If an autonomous future is to become a reality, it has to be achieved with a driverless vehicle able to operate safely without connectivity.
That’s the view of Carl Jackson, head of corporate strategy at Provizio, a specialist in advanced radar technology. He also argues an automaker that does not deploy sophisticated imaging sensors beyond the use of cameras will not achieve robust performance, even in advanced driver-assistance systems ahead of full autonomous-driving capabilities.
His view on connectivity is informed with his company’s experience in providing its 5D radar sensor solution, first presented at CES 2023, to the mining industry. He points out that an automated driving vehicle working deep underground does so with no connection to GPS or the Internet of Things.
Jackson explains: “In mining, these vehicles work autonomously underground with no access to GPS and they need lidar but also need the robustness of radar. Also, there will be huge dust clouds generated by other vehicles that even lidar can’t see through but radar can, spotting people and vehicles when cameras are obscured with dirt.
“That experience will be shared by the automakers who see that the cost point of 5D radar will persuade them to put them on the vehicle for just those occasions that connectivity is not available.”
In terms of automakers’ sensor choice, he believes the argument that cost restricts the use of anything other than camera-based systems does not now hold true for radar sensors which are becoming increasingly compact and affordable. Jackson points to the case made by Tesla, saying: “Tesla wouldn’t put lidar in its cars simply because of cost and rely on cameras instead. However, I would now argue you can’t afford not to put sophisticated sensors in your cars, especially for high-quality imagining.
“High-quality 5D radar has an advantage in that its cost is minuscule compared to that of lidar. Compared to cameras, radar gets you the depth of a scene, the velocity of the vehicle and the trajectory of the other vehicle in the picture. So, it has to go into the car and I’m very uncomfortable with the strategy to only use cameras.”
While Jackson accepts full driverless technology is some way off, the expansion of ADAS features will pave the way toward that future. He explains: “Yes, I absolutely believe there will be an autonomous future but it’s not going to happen overnight. The route to autonomy will be through ADAS that will give better safety in vehicles. I think that is the most important thing and that’s where we focus our customers on. As the market matures, as the processing power of vehicles increases significantly, sensors are improving and the business model of putting sensors in cars, which will become mandated in the future as with seat belts.
“However, a lot of the technology will come in because customers want it. We are seeing customers are willing to pay for safety features. One of the great things Tesla has done is the subscription model with over-the-air upgrades.”
Jackson thinks that as the consumer experiences the safety and convenience features of increased implementation of ADAS, the demand for the technology will also accelerate its adoption by automakers.
He says: “The OEMs are all seeing this drive and acceptance from customers. They can pay for a service or in packages, and I think the old model of where an automaker would purchase a box from a Tier One supplier where the item would just do that job for 10 years or so are over.
“It’s more about when you have more computing in the vehicle getting cheaper and cheaper, you will find the ability for OEMs to take the data from the car, interact with the vehicle and update sensors, algorithms, all over-the-air. That’s something customers will want to pay for. Manufacturers are already talking about dropping in Nvidia Drive platforms into their vehicles. All of a sudden you have a supercomputer in the back of the car.”
