Our Approach

We love bringing ideas to life. We don’t follow a rigid process to do this, as we believe it stifles innovation and reduces the scope for surprise discoveries. However, our general approach ensures we maximise the value of our research and outputs.

We believe that iteration is key - Create, Try, Learn, repeat.

Each step explores the limits of technology, deepens understanding of the problem space and highlights implications for the broader system.

  • Discover

    We start with either a problem to solve, an approach that we can use in a novel way, or an idea that emerges from “what happens if…?”.

    TRL 1-2

  • Explore

    Next, we rapidly iterate the concept to find its strengths and limitations in the context of the problem space.

    TRL 3-4

  • Prototype

    Finally, we take what we’ve learned and apply it in more realistic situations, adjusting as we go. Real world experimentation is the acid test.

    TRL 5-6

Our Capabilities

Edge Intelligence

  • A system has to be able to make sense of its environment to operate effectively within it. We have deep expertise in processing sensor data in novel ways, including in real time on lightweight hardware. We’ve worked with a range of cameras (visible and IR), radars and lidars.

  • So much information is conveyed and stored in written or verbal forms. Whether it’s mining documents for insights or providing users with an intuitive natural language interface to a system, we can explore the possibilities.

  • Manual tasking of systems limits effectiveness and scalability. We have successfully demonstrated planning and task execution algorithms that enable systems to meet assigned goals with tailorable levels of human input.

  • So much of the AI field is focused on processing intensive software but what happens when you don’t have access to the cloud or you’re operating on a small UAS? We have successfully demonstrated cutting edge low power solutions to enable persistent intelligent sensors and small UAS.

Optimisation & Planning

  • The quantity of available data about our world is continually growing, from satellite imagery to high resolution lidar scans and terrain maps. Our ability to analyse this data is a prerequisite a broad range of capabilities that operate in the physical world. Whether you’re looking for objects of interest in vast areas or planning the routes of autonomous vehicles, we have valuable practical experience to offer.

  • Sat navs mean we take route finding for granted in our daily lives. But what happens when you’re trying to navigate off road in a rapidly changing environment such as an area devastated by a hurricane? We’ve been guiding a range of ground and air vehicles through complex environments for years, why not ask us for directions?

  • Co-ordinating multiple activities and disseminating instructions around a system can be difficult, particularly when the operating environment is rapidly changing. Algorithms can help users make better decisions and scale systems under their control. We gained vast practical experience through Dstl’s Autonomous Last Mile Resupply programme and have applied that knowledge to a variety of challenges and contexts since then.

  • In a world where we’re all being asked to do more with less, how do you maximise the effectiveness of your available assets? We apply geospatial analysis and performance modelling techniques to optimise asset deployment in line with your goals.

Autonomous Systems Engineering

  • Autonomous systems require much more than intelligent algorithms to be effective. How should information flow around the system? Should we choose a distributed or centralised model? How does the system meet non-functional requirements such as the need for transparency? Where is human authority required? Our practical experience of developing advanced autonomous systems and demonstrating them in real environments could be invaluable.

  • The realisation of your system architecture and design should be straightforward if you’ve followed rigorous systems engineering approaches, right? Unfortunately the power of autonomy comes with the capacity to surprise. We’ve led collaborative research projects focused on integration challenges and picked up plenty of battle scars from our development activities. From outlining approaches that manage risk and complexity, through to advising on simulation and hardware in the loop methods, we understand what works and what doesn’t.

  • It sometimes comes as a surprise to AI researchers that ROC curves are not sufficient evidence of performance to allow algorithms and systems out into the wild. Our background in regulated industries, time sitting on AI safety conference committees, leading AI assurance research projects and practical trials experience all mean we understand the assurance challenge. Whether there’s a need to assure performance, safety, ethics or something else, we’re likely to have some insight to offer.

  • Creating an autonomous system doesn’t mean that we can completely ignore the humans that need to interact with the system in some way. We’ve seen the power of autonomy working in tandem with humans and firmly believe that bringing together the complementary properties of human and machine can deliver on the promises of AI and autonomy. As such, human-machine teaming is an active research area for us.

Our Markets

Defence

Our team has nearly 30 years of experience working in the defence sector across all domains, including leading the delivery of advanced AI and autonomy technology demonstrator programmes.

  • Our team has been instrumental in the delivery of numerous Dstl and DASA technology demonstrator programmes over the last few years. We’ve gained a reputation for showcasing the art of the possible and delivering well beyond expectations.

  • We’ve integrated AI and autonomy-enabling technologies with a range of mature and experimental military equipment. This includes armoured vehicles, uncrewed ground and air vehicles, sensor payloads and communications systems.

  • Our team has worked with a range of weapon systems including missiles, directed energy weapons and supporting systems.

  • We have extensive experience of developing and integrating with command and control systems from domains including air defence, logistics and battlefield management.

  • Our team has experience across the ISTAR enterprise covering data acquisition, analysis, dissemination and action. We’ve worked with in service sensors and developed concepts for experimentation to inform future capabilities.

Security

Our team has worked with a range of security problems, from biometrics in Op. Herrick to AI optimised surveillance cameras. We look to apply state of the art techniques from across industry in edge solutions.

  • Our team has experience running from latent fingerprints, through 3D face recognition, into familial DNA matching. We have worked with the Police and the MOD and understand the limitations of both the software and the people that do biometric matching.

  • We understand industry standards assessment techniques for visible and thermal imagers, as well as what these approaches don’t tell you.

  • Our team has researched various aspects of the surveillance problem, from estimating camera performance to natural language tracking and energy saving modes.

  • Thanks to our understanding of both imaging systems and processing limitations, we can advise on how adversaries may attempt to defeat your system and what you can do about it.

Civil

We have experience of a broad range of industries including aerospace, rail, nuclear and renewables. By combining our technical capabilities with industry specific knowledge we can tailor our research to meet your specific goals.

  • Whether it’s a small piece of graphite in a nuclear reactor or delays over the UK rail network, we can help you analyse and predict the behaviour of complex systems under stress.

  • When statistical modelling isn’t enough, simulation steps in. We can build bespoke simulations or use existing frameworks, and we’ll be sure to tell you the real limitations and uncertainties of any approach we use.

  • In order to understanding how a complex system works, we need to observe it’s behaviour. Sensing (in any modality) offers a view into your black box.