Oculii appears to be like to supercharge radar for autonomy with $55M spherical B – .

Autonomous vehicles rely on a lot of sensors to sense the world around them, and while cameras and lidar get a lot of attention, good ol ‘radar is an important piece of the puzzle – although it has some basic limitations. Oculii, which has just raised a $ 55 million round, aims to minimize those limitations and make Radar more powerful with a smart software layer for existing devices – and sell its own too.

The advantages of radar lie in its superior range and the fact that its high frequency rays can pass through raindrops, snow and fog. This makes it vital to the perception of the environment in bad weather. Lidar and ordinary visible light cameras can be completely confused by these common occurrences. Therefore a backup is required.

The main disadvantage of radar, however, is that due to the wavelengths and the way the antennas work, it cannot map things in as much detail as lidar can. You will get very precisely localized blobs rather than detailed shapes. It still has invaluable features in a number of sensors, but if someone could add a bit more accuracy to their scans it would be a lot better.

This is exactly what Oculii does – take an ordinary radar and charge it up. The company claims to have improved spatial resolution 100 times by giving control of the system to its software. Co-founder and CEO Steven Hong stated in an email that a standard radar for a 120-degree field of view could have a spatial resolution of 10 degrees so that it can see where something is with an accuracy of a few degrees on either side. and little or no ability to determine the height of the object.

Some are better, some are worse, but for the purposes of this example that’s equivalent to an effective resolution of 12 × 1. Not good!

Handing over control to the Oculii system, which intelligently adjusts the transmissions based on what it is already perceiving, could resolve this to a horizontal resolution of 0.5 ° x 1 ° vertical, which is an effective resolution of maybe 120 × 10 results. (Again, these numbers are for explanatory purposes only and are not inherent in the system.)

This is a huge improvement and means that you can see that something is close to two objects, for example, and not one large, or that one object is smaller than another nearby, or that – with additional computation – it is moves in one way or another at this or that speed relative to the radar unit.

Here’s a video demonstration of one of their own devices that’s way more detailed than you’d expect:

How exactly this is done is part of Oculii’s proprietary magic, and Hong hasn’t gone into much detail on how exactly the system works. “Oculii’s sensor uses AI to adaptively create a ‘smart’ waveform that adapts to the environment and embeds information over time that can be used to significantly improve resolution,” he said. (Incidentally, the integration of information over time gives it the nickname “4D”.)

Here’s a little sizzling roll that gives a very general idea:

Autonomous vehicle manufacturers have yet to hit a canonical set of sensors that AVs should have, but something like Oculii could give radar a more prominent place – its limitations sometimes mean it is relegated to the frontline or in such a situation for emergency brake detection. With more detail and more data, radar could play a bigger role in AV decision making systems.

The company is definitely doing business – it works with tier-1s and OEMs, one of which (Hella) is an investor, which gives a sense of confidence in the Oculii approach. It also works with radar manufacturers and has some commercial contracts dealing with a 2024-2025 schedule.

Credit: oculii

It’s also about making your own all-in-one radars and leveraging hardware-software synergy. It is claimed that these are the highest resolution radars in the world and I don’t see any competitors to disagree. The simple fact is that radars don’t compete much for “resolution” but rather for the precision of their range measurement and speed detection.

An exception could be Echodyne, which uses a metamaterial radar surface to direct an adjustable radar beam anywhere in its field of view, examine objects in detail, or quickly scan the entire area. But even then, its “resolution” is not so easy to estimate.

In any case, the company’s new Eagle and Falcon radars could be tempting to manufacturers working to put together state-of-the-art sensor suites for their autonomous experiments or production driver assistance systems.

It is clear that it is worth investing seriously in space if radar is used as a major part of autonomous vehicles, robots, airplanes, and other devices. The $ 55MB round certainly shows that well enough. The Oculii press release states: “Led by Catapult Ventures and Conductive Ventures with the participation of Taiwania Capital, Susquehanna Investment Group (SIG), HELLA Ventures, PHI-Zoyi Capital, R7 Partners, VectoIQ and ACVC Partners. Mesh Ventures, Schox Ventures and Signature Bank. “

The money will enable the scaling and adjustment expected, and, as Hong added, “continues to invest in technology to provide higher resolution, longer range, more compact, and cheaper sensors that will accelerate an autonomous future.”

Comments are closed.