Chithra et al. (2025) Collaborative Human in the Loop Robotics Framework for Precision Agriculture and Crop Monitoring

Identification

• Journal: International Academic Journal of Science and Engineering
• Year: 2025
• Date: 2025-12-30
• Authors: Dr.R. Chithra, Dr.M. Raja
• DOI: 10.71086/iajse/v12i4/iajse1287

Research Groups

• Department of Information Technology, K.S. Rangasamy College of Technology, Tiruchengode, India
• Department of Information Science and Engineering, MVJ College of Engineering, Bangalore, India

Short Summary

This paper introduces a Collaborative Human-in-the-Loop (HITL) Robotics Framework for precision agriculture and crop monitoring, integrating human decision-making with robotic efficiency to overcome limitations of conventional autonomous systems. Experimental results demonstrate the HITL framework's superior performance in crop yield prediction (92%), pest detection (89%), and water use efficiency (94%), with minimal human intervention.

Objective

• To develop and evaluate a Collaborative Human-in-the-Loop (HITL) Robotics Framework for precision agriculture and crop monitoring.
• To demonstrate how integrating human decision-making with autonomous robotics can enhance flexibility, precision, and adaptability in dynamic agricultural conditions, outperforming conventional autonomous systems.

Study Configuration

• Spatial Scale: A 10-hectare (100,000 square meter) farm in rural India.
• Temporal Scale: Real-time, continuous monitoring and decision-making.

Methodology and Data

• Models used: XGBoost (for predictive analytics and regression like crop yield prediction), Convolutional Neural Networks (CNNs) (for image-related tasks like crop health and pest detection), Support Vector Machines (SVMs) (for classification tasks like pest-infested region segmentation), and Reinforcement Learning (RL) (for decision-making problems like irrigation optimization).
• Data sources:
  • Autonomous robots equipped with FLIR Systems Vue TZ20 RGB Cameras, Ouster OS1-64 Thermal Imaging Cameras, and Decagon Devices 5TE Soil Moisture Sensors.
  • Custom field data from a 10-hectare farm, including soil moisture (0-100 %), temperature (20-45 degrees Celsius), humidity (50-90%), soil type (loamy, clay), and crop type (wheat, rice, maize).
  • Real-time data acquisition controlled with ROS (Robot Operating System).
  • Human operator input via an intuitive control interface.

Main Results

• The HITL framework achieved a crop yield prediction accuracy of 92%, a pest detection rate of 89%, and a water use efficiency of 94%.
• Human intervention was required for only 10% of tasks, demonstrating a low frequency of intervention, with an effectiveness of 80%.
• Compared to traditional autonomous robotics systems, the HITL framework showed significant improvements:
  • Crop Yield Prediction Accuracy: 92% (HITL) vs 85% (Traditional)
  • Pest Detection Rate: 89% (HITL) vs 75% (Traditional)
  • Water Usage Efficiency: 94% (HITL) vs 82% (Traditional)
  • Operator Intervention Frequency: 10% of tasks (HITL) vs 30% of tasks (Traditional)
  • Operator Intervention Effectiveness: 80% (HITL) vs 60% (Traditional)

Contributions

• Proposes a novel Collaborative Human-in-the-Loop (HITL) framework that integrates human expertise with autonomous robotics for decision-making in precision agriculture under uncertainty.
• Significantly improves crop health monitoring, pest detection, and irrigation management, making them more precise and efficient than traditional autonomous systems.
• Demonstrates superior performance in key agricultural metrics (crop yield prediction, pest detection, water usage efficiency) while requiring less frequent, but highly effective, human operator control.
• Enhances the flexibility, adaptability, and responsiveness of agricultural systems to dynamic and unpredictable environmental conditions.

Funding

Not specified in the paper.

Citation

@article{Chithra2025Collaborative,
  author = {Chithra, Dr.R. and Raja, Dr.M.},
  title = {Collaborative Human in the Loop Robotics Framework for Precision Agriculture and Crop Monitoring},
  journal = {International Academic Journal of Science and Engineering},
  year = {2025},
  doi = {10.71086/iajse/v12i4/iajse1287},
  url = {https://doi.org/10.71086/iajse/v12i4/iajse1287}
}