Path planning and scheduling of multiple unmanned Ground vehicles for orchard plant protection operations

• A decision-making model is developed for UGV task allocation, routing, and schedulingIn-orchard supply bases and initial UGV assignment are optimized for operational efficiency. • Increasing UGV number and tank size enhances efficiency and reduces non-spraying energy loss. • Guidelines are offered for selecting UGV quantity, type, and pesticide load capacity. • The method improves UGV swarm efficiency and reduces energy waste in orchard operations. We propose a path planning and scheduling algorithm for multiple Unmanned Ground Vehicles (UGVs) engaged in pesticide application in orchards with multiple pesticide and power loading bases. Based on the structured operational environment of modern orchards, we develop a decision-making model for task generation, walking and spraying path generation, and scheduling of the UGV swarm, optimizing energy consumption and operational efficiency. Considering the movement of multiple UGVs between operation fields and supply bases, we establish a latency time calculation model to determine the total operation time for the UGV swarm. To alleviate swarm supply delay and reduce the operation end time difference between UGVs, we develop a PSO-based optimization strategy for the initial assignment of UGVs to different supply bases. The test results show that UGV swarm efficiency is strongly influenced by the number of UGVs, pesticide loading capacity, and supply time for orchard pest and disease control. Increasing the number of UGVs for orchard operations directly reduces total operation time, particularly when the number of UGVs increases from 1 to 3. Expanding UGV loading capacity reduces non-spraying walking energy waste by decreasing the number of operation tasks. However, deploying UGVs with lower loading capacity (and lower cost) can achieve similar efficiency to using the same number of UGVs with higher loading capacity (and higher cost). It is verified that the warm supply delay and operation end time differences between UGVs can be alleviated and reduced, respectively, by using the proposed algorithm. Compared to other methods, the total operation time could be reduced by 50 min (approximately 200 min in total) using the proposed method. Simultaneously, significant reductions in non-spraying walking energy waste in the UGV swarm can be achieved in many cases.