Tags:Autonomous Driving Platforms, Digital Twin Simulation, formal analysis, runtime verification and Safety
Abstract:
The future of the automobile industry lies in Autonomous Vehicles (AV). Despite the efforts of major companies to completely automate driving, ensuring AV’s safety remains challenging. The objective of our work is to provide a compositional simulation interconnect framework to verify the safety of Autonomous Driving Platforms (ADP). This objective is achieved by translating high-level safety requirements from ISO 26262 and ISO 21448 into verifiable properties and building a comprehensive digital twin comprising of a Runtime Verification (RV) monitor for property checking together with a car scenario simulator and ADP-under-test. We demonstrate our framework using Apollo an open-source autonomous driving platform developed by Baidu as a case-study.
Safety Analysis of Automated Driving Platforms Using Digital Twin Simulation and Runtime Monitoring