【Summary】Recogni Inc., a Silicon Valley startup developing a vision-oriented artificial intelligence platform for autonomous vehicles announced today it raised $25 million in Series A funding. The company is focused on creating high-performance and low-power consuming AI edge processing for autonomous vehicles.
Recogni Inc., a Silicon Valley startup developing a vision-oriented artificial intelligence platform for autonomous vehicles announced today it raised $25 million in Series A funding. Invetors include the venture capital arms of automakers BMW and Toyota.
Recogni is focused on creating high-performance and low-power consuming AI edge processing for autonomous vehicles. The company’s founders have extensive experience in system design, AI, computer vision, and custom silicon design.
Recogni was founded in 2017 and is headquartered in San Jose, Calif. and also has operations in Munich, Germany.
Participating in the $25 million Series A funding round was BMW i Ventures, Toyota AI Ventures, DNS Capital, Fluxnit, the VC arm of automotive lighting company OSRAM, and automotive technology company Faurecia.
Recogni aims to revolutionize perception processing for Level 2+ autonomous vehicles, which requires exceptionally high real-time processing performance while consuming little power. Recogni’s perception technology can identify bikes and pedestrians, as well as their location in real-time, much faster and from further away than any of its competitor in the market, the company claims.
Recogni designed a “Vision Cognition Processor”, which is an artificial intelligence platform for autonomous vehicles.
One of challenges for self-driving cars is the ability to process AI and machine-learning algorithms without needing a trunk full of expensive, power-hungry hardware. Recogni is focused on creating high-performance and low-power AI processing solution for autonomous vehicles by solving the endpoint inferencing problem with autonomous vehicles more efficiently.
Although most AI-based machine learning systems are currently trained offline, for autonomous vehicles, there is a need to process sensor data in real-time while navigating, which requires robust computing power and hardware. As the automotive industry transitions to autonomous vehicles, there is a growing need to drive these vehicles using more efficient, low-power processors.
“The issues within the Level 2+, 3, 4 and 5 autonomy ecosystem range from capturing/generating training data to inferring in real-time. These vehicles need data center class performance while consuming minuscule amounts of power,” said RK Anand, CEO of Recogni. “Leveraging our background in machine learning, computer vision, silicon, and system design, we are engineering a fundamentally new system that benefits the auto industry with very high efficiency at the lowest power consumption.” Edge Processing for Self-driving Cars Recogni believes that the autonomous vehicles today have hit the processing efficiency wall and are unable to transition to Level 3 and 4 autonomy without utilizing edge processing for a vehicle’s AI-based perception software.
The company claims that its system will deliver unprecedented inference performance at more than 500 times better power efficiency compared to other solutions. This increase in performance enables better edge processing.
Most of the artificial intelligence processing from tech companies like Google, Nvidia, Microsoft and Amazon perform AI inference on the cloud using powerful, power-hungry processors. The problem with using AI processing in the cloud for autonomous driving decisions are issues with speed and latency. Using edge processing allows autonomous vehicles to make driving decisions faster than humans while consuming minimal amounts of energy.
Unlike cloud computing, where data is uploaded to the cloud directly from an autonomous vehicle for additional processing, edge-processing takes place closer to where the data is being generated by a self-driving car’s sensors, which in this case is right from the vehicle.
In addition, the large amounts of data generated by a self-driving car’s sensors is often too large to send to the cloud for efficient processing.
“The ability to process sensor data on the edge efficiently and in real-time is essential in the development of autonomous vehicles,” said Marcus Behrendt, partner BMW i Ventures, a participant the investment round. “We believe that Recogni has the right approach and an experienced team to help solve these critical issues as the automotive industry continues on its path towards semi-autonomous and fully autonomous and vehicles.”
Recogni plans to use its new funding to improve its inferencing system to enable state of the art sensor fusion of vision and depth sensor data for Level 2 self-driving technologies that are already being used by automakers. The company also plans to expand its engineering team.