Quality assurance methodologies for automated driving

Abstract

For safety critical systems like cars, trains, or airplanes quality assurance methods and techniques are crucial for preventing situations that may harm people. The case of automated driving represents the next level of safety critical systems where additional challenges arise. This includes the question of how to assure that artificial intelligence and machine learning based systems fulfill safety criticality requirements under all potential conditions and situations that may emerge during operation. In this paper, we first review simulation-based verification and validation methods for such systems and afterwards discuss necessary requirements and future research challenges that have to be solved in order to bring automated driving into practice without compromising safety requirements.

Enhancing Human-Machine Teaming for Medical Prognosis Through Neural Ordinary Differential Equations (NODEs)

“…In other cases, tracing the error is much more difficult, and oftentimes impossible. This is of immediate concern for makers of industrial-scale autonomous systems such as self-driving cars [8], but also of great concern in applications that feature machine learning in the role of decision support, as is the case in clinical decision support systems. Physicians, hospital administrators and even government regulators, upon seeing the apparently brittleness of ML are likely to ask themselves "if this system has such low reliability and unpredictability, how can I ethically justify using it for my patients?"…”

Performance Comparison of Two Search-Based Testing Strategies for ADAS System Validation

“…All investigated testing strategies are explained in more detail in the upcoming chapters. The underlying tool-chain for automatic test scenario generation, execution and evaluation is based on previous work in this eld [23,14].…”

Smart Monitoring for Safety-Assurance in Autonomous Driving (S)

“…Checking that a concrete implementation of autonomous driving reaches this goal requires sophisticated verification and validation before deployment. Many researchers have been working on this topic, including Koopman and Wagner [6], Wotawa [11], Schuldt and colleagues [9], or Wotawa and colleagues [13].…”