It wasn’t too long ago that we were being told that self-driving cars would take over the roadways by 2020. Well, 2020 is only a few months away and it’s pretty clear such bold predictions will not be a reality. Indeed, we may not have self-driving cars by 2025 or 2030. As wonderful as the idea is, technology isn’t there yet.
Self-driving cars already exist on paper. They exist as a theoretical application of automation technology that requires very little human input. But there’s one problem: self-driving cars are waiting for sensor technology to catch up. Until it does, autonomous vehicles will remain the domain of films and literature.
Sensor Technology is Deficient
The New York Times ran an interesting piece in late September 2019 discussing the state of autonomous vehicles in the US. According to contributor John Quain, current sensor technology is deficient enough to have slowed down the development of self-driving cars. Quain even wrote that the automotive industry has had to hit the reset button on plans to bring autonomous cars to market.
The engineers at Rock West Solutions would agree. As experts in sensor technology for commercial applications, they recognize that the current state of technology doesn’t allow for systems capable of making the same kinds of decisions the human brain makes. Driving in rain is a very good example.
The self-driving cars of the future will rely on a variety of sensors to keep cars safe during rainstorms. Those sensors don’t yet exist. Put a first-generation self-driving car in the midst of a bad storm and it will have trouble navigating safely. As Quain so eloquently puts it, we need sensors “that can see through rain” much better than they do now.
Contrast that to the capabilities of the human brain and ocular system. Our brains can compensate for fairly heavy rains and allow us to navigate safely, even if that means driving a bit more slowly. Our eyes are able to focus beyond rain to see cars ahead. Our brains can look beyond the rain to see a traffic signal turning red.
Moving Beyond the Standards
If self-driving cars are ever to become reality, auto makers have to move beyond the four standards: video, radar, lidar, and ultrasonic sensors. Sensors have to be better at ‘seeing’ and ‘hearing’ the general surroundings. More importantly, they have to be able to perceive conditions that aren’t immediately evident.
Sensor technology is still a long way from achieving such lofty goals. But even when it finally arrives, there is yet another problem we will have to overcome: a national infrastructure that is wholly inadequate for an autonomous vehicle network. Not only are we going to have to upgrade our cars, we will have to upgrade our roads as well.
A litany of technologically advanced sensors might someday make it possible to operate self-driving vehicles in a closed loop environment. But for those vehicles to be safely taken on public roads, we will need new infrastructure capable of receiving data from every car on the road, analyzing that data, and putting it to use for safely managing traffic.
We are going to need smart roads with a litany of embedded sensors. Each of those sensors will have to be networked with the capability of communicating simultaneously with vehicles and traffic management computers. Throw in GPS and other extra features and you have a system relying on millions – if not billions – of interconnected sensors tasked with keeping us safe.
It is a noble dream indeed. Unfortunately, it is also a dream that remains a long way off.