5G UAS Delivers AI/ML Enabled HazMat Classification for Public Safety

Open Generation 5G Consortium members, OST Global and MITRE Engenuity, partnered together to demonstrate a series of new aviation and public safety use cases using 5G-enabled UAS and AI/ML at the network edge. These capabilities included beyond visual line of sight (BVLOS) operation, remote site assessment, multi-access edge compute, video processing and image classification, command and control (C2) link and video streaming over 5G.

MITRE Engenuity contributed to the project by flying a 5G-enabled UAS in BVLOS to complete a 6-mile flight using 5G for C2 and video streaming. The 5G UAS descended on the remote end of the flight path and was maneuvered remotely to execute the hazard materials (HazMat) incident scenario at a simulated incident zone located at DHSES New York State Preparedness Training Center in Oriskany, NY. All communications links including command and control (C2), video streaming and sensor data, were transmitted over 5G link.

“This proof of concept has the potential to revolutionize how drones are utilized for public safety. The use of 5G for C2 allows reliable connectivity to extend beyond visual line of sight operations, while also counting on high throughput and low latency to support real time video streaming. The use of AI/ML at the network edge allows moving complex video processing from onboard the drone to a low-latency point of computing, close to the mission. By doing so, we provide valuable situational awareness for emergency response teams, local and remote, while reducing battery consumption requirements at the drone. Additionally, bringing the processing to the edge of the network or to a cloud-based location allows aggregating information from multiple drones, and enables the implementation of intelligent algorithms that tap on these multiple streams simultaneously, in future scenarios.” says Leila Ribeiro, Ph.D, Chief Technology Director, Open Generation 5G Consortium.

OST Global contributed to the project by building and running a machine learning (ML) model capable of detecting and classifying HazMat placards as defined in the USDOT 2020 Emergency Response Guidebook. Using machine learning, OST Global was able to conduct image recognition of HazMat placards involved in the incident to provide real-time intelligence for image classification capability at the network edge to identify the class or division of a material involved in an accident.

Our Hazmat placard detection AI/ML model equipped on a drone has the potential to assess incidents and perform scene analytics to significantly enhance the safety and well-being of communities. Use of this technology can enable early identification of Hazards, improve emergency response by responding faster and allocating appropriate resources for evacuation planning. By collecting and analyzing data quick containment and mitigation of Hazmat incidents can reduce the overall cleanup and remediation costs, benefiting both the community and responsible parties. More complex scene analytics experiments need to be performed to exploit the fusion of using UAS, AI models and 5G” says Renu Chaudhry, Chief Strategy Officer at OST Inc.

“The successful demonstration of aviation and HazMat use cases over 5G illustrates the value technology can provide to public safety. The tested use cases demonstrate not only the value of 5G, but also show how implementing AI/ML responsibly can deliver positive impact and outcomes,” said Thierry E. Klein, President of Nokia Bell Labs Solutions Research and founding member of the Open Generation Consortium. “The Open Generation Consortium provides an excellent forum for cross industry collaboration and we look forward to continuing to test enterprise and industrial use cases.”

The use of AI/ML at the network edge delivers low latency required for complex computing of real time video and HazMat image classification, while avoiding battery depletion. This project also proved 5G benefits of real time video-streaming while making mission data available to public safety officials.

Technical Documentation

To learn more about this capabilities read the Technical Report “MITRE ENGENUITY OPEN GENERATION EXPERIMENT: REAL-TIME HAZMAT SIGN CLASSIFICATION OVER 5G“. To learn more about how you can join this and other research activities with the Open Generation 5G Consortium visit us https://mitre-engenuity.org/5g/open-generation/ or email us at opengeneration@mitre-engenuity.org.

Technologies Used

This Open Generation 5G Consortium use case was achieved using a customized Aurelia X6 Pro UAS equipped with the MITRE Engenuity payload using NDAA compliant Telit 980m modem. The AI/ML at the network edge used Microsoft Azure for cloud services.

The Open Generation 5G SA Private Network is powered by Nokia and Verizon Wireless. Flight operations and coordination with FAA UAS Test Field was managed by NUAIR.


A blue and yellow logo featuring 5G collaboration on a black background.
Mitre egency a foundation for public good.

Join Us
The MITRE Engenuity Open Generation 5G Consortium is a privately funded research, development and innovation community. We bring together the wireless ecosystem to share insights, research, experimentation, and innovation to advance the use of 5G for public good. To learn more on how to become a member, validate NDAA-compliant 5G devices, share spectrum, fund open R&D projects or sponsor our 5G Innovation Summit. For more information regarding MITRE Engenuity Open Generation 5G Consortium visit https://mitre-engenuity.org/5g/open-generation/ or email us at opengeneration@mitre-engenuity.org