"""Sim4SafeCAV"" will combine safety and simulation for SAE level 4 autonomous cars to significantly enhance safety analysis and use simulation to demonstrate achievement of safety targets. The outcomes will enable OEMs to establish which mix of sensors and systems can meet global safety requirements, creating efficiencies within the supply chain to accelerate products to market. The project is novel in combining Safety and Simulation activities to offer an innovative solution to meet the safety and the time-to-market objectives of L4 autonomous vehicles. Autonomous car technology is developing rapidly, UK government outlined its ambition for self-driving vehicles on UK roads by 2021\. To meet these expectations, strong joint commitment between academia and industry is needed to support the progress of simulation strategies for L4 automation (no driver back up). Autonomous vehicles are part of the international drive to 'zero deaths' and most stakeholders agree that autonomous cars will be safer. Based on this goal, OEMs have a great responsibility to design safe autonomous cars and to provide evidence that they operate safely across the full range of situations likely to be encountered. Given the impractical possibility of driving continuously for 12.5 years with a fleet of 100 cars to gain sufficient statistical evidence to argue safety, this project tackles the challenge by systematically evaluating system safety, limitations and constraints. We will use this knowledge to inform and enhance simulation capability required to drive an equivalent of 12.5 years in a virtual environment and gain sufficient evidence towards the argument for a safe Autonomous Vehicle. L4 vehicles are exceptionally complex systems, in fact hazardous events can happen due to unexpected behaviour with or without the presence of faults or due to malicious intent. New guidelines to help ensure the 'Safety Of The Intended Function' (SOTIF) are currently under development. We will aim to inform such activity with our insights. We will innovate through (1) Enhancing the SOTIF approach (STPA, top-down, bottom-up analysis, noise analysis), (2) Sensor modelling at different fidelities, (3) Advanced simulation setup for gathering evidence for SOTIF, (4) Enhancing the OEM safety validation strategy."

**An exciting project focussed on developing and maturing the simulation, modelling and physical testing supply chain for UK-centric CAM perception sensor and systems developers.** Sim4CAMSens will build a UK supply chain that will advance the quality of modelling, simulation, test and characterisation capability in the UK to accelerate and de-risk the design, development, validation and usage of perception systems sensors and algorithms for automated driving functions. The project will create clear links between the tools, methodologies, standards and safety cases. With state-of-the-art modelling and simulation environments, Sim4CAMSens will deliver much needed synthetic training data of suitable quality for the training of AI systems used in autonomous vehicles. We are bringing together an expert, world-class consortium of partners to support the development of an emerging UK-based perception sensors and systems industry by accelerating the development of perception sensors for assisted and automated driving functions There is a nascent perception sensor design and development industry in the UK, with the potential to challenge global innovation with the right investment and coordination. In parallel, the UK has developed outstanding modelling, simulation and testing capabilities for automated vehicle systems through previous Innovate UK supported projects as well as continued industry investment. We see a major opportunity for the UK by bringing these two worlds together to create a globally competitive sensor design, development, modelling, simulation and testing supply chain. The supply chain will focus on the specific use case of the development and testing of three sensor technologies within the virtual and physical test framework: 1\. RADAR (Oxford RF, Claytex), 2\. Camera technology (rFpro), 3\. LiDAR development (CSAC, Claytex).

Autonomous Vehicle (AV) pilot studies have demonstrated that AVs can handle "normal" driving; 99% of our day-to-day driving experience. However it is proving harder than expected to train AVs to deal with the multitudinous unusual events that can happen on the road, known as edge-cases. When an AV fails to understand an edge-case, driving behaviour becomes unreliable and unsafe, with examples include over-braking causing rear-end collisions, lane-keeping failures, and in the worst cases fatal high-speed collisions. It is essential for AVs to be able to handle edge-cases safely and reliably for public and consumer acceptance of AVs on the road, for the technology developers and automotive manufacturers to meet upcoming regulatory standards and so that the automotive industry can realise a return on the considerable investments made to date in AV technology. Training AV perception systems for edge-cases is challenging; the volume of real-life training data is limited. Simulation and synthetic data are widely recognised as being needed but currently available synthetic training data is not sufficiently sensor-real for simulation-trained AI to be successful in improving perception and decision-making on edge cases. DeepSafe brings together leaders in the simulation supply chain to resolve synthetic data issues inhibiting successful training for AV perception using simulation.