Autonomous Navigation of 4WIS4WID Agricultural Field Mobile Robot using Deep Reinforcement Learning

TL;DR

This work addresses autonomous navigation of a four-wheel independent steering and driving (4WIS4WID) agricultural field robot under Agriculture 4.0 conditions. It presents a unified, end-to-end DRL framework that couples a parametrized 4WIS4WID model with camera-based crop-row detection and waypoint-guided navigation to traverse multiple crop rows. The study compares continuous-action DRL algorithms and finds PPO to perform best, achieving high success rates and generalization to unseen crops, while outperforming a PD controller on a skid-steered baseline in a C-shaped path. The findings demonstrate that integrating 4WIS4WID kinematics with DRL can yield efficient, adaptable field navigation with potential sim-to-real transfer and reduced reliance on explicit map-based planning.

Abstract

In the futuristic agricultural fields compatible with Agriculture 4.0, robots are envisaged to navigate through crops to perform functions like pesticide spraying and fruit harvesting, which are complex tasks due to factors such as non-geometric internal obstacles, space constraints, and outdoor conditions. In this paper, we attempt to employ Deep Reinforcement Learning (DRL) to solve the problem of 4WIS4WID mobile robot navigation in a structured, automated agricultural field. This paper consists of three sections: parameterization of four-wheel steering configurations, crop row tracking using DRL, and autonomous navigation of 4WIS4WID mobile robot using DRL through multiple crop rows. We show how to parametrize various configurations of four-wheel steering to two variables. This includes symmetric four-wheel steering, zero-turn, and an additional steering configuration that allows the 4WIS4WID mobile robot to move laterally. Using DRL, we also followed an irregularly shaped crop row with symmetric four-wheel steering. In the multiple crop row simulation environment, with the help of waypoints, we effectively performed point-to-point navigation. Finally, a comparative analysis of various DRL algorithms that use continuous actions was carried out.

Figures (13)

• Figure 1: 3D model of 4WIS4WID field mobile robot used in simulation.
• Figure 2: Configuration of Symmetric 4WS.
• Figure 3: Different components in the simulation setup.
• Figure 4: Randomized start and goal locations at the start of each episode. A solid red circle represents the goal location. White guidelines in the robot model represent where the two cameras are directed.
• Figure 5: Software Workflow