Rationale: This document provides dynamic test procedure information and guidelines for verification and validation (V&V) of automated driving systems (ADSs). The levels of automation addressed in this document include conditional (level 3), high (level 4), and full (level 5) as defined by SAE J3016; when activated, these ADSs do not rely on a human driver for monitoring and responding to the vehicle or traffic environment. (SAE J3016 defines the italicized terms in this document.) The information and guidelines contained within apply to the verification and validation (V&V) of all types of motor vehicles including light-duty, passenger, freight or transit vehicles by specifying dynamic test procedures, such as those performed on a test track, for verifying and validating ADS functions. This document does not address component-level V&V or broader top-down systems-level V&V practices.
V&V applies to verifying and validating the functionality and dependability of on road ADSs for correctness and completeness. The functionality of an on road ADS is defined by its functional specification of behavior. The dependability of an on road ADS is defined by its safety, reliability, availability, maintainability, and integrity. Functionality and dependability are impacted by the correctness and completeness of requirements, design, implementation, and operational deployment of the on-road ADSs based on our understanding of the key nominal and non-nominal use cases to which on road ADSs could be subjected.
As key nominal and non-nominal uses cases for specifying aspects of functionality of ADSs emerge and evolve, this information report is designed to complement such uses cases by specifying how ADSs embodying such functions may be V&V’d for dependability.
As ADSs evolve, different means for verifying and validating these functions will be adopted and thusly evolve. This information report harnesses best practices and information from industry to serve as a means to guide the V&V of ADSs as ADS functions evolve. This information report thus provides a generic template for the V&V of ADSs as well as provides a few specific key examples, which apply this template to specific well-defined ADS functions. Here, V&V applies specifically to the dynamic and live trackside testing of ADS functions.