A consortium of eleven organisations is building a high-fidelity computer simulator to test self-driving cars before they ever hit a real road. The problem is straightforward: connected autonomous vehicles (CAVs) need to be proven safe, but testing them on public roads is slow, expensive, and potentially dangerous. Current simulators often train vehicles on narrow, repetitive scenarios, leaving them unprepared for the messiness of real driving. OmniCAV aims to create a single, authoritative simulator that regulators, insurers, and manufacturers can trust to certify a vehicle as ready for road trials. If successful, the simulator will change how autonomous vehicles are approved. Instead of piecemeal testing, a CAV would face thousands of randomised, holistic scenarios—rural lanes, peri-urban junctions, and city streets—generated from real traffic, accident, and CCTV data. AI-trained virtual road users will interact unpredictably with the vehicle under test. A security "root-of-trust" design will protect the integrity of test inputs and outputs. The result could be a standardised certification tool that accelerates safe deployment of autonomous vehicles across the full UK road network, not just motorways.
View original technical description
"OmniCAV will lay the foundations for the development of a comprehensive, robust and secure simulator, aimed at providing a certification tool for Connected Autonomous Vehicles (CAVs) that can be used by regulatory and accreditation bodies, insurers and manufacturers to accelerate the safe development of CAVs. It brings together a team of eleven internationally renowned organisations, with decades of accumulated knowledge in the area, in order to produce a single-point-of-call simulator to establish when a CAV can safely progress from a testbed to road trial. To achieve this, OmniCAV will use highly detailed road maps, together with a powerful combination of traffic management, accident and CCTV data, to create a high-fidelity dual (traffic and driving) simulation environment, including AI-trained road users to interact with the AV under test. Scenarios for testing will be developed and randomised in a holistic way to avoid CAVs training to specific conditions, whilst maximising coverage, and the integrity of the testing environment will be taken into consideration through creation of a root-of-trust design to secure the test inputs, simulator configuration and resulting test outputs. Critically, the simulator will offer market-leading coverage of a representative element of the UK road network, through encompassing rural roads, peri-urban and urban roads, to help enable autonomy for all. Representatives of the key end-users, including a local authority, an OEM and an insurance provider, will be engaged throughout to understand their needs. The validity of the synthetic test environment compared to the real-world is of particular importance, and OmniCAV will be tested and refined through an iterative approach involving real-world comparisons and working in conjunction with a CAV test-bed. This is an ambitious project aiming to step-change the safe trialing of CAVs in a safe, holistic and challenging manner in order to accelerate their training, deployment and adoption."
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know