Leveraging AI in Radar to Achieve True Safety

December 31, 2023
Read Blog

How our Perception Radar leverages the power of AI processing across various applications to enhance the capabilities of modern vehicles in multiple dimensions.

The Evolution of Radar

Radar is not a new technology. It emerged during World War II as a military technology for airplanes and since then has undergone a major transformation; the military radar of today differs significantly from its original form. The first notable use of radar in automotive was in 1998: Mercedes-Benz S-Class was the first vehicle to debut radar technology for enabling adaptive cruise control features.

Benefits of Radar Technology

Radar technology is unique in its ability to directly measure the distance and relative velocity between two objects – it detects range, depth. Because of this, it can easily calculate whether a vehicle should proceed, accelerate, or brake based on the velocity of the object it is detecting, and then send an alert notifying the driver. These advantages are the primary role of the radar as part of the sensor suite. Radar is also an affordable sensor and can offer redundancy to camera technology, which has different strengths, making the vehicle more accurate and safer.

Moving Towards Autonomy

When moving in the direction of autonomy, alerts no longer suffice: the vehicle needs to be able to make decisions. To do so, a vehicle needs to have sensors that achieve the highest level of detection and accuracy to truly understand the driving scenario. That means the sensor needs to be able to “see” objects precisely during the day and night, in fog and snow, and in every other complicated weather condition at both long and short range.

Traditional Radar Versus Perception Radar

Radar inherently has a lower resolution than camera technology. Traditional radars cannot play a major role in advanced driving applications without first improving resolution to detect smaller objects like a child crossing the street, a tire on the road and other challenging use cases.
In MIMO technology, the overall spatial resolution is proportional to the number of virtual antennas, also known as the radar’s channel count. Therefore, by increasing the channel count from the typical 192 to 2,300 channels (over 10x improvement) Arbe’s technology promises an equivalent improvement in resolution, making its radar a perception-level sensor.

Camera vs. Radar

Camera is not a competing sensor to radar but a complimentary one, since each has orthogonal strengths and weaknesses. A vision- only approach can’t solve the challenge of autonomy. The sensor suite needs true redundancy and therefore needs both sensors. Today, the perfect “sweet spot” is for vehicles to incorporate multiple sensors, providing the necessary redundancy for true safety. A sensor fusion approach will ultimately be the safest and most powerful approach to full autonomy.

How We Leverage AI in Perception Radar Processing

The Perception Radar converts analog raw data into a digital data set for processing, employing its exceptional long-range detection capabilities, direct distance and relative velocity measurement, and the recognition of connections between frames.

The power of AI is harnessed to optimize performance, employing advanced features such as multipath suppression to eliminate false alarms, thereby elevating image reliability. Notably, the Perception Radar employs clustering to group point cloud detections into coherent objects, ensuring accurate tracking. This innovative radar technology analyzes the object’s speed and direction, providing the vehicle with a clear understanding of an object’s current location and its anticipated position, in fractions-of-a-second accuracy. Remarkably, even in scenarios with zero connectivity, the Perception Radar can estimate the vehicle’s velocity and position in the mapped environment.

One of the distinctive capabilities of the Perception Radar is its achievement of comprehensive 360° perception through intelligent data sharing among multiple radars strategically placed at the front, rear, corners, and sides of the vehicle. The integration of these radars ensures a seamless overlap in the fringes, enabling smooth tracking of objects of interest from one radar to the next. This redundancy is further reinforced by the application of two distinct perception algorithms, validating the location of the same object through a meticulous and robust process. Ultimately, Arbe’s Perception Radar emerges as a cutting-edge solution, enhancing not only the reliability of data but also the vehicle’s awareness and responsiveness in diverse driving scenarios.

Perception Radar leverages the power of AI processing across various applications, enhancing the capabilities of modern vehicles in multiple dimensions.

