Get one step closer to the autonomous car with the radar of the future
Obstacles on the road to the autonomous car, which were once swept away with an optimistic hand wave at the speeches of manufacturers and equipment manufacturers, today resemble impenetrable walls. Recently, Luc Julia, an artificial intelligence specialist at Renault, was completely pessimistic. “A fully autonomous car will never exist. I can demonstrate this mathematically.” We are far from the promises of Ford, which announced in 2016 that the car without pedals and steering wheel will be marketed in 2021… We are still waiting!
It must be said that there is always an obstacle in the way. First, there is a European regulation that limits the speed to 60 km/h in autonomous driving mode. And technical obstacles, whether related to software (it is impossible to foresee all situations that may arise during a trip) or sensors. After that, announcement effects gave way to wisdom. Progress is incremental, step-by-step, across five defined levels of automation. The next step, which is just about to surface, will be to allow cars to be driven without drivers in perfectly marked environments (usually highways that are properly mapped and road signs are impeccable). This corresponds to level 3.
Expensive and weather affected lidars
Some brands, starting with Mercedes-Benz, already offer Level 3 automated systems in markets where they are legally possible. But there are still many pitfalls, starting with the sensors. Cameras (currently the only means of reading the road on a Tesla) can easily be disturbed by bad weather… Or mosquitoes! As for the radars, they suffer from average definition. Lidar remains, more precisely. However, these are currently very expensive, difficult to integrate into the housing, and because they use laser beams, they are sensitive to climate conditions.
French startup Greenervawe seems to have squared the circle with what it calls 4D radars. Young shooting has been working on the development of its technology for fifteen years, which ensures that it is not more expensive than existing radars. The project is credited with finding industrial Plastic Omnium as a partner. For several years, the equipment manufacturer has been trying to move beyond its status as a simple supplier of plastic parts to become a first-class equipment supplier, capable of producing complete subassemblies, with technologies that contribute to the performance of the job. car. Thus, Greenervawe’s 4D radars will be part of the front panel that integrates the various sensors. Marketing is expected by 2026 or 2027.
Multiply the receiving antennas
More than a technological revolution in hardware, Greenerwave’s technology is on the way to data processing that intends to make a difference. “Today, the sensors are packed into a small surface (often hidden behind a logo in the middle of the grill or a smooth plate, editor’s note),” explains Geoffroy Lerosey, co-founder of Greenerwave. “By concentrating transmitters and above all receivers on a small surface, it is difficult to achieve good resolution. That is why today’s onboard systems cannot distinguish between two objects separated by an angle of less than 1-2 degrees.”
Greenervawe therefore multiplies the receiving antennas over the entire width of the front face. “By using a wider reception zone, it is possible to better interpret the feedback from the waves due to the different angles of arrival at different antennas. This allows us to increase the resolution by a factor of ten, similar to the resolution of a lidar… Moderate additional costs. Indeed, signal processing the electronics are the same as a conventional system and the antennas are very reasonably priced.”
With their combination, Plastic Omnium and Greenerwave hope to significantly improve the performance of automated driving systems. At first, innovation will be reserved for high-end models, small series production is necessarily more expensive than existing systems. However, the ever-increasing demand for driver assistance systems, mainly under regulatory constraints, promises a bright future for this technology.
