Siddartha Khastgir
Head of Safe Autonomy
Siddartha Khastgir is the Head of Verification & Validation for Connected & Autonomous Vehicles (CAV) at WMG, University of Warwick. He leads several collaborative R&D projects with industrial and academic partners nationally and internationally. His research focuses on generating safety evidence and arguments, test scenario generation, simulation-based testing, and safety of AI systems. Leveraging the cross-domain nature of safety, he is also involved in safety research in aviation, marine and healthcare.
Siddartha is an active member of various national and international standardisation and regulatory groups, including ISO, SAE and ASAM. Currently, he represents the UK on several ISO technical committees and is the lead author for two new ISO standards for aspects of automated driving systems. He sits on the United Nations Economic Commission for Europe (UNECE) committees on safety of automated driving.
Prior to joining WMG, Siddartha was with FEV GmbH in Germany, leading automotive software development and testing for series production projects.
He has received numerous national and international awards for his research contributions, including the prestigious UKRI Future Leaders Fellowship in 2019 focused on safety evaluation of CAVs.
Presentation
Objectifying Safety for Automated Driving: Defining the stopping criteria through a Leading Indicators Approach
2025 has the potential of being a pivotal year for Automated Vehicles (AV) ecosystem, with various large-scale deployments in the Americas and Europe & Asia moving to the later stages of regulation, including the secondary legislation of the UK’s Automated Vehicles Act. However, one of the key challenges associated with understanding safety of AV is in establishing the safety threshold or benchmark. Various approaches have been suggested, such has Careful & Competent Human Driver or comparison to historic accidents; however, a well understood and accepted objective definition of safety for AV still evades of the industry.
In this talk, we present an objective leading indicator-based approach to defining safety for AV and stopping criteria for their testing.
