Road Safety: It’s All About ADAS

Article By : Rob Stead

The sine wave of safety has brought the industry back into focus. I believe the straight-to-Level-5 players will have their time, but right now it’s all about ADAS...

I’m a big believer in cycles. Virtually everything is cyclical. The sine wave is one of the most fundamental functions in mathematics, representing relationships between subatomic particles right up to the nature of the universe itself. But the elastic, return-to-origin motion that the sine wave represents is illustrated throughout society, politics, the economy and, in my view, almost all aspects of life. With all the hype about robotaxis and the utopic future of mobility over the past five years or so, we have lost sight of what autonomy was all about in the first place. Right now, the elastic motion of the sine wave is bringing us back around to focus on what the original objective was, namely, safety. The Covid-19 pandemic has created a challenging environment for the automotive industry, but it’s also helped us regain focus on what matters most. Autonomy was meant to be a by-product of safety: We’d build the safest cars we could, and by default they would be fully autonomous because computers would be better at driving than humans. I still believe that this will be the case in the long term, but it was always going to be a big challenge, even in boom times with seemingly unlimited budgets. In the engineering community, it’s become clear more recently that autonomous driving is really, really hard, and in a post-Covid-19 economy solving the last 5% of technical challenges could be slower than previously thought. But that doesn’t mean we suddenly give up on safety. Automakers, regulators, and consumers alike still want to see the road become a safer place, and despite the squeeze in some areas of automotive manufacturing, it’s encouraging to see safety programs remain a top priority. I’ve always said the key was to focus on incrementally improving advanced driver assistance systems, and if we stick with driver assistance, the action the car takes in an emergency is mostly limited to bringing it to a safe stop. Safety has many facets, but if we focus on ADAS, it’s primarily about perception:
  • Sensing — improving sensor capabilities and optimizing image quality
  • Robustness — building in system redundancy, SOTIF (Safety of the Intended Functionality), and a continuous development approach
  • Monitoring — real-time assessment of system performance, understanding degradation pathways and impacts
Car safetyThere’s some amazing tech already on the road, and having updated my car last year I now have personal experience of automatic emergency braking (AEB) and can confidently say it far outperforms any accident avoidance a human could perform. There are also studies that show systems sometimes don’t perform as intended, so ADAS is clearly still a work in progress. And there are still many ADAS elements that are still in the R&D phase and not yet deployed on the road, actively saving lives. By focusing on the performance factors above, we can improve customer-ready ADAS and make a real impact today, well before we have fleet-level deployment of robotaxis and other self-driving vehicles. Don’t forget that even ADAS is the cutting edge of computer vision. High speed, high risk, multiple sensor modalities — there are many components and many external factors to consider. It’s hugely complex, so developing robust ADAS functions is really a more productive use of resources. But — and this is a big but — it must be done within the context of future demand and safety strategies. The biggest dividing factor between traditional OEMs and the new breed of automakers — whether established players like Tesla or pre-commercial brands such as Zoox — is that the new players are not constrained by legacy electronics and software platforms. As ADAS connects more and more of the dots to present a full safety picture, there’s an increased need to approach this work by using a common software platform and a continuous development approach. This is core to the way Tesla operates, but it’s a paradigm shift for many OEMs. The great news is that it’s happening, as evidenced by recent announcements from major OEMs. It’s not easy work, but it’s essential. And once this approach becomes mainstream, we’ll see safety capabilities advance at a much higher rate. The context I mentioned also must include more advanced motion planning, and a continuous development approach is all but essential here. There’s lots of interesting research going on to optimize digital driver behavior, whether in collision avoidance, to ensure a smooth traffic flow, or in a motorsport environment like Roborace. I’m excited to read about these developments and look forward to seeing them integrated into ADAS safety features. From a market perspective, I think one of the keys to driving demand for these technologies is selling safety as a desirable feature. If the market wants it, industry will follow. I’ve heard that even muscle car buyers in the U.S. are impressed by the radar panel on the front that will help them avoid a crash, and brands like Volvo have built a reputation on safety for many years. The sine wave of safety has brought the industry back into focus. I believe the straight-to-Level-5 players will have their time, but right now it’s all about ADAS. This article was first published on EE Times Europe

Leave a comment