Collaborative path planning and task allocation for multiple mowing robots in the standard orchards

Path planning and task allocation are the key technologies of multi-machine collaboration. Current approaches focus on field operations, but actually orchard operations are also a promising area. In order to improve the efficiency of orchard mowing, a cooperative operation scheduling method was proposed for multiple mowing robots in the dwarf dense planting orchards. It aims to optimize the non-working time of the robot in the intra-plot paths and inter-plot routes. Firstly, a genetic algorithm with multi-mutation and improved circle algorithm (MC-GA) was proposed for path planning. Subsequently, an ant colony optimization algorithm with mixed operator (Mix-ACO) was proposed for task allocation. With regard to the shortage of robots, a local search algorithm was designed to reassign work routes. Simulation experiment results show that MC-GA can significantly reduce the total turning time and the number of reverses for the robot. Mix-ACO can effectively allocate tasks by generating multiple work routes and reduce the total transfer time for the robot fleet. When the number of work routes exceeds the number of mowing robots, the local search algorithm can reasonably reallocate multiple routes to robots, reducing the difference in task completion time of the robot fleet. Field experiment results indicate that compared with the reciprocating method, SADG, and GA, MC-GA can reduce fuel consumption rate by 1.55%-8.69% and operation time by 84-776 s. Compared with ACO, Mix-ACO can reduce the total transfer time by 130 s. The research results provide a more reasonable scheduling method for the cooperative operation of multiple mowing robots.