U.S.-German startup Recogni will launch with a $25 million war chest raised in Series A funding to make “fully-autonomous vehicles a reality.” This is thanks to the company’s vision-oriented artificial intelligence platform, as Robotics Automation and News reports.
Recogni—which stands for Realtime Object Recognition—has been in “stealth mode” while they raise money and develop their “high-performance, low-power” AI system. Great-Point Ventures led the funding round which also included contributions from Toyota AI Ventures; BMW’s i Ventures; automotive technology company Faurecia; Fluxunit, the venture capital arm of photonics and lighting company Osram; and DNS Capital.
The automotive industry is making a strong move toward autonomous vehicles. To drive this revolution, a network of computers is integral in operating the cars efficiently on a conservative energy budget. The AIs receive their training offline. However, they need to process sensor data in real-time as the car cruises down the road.
But many of these autonomous vehicle systems have come to an impasse. International engineering association SAE has set forth six levels of autonomy. The spectrum runs from Level 0 (no autonomy) to Level 5 (full autonomy). But what’s called the “processing efficiency wall” has prevented current systems from achieving Level 3 and beyond.
“While scoping the market, we realized that most of the neural network accelerator technologies are either optimized for performance or power, none are optimized for both,” Ashok Krishnamurthi, managing partner at GreatPoint Ventures says in a statement.
Changing the Game
Recogni plans to change the game with their “Vision Cognition System.” The new tech utilizes a suite of image sensors to pick out small objects like traffic lights, bikers, and pedestrians at more than 200 meters away in real-time. In addition, Recogni can discern the difference between a red, yellow or green traffic light.
The other revolutionary aspect of Recogni is its efficiency. Current systems eat up tens of kilowatts in data processing. Recogni claims that its system can process the same amount of information with less than 100 watts. With this innovative new technology, Recogni aims to take autonomous vehicles into Level 3 and beyond.
Founded in 2017, Recogni’s team consists of a diverse group of experts in a wide range of fields including vision science, AI, software, silicon design, and more. The startup plans to use their funds to expand on its world-class team and lead the autonomous vehicle industry into the new decade.
The company is currently in correspondence with a number of automotive manufacturers. They hope to provide the industry with the systems it needs, from modules to software. With the new tech, Recogni envisions robotic taxis by 2024. Next up, the company is hopeful that personally-owned autonomous vehicles will follow a few years later.
With so much news of autonomous vehicle foibles popping up in newsfeeds, it’s refreshing to see a company that has a vision to safely move the technology towards the realm of reality.