key components for autonomous driving

Over the past 10 years, the development of self-driving cars has seen increasing interest. More than $106 billion has been invested in the sector, mainly used to improve advanced driver assistance systems that manage braking, obstacle detection and other critical vehicle functions.

These new systems are critical: they must instantly process data to react to unexpected events such as a sudden stop in traffic. To function properly they require multiple interconnections within the vehicle and high-performance microprocessors. They need a more centralized electrical and electronic (E/E) architecture than traditional vehicles, as well as more sensors and electronic computing elements.

While current systems provide for multiple control units (ECUs) distributed throughout the car, future cars will be designed with a different logic, based on interconnected sectors. The management of the car will be divided into zones, positioned in strategic places and in dialogue with each other. There will be no more thousands of chips and meters of cable. Car wiring is estimated to decrease by 20%.

The strong focus on autonomous vehicles has already changed demand patterns for automotive semiconductors, with strong growth in sales of special silicon microprocessors tailored to these specific applications. These chips are more efficient and lead to a rapid increase in performance within vehicle systems. They allow the execution of complex software and analysis functions, such as those that allow the integration of sensors, cameras, lasers, LiDAR (Light Detection And Ranging) and other devices. They are also able to learn new information through deep-learning and actively train artificial intelligence, in order to be constantly evolving and ensure an unprecedented level of security. Autonomous driving is destined to become more reliable than human’s driving, thanks to the greater calculation capacity. According to some estimates it will be one sextillion higher than the man’s one.

However, these custom microprocessors are made by only a few semiconductor companies, and many self-driving car manufacturers face the scarce availability of chips and the difficulty of obtaining products that exactly match their needs. Faced with these constraints, some automakers have begun to design chips in-house to reduce development time and gain greater control over production. This solution offers several benefits: optimized performance for specific algorithms and shorter development times, thanks to continuous functionality improvements. The internal design also allows you to define the software stack and allows greater control over the design of microprocessors so that you can create customized solutions that could differentiate the performance of one car manufacturer’s self-driving vehicles from those of another.