We asked three automotive sensors and data experts to name what technologies are going to make cars more connected. These are their answers.
Cars are now true innovation hotspots, where cutting-edge technologies to improve driving and travel experiences are being rapidly developed. Cars are becoming more and more like mobile computational units equipped with sensors detecting, tracking, and sharing information at every moment. Like many other objects – home appliances, for example – cars are increasingly connected to the surrounding environment. They use artificial intelligence (AI), they can exchange information with the road, and with other vehicles too.
If the cars of the future will be more and more connected, it will be thanks to the technologies being developed today. On January 23, theHigh Tech Campus Eindhoven organized a seminar on the future of connected cars, giving attendees a sneak preview of what the future of the connected car will look like. Innovation Origins visited the conference and asked the three panelists to pick a technology that will define tomorrow’s connected cars.
Since Chat GPT’s boom, more people are becoming acquainted with the capabilities of AI. Chat GPT is an example of conventional cloud-based AI, meaning that computing happens in servers far from the device. What makes edge AI different from conventional AI? Edge AI computes in the device itself which allows for reduced latency, costs, and power usage. Amongst the downsides, however, are lower memory and computing capabilities.
Axelera AI is focusing on AI at the edge, targeting applications such as automotive ones. Bram Verhoef is head of machine learning at Axelera AI and believes artificial intelligence will represent the future of cars, too.
“To have fully autonomous cars, smart AI needs to be onboard. It is essential not only to interpret sensors’ information but also to predict and plan how to manoeuvre the vehicle in an optimal way. In doing so, you cannot rely on an internet connection, because fast and reliable processing is needed. Therefore, all the computing ought to be performed inside the car.”
To get to that point, many problems have to be solved. Yet, Verhoef believes that the first realm to be taken over by AI in cars is sensor processing. “Then there is also reinforcement learning for planning, which is based on all the sensor information,” he adds. Founded in 2019, Axelera AI is launching its first products on the market this year, with chips that could be the “brains” of future autonomous cars.
Radar technology might seem like nothing new for the automobile industry. Several object detection and collision systems, such as adaptive cruise control technology, use radar. However, further developments in this technology will make it even more powerful.
NXP semiconductors recently introduced a new radarchip that enables reliable long-range detection and separation between larger and smaller objects. What’s more, it comes in a smaller size than its predecessors, allowing for smaller sensing modules.
“Radar helps increase the perception of the surrounding world. With respect to other technologies, it is stronger in adverse weather conditions,” underlines Lulu Chan, technical domain lead for ADAS/Radar in automotive system innovation at NXP.
In her vision, radar is going to be the go-to technology, at least in the short term. “Radar is a technology that can give a lot of information about the surrounding environment at a low cost for a sensor. It can also have quite a small form factor. For autonomous driving, multiple sensors technologies are needed, radar ones included.”
On chip LiDAR
According to Christian A. Bachmann, Program Director Wireless Sensing at imec The Netherlands, cost-effective silicon-photonics LiDAR technology will further improve automotive sensing capabilities. LiDAR stands for Light Detection and Ranging but compared to radar, LiDAR utilizes light instead of radio waves. This technology allows mapping surrounding spaces to create 3D models. Imec, one of the world’s leading R&D centers in nanoelectronics and digital technology, is also conducting research on LiDAR.
“LiDAR is coming to the first production cars to enhance safety and to provide a better perception of the surroundings. This technology provides superior 3D mapping performance since it enables very high resolutions. Therefore, it can detect small objects – such as a piece of debris, or a car tire - on the road up to 300 meters in front of the car. LiDAR can make a real difference and improve safety in autonomous driving assistance systems, as well as enable future autonomous use cases,” says Bachmann.
However, LiDAR is still significantly more expensive than radar technology. Silicon photonics could play a crucial role in bringing costs down. In other words, the techniques used in the electronic chip industry can also help miniaturize LiDAR applications. “Integrating laser, electronics, and photonics components using microchip technology will lead to cost-efficient LiDAR solutions,” Bachmann emphasizes.
Connectivity at its finest
On the way to smarter, autonomous and connected cars, sensory technology developments will lead the path. This way, cars will have a clearer vision of the environment surrounding them. This sensory information is what will ultimately make them drive autonomously – yet safely. LiDAR, radar, and AI will all contribute to reshape cars as we know them.