Sensor Counter Deception
Overview
DIS has developed a cutting-edge threat classification technology for uncrewed air systems (UAS), as part of a counter-deception work package from DASA. The “Sensor Counter Deception” capability can identify UAS threats based on their flight behaviour against a background of neutral activity and can self-learn as the threat evolves.
The Capability
This advanced machine learning algorithm was tested using a scenario-based concept demonstrator. Here, it was tasked to “defend” a sensitive site against threat objects that intended to breach the perimeter and reach the site. The goal was to correctly identify threat targets before they reached the site without raising false positives from the ongoing benign activity in the environment.
​
An adversary algorithm was developed to test and challenge the “defender”. Its aim was to generate flight paths which could successfully breach the sensitive area without detection, deceiving the defender. The opposing algorithms were tested against one another in an adversarial manner within the demonstration environment, each successfully learning from the behaviour of the other and adjusting itself to counter the other, in an iterative and evolving strategic interplay.
Use Cases
There is a huge opportunity for further research in this area and exploitation of the capability developed by DIS, with scope beyond the specific counter-deception and counter-UAS (CUAS) applications explored.
​
Key exploitation areas identified are the;
Detection of anomalies in airspace monitoring
Evaluation of CUAS sensor networks
Use of ML in the optimisation of complex modeling and simulation
Development of a broader deception and counter-deception AI sandbox
Get in touch
To find out more about Dynamic Intelligence Solutions, our expertise or our products, please get in touch by filling in the form or by sending us an email.