Safety Evaluation of Automated Driving Systems


At a glance

Automated driving systems (ADS) have been developed and marketed in recent years and likely will be further deployed in the coming years. Regardless of the automation levels, per SAE J3016 standard of Levels 1-5 [1], the safety impact of ADS is doubtlessly an inspiration, but at the same time a concern if performance is not up to expectations. The on-road experiences of ADS are generally insufficient to allow a conclusive safety assessment.

In this study, we aim at formulating a systematic model for safety risk estimation, particularly if the targeted ADS is new to the market and there are no comprehensive data to support convincing safety performance. This could be useful for consumers to try at least to understand the safety risks and an insurer to use as a reference to assess the exposure of its customers. More broadly, it may assist public agencies in seeking a balance of risk and benefits when evaluating or accepting the deployment of ADS in its community.

principal investigatorsresearchersthemes
Ching-Yao ChanKang Li

Automated driving system, Safety Evaluation, Bayesian probability network