Design and Performance Evaluation of an Edge-AI Wildlife Monitoring System for Human-Wildlife Conflict Mitigation

Abstract

Human–wildlife conflict has become a critical issue in rural regions where human settlements overlap with natural habitats. This paper presents the design and performance evaluation of a low-cost Edge-AI based wildlife monitoring system aimed at reducing such conflicts. The proposed system integrates motion sensing, computer vision, and embedded processing to enable real-time detection of wildlife intrusion. Upon detection, the system triggers alerts and deterrent mechanisms to prevent close encounters. Experimental evaluation demonstrates that the system achieves high detection accuracy with low latency while maintaining energy efficiency suitable for solar-powered deployment. The results indicate that the proposed approach is a viable and scalable solution for rural environments.

Introduction

Human–wildlife conflict is increasing due to population growth, agricultural expansion, and habitat fragmentation, leading to more encounters between humans and wild animals such as tigers and leopards. These interactions often cause injuries, fatalities, and economic losses. Traditional methods like fencing and manual patrolling are reactive and inadequate. To address this issue, the paper proposes a low-cost Edge-AI wildlife monitoring system for rural areas that enables real-time monitoring and alerts.

The literature review highlights that traditional monitoring methods are labor-intensive and lack continuous surveillance. Camera traps improved wildlife observation but do not provide real-time alerts. AI-based object detection models, particularly YOLO, have shown high accuracy in wildlife detection, while edge computing allows local processing with low latency. Existing AI-based systems such as TrailGuard AI and MARVEL offer real-time monitoring but are expensive, whereas IoT-based systems are affordable but often generate false alarms due to limited intelligence.

The identified research gap shows that traditional methods are inefficient, camera traps lack instant alerting, AI systems are costly, and IoT solutions lack intelligent classification.

The proposed system consists of seven components: a PIR motion sensor, camera module, Edge-AI processing unit, decision-making module, alert and deterrent system, power management unit (solar and battery), and a protective enclosure. The workflow begins with motion detection, followed by image capture, AI-based animal detection, evaluation of results, and generation of alerts when necessary.

The methodology includes motion-triggered image acquisition, image preprocessing, CNN-based object detection on an edge device, confidence-based decision logic with multi-frame validation to reduce false positives, and an alert mechanism that activates sirens, LED lights, and user notifications upon wildlife detection. This approach aims to provide an affordable, intelligent, and real-time solution for reducing human–wildlife conflicts in rural areas.

Conclusion

This paper presents a low-cost Edge-AI wildlife monitoring system capable of real-time detection and alert generation. The system demonstrates that affordable embedded AI solutions can effectively mitigate human–wildlife conflict.

References

[1] Government of Maharashtra, “MARVEL: Maharashtra Advanced Research and Vigilance for Enhanced Law Enforcement,” Technical Report, 2024. [2] J. S. Dertien, E. J. Milner-Gulland, et al., “Mitigating human–wildlife conflict using real-time AI camera systems,” BioScience, vol. 73, no. 10, pp. 748–760, 2023. [3] A. Kumar, R. Mehta, and P. Shah, “Real-time wildlife detection using YOLO and embedded systems,” Procedia Computer Science, vol. 218, pp. 195–204, 2023. [4] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “YOLOv4: Optimal speed and accuracy of object detection,” arXiv preprint arXiv:2004.10934, 2020. [5] A. S. Kshirsagar and S. B. Patil, “IoT-based animal intrusion detection system for agriculture,” International Journal of Scientific Research in Science and Technology, vol. 7, no. 3, pp. 89–94, 2020. [6] J. Norouzzadeh, A. Nguyen, M. Kosmala, A. Swanson, M. Palmer, C. Packer, and J. Clune, “Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning,” Proceedings of the National Academy of Sciences (PNAS), vol. 115, no. 25, pp. E5716–E5725, 2018. [7] M. Satyanarayanan, “The emergence of edge computing,” Computer, vol. 50, no. 1, pp. 30–39, 2017.

Copyright

Copyright © 2026 Mayur Ramteke, Dr. Vijayshri Mahobiya, Swapnil Chaudhary. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.