Authors : Disha S. Wankhede; Rahul S. Ghodake; Prajakta S. Gaikwad; Ram V. Gavade; Sharvil V. Ghasad

Volume/Issue : Volume 11 - 2026, Issue 5 - May

Google Scholar : https://tinyurl.com/5wvn3wdc

Scribd : https://tinyurl.com/h7ajz3c4

DOI : https://doi.org/10.38124/ijisrt/26May1140

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Remote and disaster-prone sites demand a fast, dependable hazard detection method that does not rely on fragile ground infrastructure. Classical UAV systems usually depend on cloud processing or high-bandwidth video links with time lags and are inoperable without communication networks. In this paper we present an Edge AI technology for real-time disaster recognition and alert through an autonomous, sub-250g UAV system. The device is integrated with a Raspberry Pi and camera module to process an efficient object detection model-mobilenet onboard. The system shifts the computational load of raw video data to the ground and transmits compact- structured telemetry packets derived in real-time from locally processed visual data, including the class of object detected, confidence scores, as well as GPS coordinates via a low energy LoRa transceiver. With this method we significantly reduce the requirement for bandwidth and power while leaning on a communication range of over 500 meters. The experiments show that the system runs at on-line (real-time) performance in any typical small/medium building (8-12 fps@320x320), additionally indicate and locate fire and damage indicators. The suggested solution provides a first responders tool that is cheap and quickly deployable with or without internet.

Keywords : Disaster Management, Edge AI, LoRa Telemetry, MobileNet, Object Detection, Raspberry Pi, Unmanned Aerial Vehicle (UAV).

References :

1. S. Sharma and R. Kumar, "Edge-Based Autonomous Fire and Smoke Detection Using MobileNetV2," IEEE Access, vol. 9, pp. 11234-11245, 2021.
2. M. P. Manuel, M. Faied , and M. Krishnan, "A LoRa-Based Disaster Management System for Search and Rescue," IEEE Internet of Things Journal, vol. 11, no. 4, 2024.
3. A. Al-Kaff et al., "Deep learning-based autonomous fire detection system," Journal of Real-Time Image Processing, vol. 16, no. 5, pp. 1575-1588, 2019.
4. Y. Chen and L. Zhang, “Lightweight Neural Networks for Embedded Robotic Vision,” IEEE Embedded Systems Letters, vol. 12, no. 3, pp. 85–89, 2020.
5. T. Nguyen, D. Pham, and K. Do, “Real-Time Object Detection on Raspberry Pi Using Optimized CNN Models,” Procedia Computer Science, vol. 190, pp. 617–624, 2021.
6. F. H. C. Santos Filho, A. Silva and R. S. F. Filho, “Performance of LoRaWAN for Handling Telemetry and Alarm Messages,” Sensors, vol. 20, no. 11, p. 3061, 2020.
7. W. D. Paredes, A. Bonatti and M. A. Labrador, “LoRa Technology in Flying Ad Hoc Networks: A Systematic Review and Technical Overview,” IEEE Access / PubMed Central, 2023.
8. F. Mavilia, F. Girolami and A. Rota, “An experimental dataset using UAVs and LoRa technology,” Tech. Rep., 2025.
9. S. Boddu and A. Mukherjee, “Efficient Edge Deployment of Quantized YOLOv4-Tiny for Aerial Emergency Object Detection on Raspberry Pi 5,” arXiv preprint, Jun. 2025.
10. J. Kuzmic, “Object Detection with TensorFlow on Hardware with Limited Resources,” SCITEPRESS - Proceedings, 2023.
11. N. J. H. Marcano, “An experimental study of 2.4 GHz LoRa path loss for air-to-ground links,” Tech. Rep., 2024.
12. Kulkarni, R., Manikandaprabhu, P., Wankhede, D.S., Kumar, B.V., Vasuki, M. and       Rasmi, A. (2027). AI-Driven Optimization of Renewable Energy Systems. In Artificial Intelligence and Biodiversity (eds U.K. Lilhore, S. Simaiya, S. Dalal and M. Margala). https://doi.org/10.1002/9781394384990.ch7
13. Kale, P. D., Patil, S. B., Tulaskar, D. P., Deshmukh, M. T., Wankhede, D. S., Tawani, S. S., & Shahade, A. K. (Corresponding author). (2025). Design of fractal-inspired quadband microstrip antenna for multi-standard wireless applications. International Review on Computers and Software, 15(3), 163–173. https://doi.org/10.15866/irecap.v15i3.26274
14. T. Dhavale, S. Kotalwar, S. Phase, S. Shirsat, V. K. Kolekar and D. S. Wankhede, "Secure Web-based Ride-Sharing Platform using Route Matching," 2026 6th International Conference on Expert Clouds and Applications (ICOECA), Bengaluru, India, 2026, pp. 827-834, doi: 10.1109/ICOECA68095.2026.11485341.
15. Gaikwad, V. S., Sable, N. P., Wankhede, D. S., Mishra, V., Karnik, M. P., Ambhore, N., & Manikjade, A. (2023) Securing cloud-based IoT: Exploring the significance of lightweight cryptography for enhanced security. In Internet of Things enabled Machine Learning for Biomedical Applications (pp. 273-294). CRC Press.