# Multi-Arm Robot Task Planning for Fruit Harvesting Using Multi-Agent Reinforcement Learning

## TL;DR

This paper presents a multi-agent reinforcement learning approach within a Markov game framework to optimize task planning for a four-arm harvesting robot, improving efficiency in orchard fruit harvesting.

## Contribution

It introduces a novel MARL-based task planning strategy that accounts for arm conflicts and dynamic tasks, reducing computational complexity compared to traditional scheduling methods.

## Key findings

- MARL-based planning outperforms existing methods in simulations
- The approach effectively manages arm conflicts and dynamic orchard conditions
- Experimental results show improved harvesting efficiency

## Abstract

The emergence of harvesting robotics offers a promising solution to the issue of limited agricultural labor resources and the increasing demand for fruits. Despite notable advancements in the field of harvesting robotics, the utilization of such technology in orchards is still limited. The key challenge is to improve operational efficiency. Taking into account inner-arm conflicts, couplings of DoFs, and dynamic tasks, we propose a task planning strategy for a harvesting robot with four arms in this paper. The proposed method employs a Markov game framework to formulate the four-arm robotic harvesting task, which avoids the computational complexity of solving an NP-hard scheduling problem. Furthermore, a multi-agent reinforcement learning (MARL) structure with a fully centralized collaboration protocol is used to train a MARL-based task planning network. Several simulations and orchard experiments are conducted to validate the effectiveness of the proposed method for a multi-arm harvesting robot in comparison with the existing method.