, while maximizing yield per unit area, present significant challenges for traditional weed management due to narrow row spacing and the sensitivity of young trees to chemical drift. This paper proposes a novel framework for coordinated swarm robotics designed for targeted herbicide application within these complex environments. Traditional methods often rely on broadcast spraying, which can negatively impact the growth of young trees and alter weed composition over time. By leveraging swarm intelligence and coordinated path planning, we present a system where multiple autonomous agents collaborate to identify and treat weeds with high spatial precision. Our methodology integrates multi-target search algorithms with artificial immune system-inspired coordination to ensure robust coverage and error detection. We utilize a decentralized communication architecture that allows the swarm to adapt to the specific structural constraints of super high-density olive and apple orchards. Simulation results demonstrate that the swarm approach significantly reduces herbicide volume compared to conventional methods while maintaining high efficacy. Furthermore, the system exhibits collective energy homeostasis and adaptive error detection, essential for long-duration field operations. This research contributes to the advancement of smart farming by providing a scalable, efficient, and environmentally sustainable solution for weed control in intensive horticultural systems. The findings suggest that coordinated robotic swarms can overcome the limitations of single-agent systems in HDO, paving the way for fully autonomous orchard management as part of the broader transition toward Digital Agriculture.