Autonomous vehicles are revolutionizing the automotive industry, with safety being the primary concern. To ensure real safety, Advanced Driver Assistance Systems (ADAS) must be fully redundant. Currently, most OEMs, Tier 1 suppliers, and L4 autonomous driving companies focus on software and hardware backup for redundancy.
In terms of software redundancy, companies like GAC Group are implementing algorithm redundancy strategies to enhance the reliability of systems like Automatic Emergency Braking (AEB). Hardware redundancy, on the other hand, involves dual or multi-redundant designs for different functional components such as perception, decision-making, and power supply systems.
Execution redundancy, particularly in braking and steering systems, is crucial for ensuring correct responses in critical situations. Braking systems are evolving from electro-hydraulic systems to brake-by-wire systems for improved response times and inherent redundancy capabilities. Similarly, steering systems are transitioning to steer-by-wire systems for quick response, flexible installation, and high collision safety.
One example of a cutting-edge steer-by-wire system is the NIO NT 3.0 Platform, which features dual power supply, communication, hardware, and software redundancy. This system, which eliminates mechanical steering columns, boasts 2.2 times higher reliability than traditional electric power steering systems.
As the industry progresses, integrated redundant designs are becoming increasingly popular. The CATL Integrated Intelligent Chassis (CIIC) integrates various vehicle systems into a highly intelligent skateboard chassis, promoting scalability and system safety. Similarly, the BYD e4 Platform utilizes a three-in-one driving-braking-steering technology for enhanced safety redundancy capabilities.
Control redundancy is also critical for fail-operational requirements in autonomous driving. Solutions like multi-ECU redundancy and single-chip redundancy are being explored to ensure system reliability in the event of sensor failures.
In conclusion, the future of autonomous driving lies in integrated and fully redundant designs that prioritize safety and reliability. With advancements in software and hardware technologies, the automotive industry is moving towards a new era of autonomous vehicles that prioritize safety above all else.