Comparative Evaluation of VR-Enabled Robots and Human Operators for Targeted Disease Management in Vineyards

TL;DR

This paper evaluates immersive VR-controlled robots versus humans in vineyard disease management, showing VR robots excel in treatment and navigation tasks despite slower manual scanning, highlighting VR's potential in precision agriculture.

Contribution

It introduces a VR-based control interface for agricultural robots and demonstrates its advantages in treatment and navigation efficiency over human operators.

Findings

VR robots complete treatment tasks 65% faster than humans.

Immersive VR robots are 38% faster in yield-map navigation.

Manual scanning remains slower for VR robots compared to humans.

Abstract

This study explores the use of immersive virtual reality (VR) as a control interface for agricultural robots in vineyard disease detection and treatment. Using a Unity-ROS simulation, it compares three agents: a human operator, an immersive VR-controlled robot, and a non-immersive VR-controlled robot. During the scanning phase, humans perform best due to agility and control speed. However, in the treatment phase, immersive VR robots outperform others, completing tasks up to 65% faster by using stored infection data and optimized path planning. In yield-map-based navigation, immersive robots are also 38% faster than humans. Despite slower performance in manual scanning tasks, immersive VR excels in memory-guided, repetitive operations. The study highlights the role of interface design and path optimization, noting limitations in simulation fidelity and generalizability. It concludes that immersive VR has strong potential to enhance efficiency and precision in precision agriculture.