A Fault Injection Tool for Identifying Faulty Operations of Control Functions in Automated Driving Systems

Along with an increasing usage of intelligent driver-assistance service in automotive vehicles, the guarantee of dependability and trustworthiness of the system functions for operation perception and control becomes particularly critical. To cope with the risks, an investigation for determining how such system functions would be affected by different operational conditions, including the deviations of expected inputs, software, and hardware anomalies, becomes necessary. This paper describes a tool that augments Simulink/MATLABTM with additional fault models for effective analysis of error behaviors of automated control systems. This constitutes a basis for simulating different fault scenarios, collecting the operational data, and thereby the design of embedded online functions for error detection and treatment. This approach is demonstrated by a case study with an automated emergency braking system.