The Role of Radar in Autonomous Vehicle Technology

Self-driving cars depend on a multitude of sensors to understand their surroundings, and while cameras and lidar frequently receive significant attention, traditional radar remains a vital component – despite inherent limitations. Oculii, having recently secured $55 million in funding, is focused on mitigating these drawbacks and enhancing radar capabilities through a sophisticated software layer compatible with current devices, alongside the development of its own hardware.

Radar's Strengths and Weaknesses

Radar excels in its extended range and its ability to penetrate obstacles like rain, snow, and fog – making it indispensable for environmental perception during adverse weather conditions. Both lidar and conventional cameras can be rendered ineffective by these common occurrences, necessitating a reliable backup system.

However, a primary disadvantage of radar lies in its inability to create detailed images comparable to lidar, due to the nature of its wavelengths and antenna operation. The data typically consists of precisely located points rather than intricate shapes. While still providing valuable data within a sensor suite, improvements to its scanning fidelity would be highly beneficial.

Oculii's Approach to Radar Enhancement

Oculii aims to significantly improve existing radar performance. The company asserts a 100-fold increase in spatial resolution achieved by implementing its software to control the system. According to co-founder and CEO Steven Hong, a conventional radar with a 120-degree field of view might offer a 10-degree spatial resolution, limiting precision to a few degrees and providing minimal elevation data.

This equates to a relatively low 12x1 resolution.

By utilizing the Oculii system, which dynamically adjusts transmissions based on real-time perception, this resolution could be elevated to 0.5° horizontal x 1° vertical, potentially achieving a 120x10 resolution. (These figures are illustrative and not definitive system specifications.)

This substantial improvement enables the system to differentiate between closely spaced objects, discern size differences, and – with further processing – determine an object’s velocity relative to the radar unit.

Demonstration and Technology

Here’s a video demonstration of one of their own devices, showing considerably more detail than one would expect:

The precise methodology behind this enhancement remains proprietary to Oculii. Hong explained that their sensor employs AI to generate an “intelligent” waveform that adapts to the environment, embedding information over time to significantly improve resolution. This temporal integration is the basis for the “4D” designation.

Here’s a little sizzle reel that gives a very general idea:

Impact on the Autonomous Vehicle Industry

While a standardized sensor configuration for autonomous vehicles has yet to emerge, Oculii’s technology could elevate radar’s prominence. Currently, radar is often limited to emergency braking detection. Enhanced detail and data could allow radar to play a more substantial role in AV decision-making processes.

The company is actively forging partnerships with Tier-1 suppliers and OEMs, including investor Hella, demonstrating confidence in its approach. It is also collaborating with radar manufacturers and has secured commercial contracts with a projected timeline of 2024-2025.

Image Credits: Oculii

Developing Proprietary Radar Hardware

Oculii is also venturing into the development of all-in-one radar units, integrating hardware and software for synergistic performance. They claim these units represent the world’s highest-resolution radars, a claim currently unchallenged, as radar competition primarily focuses on rangefinding and speed detection precision rather than resolution.

Echodyne, utilizing a metamaterial radar surface for customizable beam direction, presents a potential exception. However, even its “resolution” is difficult to quantify.

The company’s new Eagle and Falcon radars may appeal to manufacturers developing advanced sensing suites for autonomous research or production driver-assist systems.

Investment and Future Outlook

Given radar’s anticipated importance in autonomous vehicles, robotics, aircraft, and other applications, substantial investment in the field is warranted. The $55 million Series B funding round underscores this point. The round was co-led by Catapult Ventures and Conductive Ventures, with participation from Taiwania Capital, Susquehanna Investment Group (SIG), HELLA Ventures, PHI-Zoyi Capital, R7 Partners, VectoIQ, ACVC Partners, Mesh Ventures, Schox Ventures, and Signature Bank.

These funds will facilitate scaling operations, expanding the team, and continuing to refine the technology to deliver higher resolution, longer range, more compact, and more affordable sensors, ultimately accelerating the advancement of autonomous systems.