STREETWISE

Rules based AI performance validation

The STREETWISE project was based on the concept for a fully automated, shared transport service, targeted at replacing the urban commuter car. The project developed and demonstrated the technology, safety validation methods, insurance and service models required to make it a commercial success. TRL’s role was to develop the safety case for the on-road trials, suitable for a major insurer to underwrite it.

We also analysed UK collision data in order to define priorities for simulation-based testing. The definitions are based on a new Scenario Description Language (SDL), which incorporates both static (road geometry) and dynamic elements (vehicles, cyclists, pedestrians). We then curated a database of test scenarios that will be used to test and validate the performance of automated vehicles, in particular ensuring that they respond accurately and appropriately to a vast range of scenarios encountered on UK roads.