A biologically-inspired threat monitoring system.
STARTLE detects anomalous or threatening conditions by emulating the mammalian conditioned-fear response mechanism. It provides a mechanism of cued-attention which has significant performance and efficiency benefits over conventional techniques.
STARTLE for Maritime Air Defence
Integrated into the Defence Science and Technology Laboratory (Dstl) sponsored Open Architecture Combat System (OACS), STARTLE for Maritime is being developed to help the command team assess tracks in increasingly complex and dynamic mission environments. In essence, STARTLE is a technology that can help humans deal with complexity better. It is a step further in maritime air defence and allows machines and humans to work together as a team.
STARTLE for Autonomous Vehicles
STARTLE can provide enhanced situational awareness and an early threat warning to both an autonomous vehicle and its remote operator. Making use of existing hardware, STARTLE intelligently processes information from multiple on-board sensors, cueing systems to assess and confirm potential threats to the vehicle.
STARTLE for Health and Usage Monitoring Systems (HUMS)
The threat detection methods in STARTLE also make it ideal for the detection of anomalous subsystem conditions within a complex system; for example in internal system status monitoring of manned or autonomous platforms.
STARTLE for Computer Network Defence
STARTLE can also be used for the detection of sophisticated computer network threats. It has been demonstrated successfully in detecting penetration test activity in real-world computer networks, and can use intercept-derived data from local and cloud-based storage.
- Autonomously detects and characterises threats
- Automatically directs sensor assets
- Processing power freed up for other tasks
- Reduces operator workload
- Neural network – gives high performance,lightweight monitoring and alerting
- Rule system - compiled by domain experts and allows the system to request the most appropriate data to confirm potential threats. Cued sensor re-tasking obtains additional data to be collected from a particular sensor / data store / processing algorithm and gives traceability which allows reasoning to be validated
- Neural Network training can be based on real or synthetic environment derived data, allowing a wide range of potential operational scenarios to be investigated
- Learning – potential to accommodate on-the-job learning in future systems
For more information on how your systems can benefit from this monitoring system please contact us.