Ensuring Safe Autonomy: Navigating the Future of Autonomous Vehicles

Patrick Wolf

2024-03-28

Abstract:Autonomous driving vehicles provide a vast potential for realizing use cases in the on-road and off-road domains. Consequently, remarkable solutions exist to autonomous systems' environmental perception and control. Nevertheless, proof of safety remains an open challenge preventing such machinery from being introduced to markets and deployed in real world. Traditional approaches for safety assurance of autonomously driving vehicles often lead to underperformance due to conservative safety assumptions that cannot handle the overall complexity. Besides, the more sophisticated safety systems rely on the vehicle's perception systems. However, perception is often unreliable due to uncertainties resulting from disturbances or the lack of context incorporation for data interpretation. Accordingly, this paper illustrates the potential of a modular, self-adaptive autonomy framework with integrated dynamic risk management to overcome the abovementioned drawbacks.

Robotics

What problem does this paper attempt to address?

This paper discusses the challenges of safety verification for autonomous driving vehicles, especially in complex environments and under perception uncertainties. Current approaches suffer from performance limitations due to overly conservative safety assumptions. The authors propose a modular and self-adaptive autonomy framework that combines a dynamic risk management system to overcome these limitations. They emphasize the importance of situational awareness in safety decision-making and highlight the adaptability of humans in handling unforeseen circumstances as inspiration for autonomous driving systems. The paper discusses approaches in behavior-based robotics such as behavior networks, which excel in control and perception but lack systematic risk assessment and standardization.

The paper mentions two scenarios to illustrate that relying solely on nominal safety (e.g., RSS rules) may be insufficient to handle all situations. In these scenarios, dynamic risk management can improve safety by considering a broader range of contexts and making more granular decisions. The authors believe that integrating behavior-based systems with dynamic risk management has the potential to significantly enhance the safety of autonomous driving systems.

Future efforts will focus on the system integration, technology demonstrations, and standardization of these methods.