Dynamic Intelligence Solutions (DIS) are pleased to announce the award of a contract to explore novel Machine Learning techniques for counter-UAS threat detection.
The project, entitled “A self-learning heterogeneous sensor cluster for CUAS threat identification”, was awarded as part of the Defence Science and Technology Laboratory (Dstl) DASA competition “Autonomous sensor management and sensor counter deception” and will be delivered with OSL Technology as our supporting partner. Under this, DIS will develop a concept demonstration of a self-learning sensor cluster, capable of identifying hostile drone behaviour from a background of benign activity in order to protect a sensitive site.
This sensor cluster will be trained using adversarial reinforcement learning techniques to learn what threat behaviour looks like against an intelligent adversary, incorporating sensors of different modalities to provide resilience against deception. The open-source Stone Soup framework (https://github.com/dstl/Stone-Soup/) will be used as the testbed for this groundbreaking research.
The outcomes of the project will feed into the future of DIS’ Octa capability and its use in OSL’s FACE® platform to provide unparalleled situational awareness and airspace protection.