After its successful $30M Series A round of funding, Innovusion has just announced the release of Cheetah, a new image-grade high resolution lidar sensor, in a bid to deliver the safest performance in autonomous vehicle mobility.
Featuring an innovative dual rotating polygon optical architecture, Cheetah enables Innovusion to claim that they have developed “the safest and most reliable detection available”. By combining their own detector electronics, advanced optics and well-developed algorithms, Innovusion has managed to surpass the capacity of other lidar systems with Cheetah. According to Innovusion’s Vice President of Sales and Marketing, Steve Ehrsam, the company came up with this architecture by taking different methods into account. They investigated such technologies as MEMS, Galvo, dual-Galvo, and Solid State, and got a thorough understanding of what building blocks were available to deliver within the 3-year time frame they set.
“The challenge of a polygon system is to scan x and y efficiently,” said Ehrsam, “Our founders’ vast experience in optical and mechanical allowed us to build such a device.”
Ehrsam described their definition of safety to have three components: accuracy, reaction time and reliability.
“Because of the unique architecture of our system, we can achieve remarkable performance in all three areas.”
Cheetah features a resolution of 300×300 (over the entire 100deg x 40deg field of view), which Ehrsam claims is two times that of the competition. This resolution allows the company to accurately identify objects and not “miss” critical ones. The sensor is capable of seeing small, low reflectivity objects as far away as 200m, and can see objects all the way out to 280m and receive multiple points back. This capability allows a customer’s perception systems more time to identify and react to potential road hazards, making the system safer.
“Due to our elegant and robust design utilizing mature technology, our system is very reliable and simple to build,” Ehrsam continued. “Reliability is a critical component to the safety of autonomous mobility systems.”
Headquartered in Los Altos, California, Innovusion was founded in 2016 by Junwei Bao and Yimin Li, who worked together at Baidu, a Chinese multinational technology company specializing in Internet-related services and products, and artificial intelligence. As the Head of Baidu’s Autonomous Driving Unit at the time, Junwei had the opportunity to evaluate lidar sensors in the market. Ehrsam explained that Junwei came to two major conclusions; that existing designs were expensive and that a sizable gap existed between what was available and what was needed (or what could be delivered).
For automous vehicles and automated driving, the Society of Automotive Engineers (AE) have defined five classification levels for vehicles, each defining an increasing amount of automation. As automation increases, especially after level 3, where driver involvement becomes minimal, safety is paramount.
“[Junwei found] that the technology being used wasn’t good enough to develop something for [SAE] level 3 or 4,” Ehrsam explained, “As a result, they developed their own system to be better and more efficient – which was the basis for Innovusion.”
In the future, the company plans to develop additional versions of the Cheetah sensor with increased resolution and flexibility, and is working towards partnerships with OEMs.
“We have seen a large upsurge in the installation of our products in the non-automotive space including autonomous trucking, autonomous construction equipment, autonomous mining equipment, high speed rail and V2X traffic monitoring systems for city infrastructure, all of which require the picture-like, long distance, high resolution and high reliability our system provides,” Ehrsam added.
“By demonstrating our progress and capability, the adoption of Innovusion’s high definition LiDAR by more industry leaders will build up confidence that level 3 and 4 is achievable in the near future.”
Innovusion’s Cheetah is available now with a single unit price of $35,000 for small quantities. To get a 360 degree field of view, multiple units are necessary.