In the recent years, Artificial Intelligence (AI) has emerged as a transformative force in the automotive industry, promising to revolutionize how vehicles are driven and the way they interact with their surroundings. AI technologies, such as machine learning are promised to enable vehicles to make complex decisions, navigate challenging environments, and even communicate with other vehicles and infrastructure.
However, using AI brings along critical safety concerns. Ensuring that AI-powered road vehicles operate safely and reliably is essential to avoid accidents and protect human lives. This is where the soon to be released ISO 8800 will play an important role.
In an interview with INVENSITY’s Head of Automated Driving and safety expert, Sascha Hackmann, we investigate the significance of AI technology in automotive industry and the crucial questions concerning the intersection of AI and the safety of road vehicles.
For more information about the topic and what we have to offer, contact our Head of Automated Driving department, Sascha Hackmann, directly.
INVENSITY: How significant do you think AI technology is in the context of road vehicle safety?
Sascha Hackmann: Very significant. Since more automated driving vehicles need a better situational awareness, and more responsibility is handed over to the machine with increased level of automated driving. This can be achieved more effectively through continuous learning, which is why machine learning algorithms will be more and more included in SW development for embedded microcontrollers inside vehicles.
INVENSITY: What are some potential benefits of using AI in road vehicle safety, and what are some of the associated risks and challenges?
Sascha Hackmann: AI is able of shaking up road vehicle safety with a wide range of benefits. For instance, AI-powered systems can analyze data from sensors and cameras in real-time to detect potential hazards, such as pedestrians, vehicles, or obstacles, and take appropriate actions to prevent accidents. Furthermore, AI can enable features like adaptive cruise control, lane departure warnings, and automatic emergency braking, which will assist drivers in making safer decisions and reacting to sudden changes on the road.
AI algorithms can predict and prevent collisions by analyzing the behavior of nearby vehicles and pedestrians, allowing vehicles to take evasive actions.
Another benefit can be AI-based computer vision systems, which will improve visibility by identifying objects that might not be easily visible to human drivers, such as low-light conditions or blind spots. Other beneficial topics are traffic management, driver monitoring, and parking assistant.
However, one should not forget the risks and challenges that AI brings along in road vehicle safety. Among some of the most important risks and challenges we have reliability, data quality and diversity, cybersecurity, and ethical, legal and privacy concerns.
For instance, ensuring effective communication between AI systems and human drivers is a very crucial point. Drivers need to understand when and how AI systems are making decisions, and they must be prepared to intervene when necessary. Or as another example, overreliance on AI systems could lead to complacency among drivers, who might assume that the AI can handle all situations. This could lead to delayed responses or failures to take control when necessary.
INVENSITY: Can you explain what ISO 8800 is and what it aims to achieve in terms of AI and safety in road vehicles?
Sascha Hackmann: ISO 8800 will be an automotive standard that will define how to achieve functional safety with the implementation of machine learning algorithms. It is still planned to be released within 2023 and will broaden the relevant standards for automotive safety like ISO 26262 (addressing Functional Safety), ISO 21448 (addressing Safety of the intended functionality) or ISO 21434 (addressing Cybersecurity). It will create a new framework which will address all phases of the safety lifecycle for AI development (requirements derivation, data quality / completeness, control / mitigation of failures, AI supporting tools, V&V techniques, creation of an assurance argument).
INVENSITY: How do you envision the guidelines provided by ISO 8800 being used by organizations involved in the development and validation of AI for SOTIF in road vehicles?
Sascha Hackmann: AI safety guidelines for SOTIF in road vehicles would be a comprehensive roadmap that organizations can follow to ensure that their AI systems are developed, tested, and validated with safety as a top priority. These guidelines would provide a structured approach to addressing potential hazards and ensuring that AI systems behave as intended in complex real-world scenarios.
INVENSITY: How might ISO 8800 impact the broader landscape of road vehicle safety and AI technology in the coming years?
Sascha Hackmann: It could create a more structured and reliable foundation for the integration of AI technology into vehicles. It would influence technology development, regulatory compliance, consumer confidence, and various other aspects of the automotive industry and transportation ecosystem.
In technology development, the standard could drive the development of more advanced and reliable AI algorithms specifically designed for road safety. This might encourage the automotive industry to invest more in research and development to meet the new safety criteria. Furthermore, the standard could serve as a reference point for regulatory bodies when developing rules and regulations related to AI-powered vehicles. This alignment between industry standards and regulations could create a smoother path for the adoption of AI technology.
As safety concerns are one of the main barriers to widespread usage, a strong AI safety standard could enhance public trust in autonomous and AI-powered vehicles since a recognized standard could better deal with some of these concerns.
INVENSITY: What advice would you give to organizations or individuals looking to stay up-to-date on developments related to AI and safety in road vehicles?
Sascha Hackmann: There are different ways to stay up to date. Follow industry news and publications, and attend conferences, workshops, and webinars. Apart from that, joining professional organizations, reading and analyzing newly published standards and engaging with research papers are of great help.
INVENSITY: Are there any other emerging trends or developments in the field of road vehicle safety that you think are worth keeping an eye on in the coming years?
Sascha Hackmann: Definitely. There are developments like connected vehicles (V2X communication) which offer real time information on traffic status, road conditions, and potential hazards, or the integration of advanced sensors, more precise radar systems, and enhanced LiDAR technology which can result in better hazard identification and collision prevention.
Apart from that, there is a big effort in development of Human-Machine Interaction which aims to create user-friendly methods for drivers and passengers to interact with AI systems and comprehend their responses. Additionally, the implementation of biometric sensors such as eye-tracking devices and heart rate monitors can help in recognizing driver distractions and drowsiness, therefore contributing to overall driving safety.