Researchers propose JTA to improve safety-critical software validation

Researchers have unveiled the Joint Testability Architecture (JTA), a novel framework designed to enhance the validation of safety-critical software systems. Published on arXiv, the work by Wenyao Xue, Jiandi Wang, and Yichen Wang tackles a persistent challenge in the field: ensuring that critical scenarios are tested under controlled conditions while maintaining clear attribution of abnormal outcomes.

JTA introduces a unified approach by treating the scenario, test system, and system under test as a single analytical object, characterized across three dimensions—controllability, observability, and isolability. The architecture leverages scenario contracts, joint capability assessment, and validation blind-spot identification to map capability gaps to actionable improvements in control points, evidence organization, and attribution boundaries. For example, analysis of ArduPilot’s failsafe validation revealed that link-loss scenarios are relatively mature, while state-estimation anomalies remain harder to validate due to weaker evidence alignment.