  • Object Tracking: The radar precisely determines an object’s location in relation to the vehicle, comprehending vital parameters such as speed, size, shape, altitude, orientation, and turn rate. This sophisticated tracking functionality is instrumental in powering advanced safety applications, contributing to a comprehensive understanding of the road environment and facilitating safer navigation.
  • Target List for Fusion: a detailed target list that includes spatial coordinates, object velocities, and a confidence parameter for each target. This comprehensive information aids in seamlessly fusing data for a more holistic perception of the vehicle’s surroundings.
  • Classification: Utilizing advanced classification algorithms, the radar AI is used to distinguish between various objects, identifying whether they are pedestrians, vehicles, two-wheelers, or larger vehicles such as trucks and buses.
  • Free Space Mapping: A critical feature lies in the radar’s ability to create a detailed map of free and occupied spaces. This map is indispensable for decision-making algorithms involved in path planning, providing the vehicle with crucial information about the availability of space for navigation.
  • Path Planning: The AI-driven capabilities of the Perception Radar extend to path planning, relying on its ability to predict the future location of objects. This forward-looking approach enhances the vehicle’s ability to navigate effectively and make informed decisions about its trajectory.
  • SLAM (Simultaneous Localization and Mapping): The radar not only detects obstacles but also performs Simultaneous Localization and Mapping, continuously generating and updating a map of the vehicle’s surroundings. This dual functionality contributes to precise navigation by localizing the vehicle within the evolving map in real-time.

Perception Radar emerges as a comprehensive solution, leveraging AI to empower vehicles with advanced perception and decision-making capabilities across a spectrum of critical applications.

————————————

 

 

This blog contains “forward-looking statements” within the meaning of the Securities Act of 1933 and the Securities Exchange Act of 1934, both as amended by the Private Securities Litigation Reform Act of 1995. The words “expect,” “believe,” “estimate,” “intend,” “plan,” “anticipate,” “may,” “should,” “strategy,” “future,” “will,” “project,” “potential” and similar expressions indicate forward-looking statements. Forward-looking statements are predictions, projections and other statements about future events that are based on current expectations and assumptions and, as a result, are subject to risks and uncertainties, including the risk that the proposed regulation will be adopted in a form which increases to market for Arbe’s radar sensors,  that other companies may offer products which are less expensive or more attractive to the market, that if the regulation is adopted, the effective date will be in the distant future and may not have any significant effect on market for Arbe’s products and the risk and uncertainties described in “Cautionary Note Regarding Forward-Looking Statements,” “Item 5. Operating and Financial Review and Prospects” and Item 3. Key Information –Risk Factors” Arbe’s Annual Report on Form 20-F/A for the year ended December 31, 2022, which was filed with the Securities and Exchange Commission on May 16, 2023 as well as other documents filed by Arbe with the SEC. Accordingly, you are cautioned not to place undue reliance on these forward-looking statements. Forward-looking statements relate only to the date they were made, and Arbe does not undertake any obligation to update forward-looking statements to reflect events or circumstances after the date they were made except as required by law or applicable regulation. Information contained on, or that can be accessed through, Arbe’s website or any other website is expressly not incorporated by reference into and is not a part of this press release.

Connect to learn more
Physical AI’s Real Bottleneck Is Perception
June 8, 2026

Physical AI’s Real Bottleneck Is Perception

Why ‘Good Enough’ Detection Fails in the Mine and on the Farm
April 27, 2026

Why ‘Good Enough’ Detection Fails in the Mine and on the Farm

Level 3 Autonomy: The Next Leap in Driving Confidence
November 10, 2025

Level 3 Autonomy: The Next Leap in Driving Confidence

Phantom-Free Radar
August 11, 2025

Phantom-Free Radar

Mastering 20 FPS Performance
July 14, 2025

Mastering 20 FPS Performance

Extending the Safety Horizon: Where Every Second Counts
May 26, 2025

Extending the Safety Horizon: Where Every Second Counts