Ever run a marathon? Then you’ll appreciate there’s no substitute for putting in the hard miles (or kilometers) well before you take your place on the starting line.
Precisely the same principle applies when it comes to developing advanced safety systems. The process of testing technologies such as lane assist invariably means putting a lot of drive time on the clock.
So just how far do we go? Inevitably, the precise figure depends on the nature of the technology involved. However, it’s not unusual for Aptiv to travel a million miles testing a new solution. What’s more, as we move into the supremely challenging field of autonomous vehicles that number is rising inexorably.
But what’s the rationale behind such an extraordinary testing regime? Surely, in an age of Artificial Intelligence (AI), machine learning and virtual reality, we can find the answers we need without ever having to step beyond the confines of a well-equipped technical center.
There’s no doubt, in the process of creating world-class products, we make full use of sophisticated lab technology. But we still need first-hand experience of everything that the real world is likely to present our systems. That means the technology needs to stand up to a lifecycle of 15 years and more.
Why? Well, to start with, consider the consequences of a system failure. The implications of autonomous emergency braking missing a potentially lethal threat are all too obvious. Equally, false alarms generated by systems, such as blind spot warning, would quickly erode confidence among drivers.
Now factor in the sheer number of variables that our systems must recognize and react to. The list is near-endless, but includes every conceivable weather condition, in daylight and night-time, through landscapes ranging from the urban jungle to vast and featureless deserts to multi-lane highways and winding country lanes.
Put simply, our solutions must respond to a vast array of challenges. For example, is the impending impact detected by our radar system the bumper of another vehicle or a rock thrown up by the car in front? What happens if snow builds up on that radar? Can we confirm the system is working correctly? In an urban environment, there will be plenty of buildings, street signs and similar objects to provide a positive reaction to the radar. In the rural countryside, it’s a completely different story. Here there could be next-to-nothing to reflect energy back; nothing to demonstrate the radar is functioning as it should.
Clearly, every one of those miles covered in testing is invaluable. But the learning process doesn’t stop there. After all, we first supplied a radar-based system back in 1999. Over the past 30 years, the number of miles our technologies have covered run into the billions, all of which further contribute to the reserves of experience and expertise.