AI-Powered Solutions for detecting Chemical Threats via Drones

Kongsberg Geospatial, in collaboration with SFL Scientific, is set to unveil an innovative AI-based system designed to enable autonomous Unmanned Aerial Vehicles (UAVs) to detect and identify various Chemical, Biological, Radiological, Nuclear, and Explosive (CBRNE) threats from aerial perspectives.

Live Demonstration of Cutting-Edge Technology

The two organizations will showcase footage from live trials during an online seminar organized by the Association for Unmanned Vehicle Systems International (AUVSI). This event is scheduled for Wednesday, November 10th, 2021, at 3:00 PM EDT. Participants can register for the seminar to witness the technology in action.

Real-time Detection of Chemical Hazards

During the presentation, attendees will see how a medium-sized commercial drone, equipped with a specialized sensor package, can autonomously detect chemical threats. This system is capable of identifying and visualizing invisible hazardous plumes in real-time.

Advanced Sensor Fusion and AI Processing

The technology integrates data from a visual, thermal, and multichannel chemical sensor (the FLIR MUVE C360) with environmental and locational information, processed by a dedicated AI unit (the NVIDIA Jetson). Utilizing advanced Generative Adversarial Networks (GANs) and Graph Deep Learning models, the system autonomously identifies potential threats.

Comprehensive Threat Management

Additionally, the system incorporates subsystems for route planning, enabling it to effectively address any detected threats. The synthesized sensor data, including threat identification and suggested routes, will be presented within Kongsberg Geospatial’s IRIS UxS ground Control Station (GCS). This airspace situational awareness platform is designed to equip drone operators with the necessary insights to safely manage multiple drones Beyond Visual Line-of-Sight (BVLOS).

Enhancing Autonomous Operations

The collaboration aims to illustrate how AI and machine learning can evolve to empower drones in executing complex, mission-critical operations with greater autonomy. “Swift identification of chemical and visual threats is vital for various civilian and federal operations,” stated Michael Segala, PhD, CEO of SFL Scientific. “Future devices will be equipped to autonomously identify, locate, and prioritize decisions regarding threats and unusual activities, enhancing the safety and efficiency of first responders in potentially dangerous environments.”

insights from Industry leaders

dr. Segala will elaborate on how SFL Scientific’s groundbreaking work utilizes next-generation AI technology to analyze raw sensor data in real-time. Meanwhile, Rex Hayes, Director of Unmanned Systems at Kongsberg Geospatial, will discuss how the IRIS system integrates sensor data and AI interpretations, including route planning, into a unified operational picture, thereby alleviating the cognitive burden on UAS operators.

“IRIS enables UAS operators to maintain a cohesive operational overview of the airspace while minimizing the mental strain associated with interpreting diverse sensor data. By harnessing AI to enhance system autonomy, operators can concentrate on their missions and improve overall effectiveness,” remarked Mr. Hayes.