|
Our technology capabilities provide smart decision-making for the military and civilian intelligent systems of the future.
Our research spans the range from short feasibility studies, through concept development, standardisation and prototyping, to implementation and deployment support.
| Communications |
There is a growing requirement for individual systems and devices - in the home, in industry, in the defence domain - to be equipped with greater embedded intelligence. Such devices will need to communicate with others for full user benefits to be realised. We provide communications consultancy for networks of autonomous devices and systems. Our services include architectural analysis, protocol analysis, and modelling and simulation, backed up with strong autonomous system domain knowledge. Our world-class commercial communications domain knowledge offers significant cross-domain knowledge transfer, with particular insights and benefits to our defence domain customers.
|
|
Intelligent Networks
|
Smart algorithms from autonomous systems and communications networks research enable robust, self-managing networks to be developed. Our position at the cutting edge of autonomous systems and networks research allows us to provide innovative solutions into one domain based on research undertaken in seemingly unrelated domains. |
|
Vehicle Route Planning
|
A key challenge for an autonomous vehicle system is the requirement to plan in dynamic and unstructured domains. We have developed architectures and algorithms for modelling the world and for dynamic reasoning at multiple levels to generate and maintain robust route plans. Our AI algorithms are naturally domain independent, and may be tailored to optimise for domain-specific criteria such as energy efficiency. |
|
Anomaly Classification
|
Real-world systems monitor their own health and their environment as part of typical operation. In many cases the amount of monitoring information to process is overwhelmingly great. We have developed a lightweight bio-inspired anomaly classification architecture, based on the mammalian limbic system. Our architecture can be tailored to introspective health monitoring, or for external threat monitoring. |
|
Power & Energy Management
|
Power and energy management is a significant engineering challenge for the present and the future. Our power management research draws from a broad basis, ranging from autonomous own-vehicle power management to network resource management algorithms. We are able to offer concept analysis and development for power management at many consumer scales, both in the industrial and defence domains. |
| Virtual Network Testbed |
The Virtual Network Testbed is a flexible research testbed which is used to evaluate technologies on a controlled but representative network topology. The testbed provides a rapid and easily deployed test environment for assessing robustness, scalability and performance of protocols and algorithms across decoupled network entities. Current research hosted on the testbed includes Future Communications Architectures for the UK MoD. |
| Network Modelling Support Environment |
The Network Modelling Support Environment - NMSE - is a Distributed System Modelling Toolkit that provides a strong rapid-prototyping framework for behaviourally rich applications. NMSE offers a powerful platform for both simulation work and for rapid development of concept demonstrations. NMSE-based applications have been used by NATS, Eurocontrol, MoD and other government agencies. |
| Autonomous Soaring |
The Autonomous Soaring concept describes energy-efficient route planning for autonomous airborne vehicles. An AI planning engine onboard a glider produces route plans to be followed, based on planned and serendipitous energy harvesting from the surrounding atmosphere, without compromising overall mission objectives. Autonomous Soaring has been successfully shown in field trials and software demonstrations. The Autonomous Soaring concept is equally applicable to ground-domain route planning. |
| STARTLE |
STARTLE is a cued threat detection system which provides local protection of autonomous vehicles.
The system exploits existing sensor fits and piggybacks on existing data streams to provide early warning of potential threats to own vehicle. STARTLE combines two AI techniques - a lightweight classifier and a goal prover - to direct sensor assets and select processing algorithms to efficiently assess possible threats. The system thus allows for more efficient and more focused use of available processing power. STARTLE can be applied to autonomous external threat monitoring and internal status monitoring for manned and unmanned systems.
|
|
