Yes, we’re talking about radar, the same technology that began as a curiosity at the turn of the 20th century, helped Britain fend off the Luftwaffe during WWII, and long allowed weather forecasting and enabled air traffic controllers to keep our sky. safe.
Today, radar is no longer reserved for airplanes and military installations. A number of new companies, both on the hardware and software side, are making radar an integral part of safety systems to detect cyclists and pedestrians.
The need for such technology is urgent: Pedestrian fatalities in the United States have increased in recent years, even as Americans have driven fewer miles on the road. And the increasing levels of range in new vehicles that enable features like collision warning, automatic braking and blind spot detection, not to mention the driverless cars of the future, depend entirely on advanced sensory systems. Such systems are also essential for automakers to keep their promises to integrate automatic braking systems into all vehicles by 2022.
To reduce and prevent the carnage on our roads, companies like Mobileye, a subsidiary of Intel, are working on chips bristling with tiny radar antennas. General Motors recently invested in Oculii, a startup entirely dedicated to math and software that uses machine learning to shape the types of signals used by automotive radar systems. Software company MathWorks is developing algorithms that can allow automakers to integrate data from radar and other sensors into a reliable picture of the world around a vehicle.
For engineers working on vehicle sensors, now is a time of rapid change, says Erez Dagan, executive vice president of product and strategy at Mobileye. Cameras used in automobiles continue to have a higher resolution and are able to detect a wider range of natural light than before. Lidar, which bounces lasers off surrounding objects to “see” the world in 3D, is becoming cheaper than before. (Lidar is common in prototype robot taxis, like those from Waymo, sister company to Google Cruise and Amazon’s Zoox.)
Radar, which bounces radio waves off objects – the term originated as an acronym for “radio detection and ranging” – has been used on some first-generation security systems in vehicles since the 1990s. Automotive radar systems have a number of advantages. They are tough enough to survive years of jostling and temperature swings when fitted to cars. They are much, much cheaper than lidar, good at instantly measuring the speed of objects, and able to scrutinize the types of adverse weather conditions, like fog and rain, that can bypass both cameras and lidar systems. But until recently they had one major drawback: They only have a fraction of the resolution of these other systems, which essentially means the images they produce are much more blurry.
Oculii’s technology works by changing the shape – also known as the waveform – of the radar signal sent by the radar on cars. The physics are complicated, but by changing the nature of the radar signal depending on the type of objects it bounces off, it can resolve objects whose shape would be impossible to “see” otherwise. The result, according to managing director Steven Hong, is that existing automotive radar sensors, which cost around $ 50 each, can generate three-dimensional images of a car’s surroundings with much higher resolution. will be released in 2023.
Taking advantage of the chip-making capabilities of parent company Intel, Mobileye is working on individual microchips covered with nearly 100 tiny antennas. Using artificial intelligence software to process the noisy signals they receive, Mobileye says its systems can do things like identify pedestrians, at least in the lab. This is something that previously could only be achieved with cameras and lidar.
There is no unanimity among automotive technologists on how to configure cameras, lidar and radar that will become the standard way to achieve various security or autonomous driving systems, but almost all agree that the best solution will be a combination of these.
The resolution that even the best automotive radar can achieve is only as good as the worst lidar systems available, says Matthew Weed, engineer and senior director of product management at Luminar, which manufactures lidar systems for automobiles. Luminar’s system, which Mr. Weed says is superior to radar for most applications, costs $ 1,000, however.
Mr Weed says Luminar’s lidar systems could justify their cost by being so good that they could lower driver insurance costs by preventing pedestrian accidents and fatalities. Even with such a system on a car, radar would be a good back-up when it breaks down or cannot handle inclement weather, he adds.
Mobileye uses lidar, cameras and radar in its most advanced systems. CEO Amnon Shashua said that even though the price of lidar systems has come down, their cost is still 10 times higher than radar, and likely will remain so for the foreseeable future, due to the complexity of the hardware involved.
Elon Musk’s Tesla has made a bet that the company can achieve true autonomous driving in its vehicles using only cameras.
Cameras have the advantage of extremely high resolution, and they are affordable and compact thanks to years of advancements in smartphone cameras. But for a system that can achieve the highest safety standards, and even full battery life, cameras need backup sensors that fail under conditions different from those they encounter, Dagan adds.
Take the fog, which looks like a hindrance to camera-based and lidar systems, potentially causing vehicles to stop when they shouldn’t. In research published in 2020, radar-based automotive sensors had no difficulty penetrating fog and correctly identifying stopped vehicles hidden inside, says Dinesh Bharadia, assistant professor of engineering at the University. from California to San Diego who contributed to the work.
Dr Bharadia says his team discovered that a key uses multiple radars, spaced at least five feet apart on a vehicle. It’s the same principle at work in the ever-increasing number of cameras on the back of our smartphones, he adds. It is possible to create an “image” of the surroundings of a car using several low-cost radar sensors, just as our phone can use several inexpensive small cameras and then recombine the images they collect. into something much sharper.
Bringing together all of a car’s sensors into one cohesive view of the reality outside a vehicle requires merging all that data, says Rick Gentile, an engineer who worked on radar systems for defense applications and who is now a product manager at MathWorks, a software company that creates tools to help process data. For example, although the radar can detect the presence of a sign in front, it cannot see its color, which is essential to quickly identify what type of sign it is.
For so-called robot taxis, the way to fill the gaps in the capabilities of each type of sensor is to use them all. The goal is âfull redundancy,â says Dagan, so that even if one sensor has an error, the others perceive the world correctly. It’s the fastest way, he says, to give vehicles senses that are at least as good as a human. (Whether these vehicles will have sufficient judgment to move safely is a separate issue.)
Until we get real self-driving vehicles – something that could take years, if not decades – automakers will have to choose between radar, lidar, and cameras, or a combination of all three, to create tracking systems. safety capable of keeping their promises to make automatic braking systems standard by 2022, and to continue to improve these systems. All three types of sensors continue to improve, but the difference in cost between them has led automakers to favor one technology or another, depending on how they think they can make up for its software and AI shortcomings.
This has led to healthy competition among manufacturers of security systems, sensors, and supporting software – who you talk to, they argue their systems are the best.
As all of these companies scramble for a place on your car, the goal of all of these technologists is to profit by dramatically reducing the number of road fatalities of all kinds when a human is behind the wheel. It’s a goal they all agree is much closer than fully autonomous vehicles.
Never miss a story! Stay connected and informed with Mint. Download our app now !!