Future research needs to assure the safety of CAVs: a point of view from the field of robotics and autonomous systems Prof. Tony Pipe – Bristol Robotics Laboratory (Tony.Pipe@brl.ac.uk) https://www.bristolroboticslab.com/
Autonomous Vehicle Projects in the West of England Started in July 2015 and June 2016 respectively: • User acceptance and impact on insurability and legality • ’Flavour’ of these The needs of older adults as users, system cross-disciplinary deployment and data security projects has evolved over 5 Started in late 2017 or 2018: MultiCAV: Transport system implications and passenger experiences of a multimodal autonomous mobility years from • investigations into system. Electric, fully-autonomous on-demand shuttle vehicles, electric buses and taxis, alongside conventional vehicles in a 'Mobility as a Service' (MaaS) platform user acceptability CAPRI : A market ready, mobility service using trusted secure PODs and systems supported with legal, regulatory, • of individual CAVs insurance recommendations. to safety ROBOPILOT: Autonomous driving for electric Light Commercial Vehicles (LCV). Integrated vehicle, fleet and • assurances for the depot delivery management system. Full assessment of vehicle safety, evaluation and hardening of cyber-physical technology security interacting with Started Spring 2019: complex CAV-Forth: Autonomise the UK’s most popular full size single deck bus. Create a high capacity AV Bus Pilot • environmental Service. Utilise a route across the Forth Bridge UNESCO World Heritage location. Optimise the safety and quality of systems the service with respect to consumer needs
Autonomous Vehicles we have utilised … or will
Simulation will be important …
How do we assure Autonomous Vehicle Safety? • Simulation has many benefits over real-world testing – Safely cause dangerous situations – Parallel test runs save time – Full control over environment state – More cost effective • Verification goals include: – Rules of the Highway Code – Social conventions – Coverage, including code and functional/scenario • However, there are weaknesses of course, real-world testing is necessary! M. Webster, D. Western, D. Araiza-Illan, C. Dixon, K. Eder, M. Fisher and A. G. Pipe (2018). A Corroborative Approach to Verification and Validation of Human-Robot Teams . https://arxiv.org/abs/1608.07403
Other Challenges and Comments • The sensor-perception pipeline is a significant challenge • Simulation approaches need to be developed further and standards activities will be crucial • Standards and legislation need to proceed rapidly • SAE autonomy table needs to be enhanced, especially in the light of emerging application domains
https://www.venturer-cars.com/ http://www.flourishmobility.com/ http://caprimobility.com/ https://arrival.com/ (ROBOPILOT project) Tony.Pipe@brl.ac.uk https://www.brl.ac.uk www.bristol.ac.uk/tsl
Recommend
More recommend
