Technology development for Autonomous Vehicles is moving from functionality to performance to reliability. However, current self-driving cars and trucks need to reach stringent safety and security levels before going broad, as strict reliability standards are critical to guarantee human protection and to drive public acceptance of the autonomous technologies.
Defined by the ISO 26262 standard, the Automotive Safety Integrity Levels (ASIL) are a rigorous risk classification scheme for self-driving reliability, ranking vehicles from A to D levels. While an average US driver makes 10,000 mistakes in 1 billion hours of driving, ASIL-D compliant vehicles can only make 10 errors during the same period. This is forcing engineers to employ state-of-the-art hardware and software design methodologies to achieve those high-reliability requirements.
Failures in embedded systems are classified into permanent (repeatable faults that happens the same way) and transients (temporary faults that can randomly occur). Safety engineers handle permanent faults with thorough testing frameworks, but they are working hard trying to prevent transient faults as they are a function of the environment. Redundancy is the more practical and fail-proof solution to mitigate failures that engineers are implementing while the self-driving is engaged —integrating multiple sensor technologies and systems (like Radar, LiDAR, ultrasonic, and cameras) to detect obstacles. Safety engineers are increasing reliability and precision for self-driving vehicles by taking advantage of advancements in other technology sectors like the rise of the Internet of Things (IoT), enabling camera sensors not only on autonomous vehicles but also at intersections and high-risk regions on roads, or the arrival of 5th Generation (5G) wireless systems, allowing high-precision positioning for lane-detection on the streets using Global Positioning System (GPS).
Nowadays, not only automotive markets are demanding more severe design specifications, but also other industries like energy, aviation, or medical are requesting complex safety and security requirements that push engineers to develop stringent secure applications with safety compliance. Functional Safety is challenging engineers to design robust E/E systems to reduce faults in harsh environments that can prevent dangerous failures, or control them —if they occur. Fully self-driving vehicles are not far from becoming a reality as engineers are using safety architectures and system design to enable full redundancy to facilitate higher levels of autonomous driving and fault tolerance in the case of failure, because it is all about self-driving reliability.