New Telemetry Paper Examines How AI Reasoning Models Enable Higher-Resolution, More Reliable Radar Perception Across Automotive, Aviation, Intelligent Transportation Systems (ITS), Robotics, and Security Applications
-- Atomathic, formerly Neural Propulsion Systems and a pioneer in physical AI-sensing technology, today announced the publication of a new white paper “Improving Radar Performance for Physical AI Systems” by Telemetry.
The report explores how AI reasoning models can dramatically enhance radar performance by overcoming long-standing challenges, such as sidelobes, ghost reflections, low dynamic range, and target flicker. The report also details how a multi-hypothesis, physics-grounded reasoning approach enables radar systems to extract significantly more usable information from existing sensors without requiring prohibitively expensive hardware upgrades.
“Physical AI systems depend on accurate, reliable perception to operate safely,” said Sam Abuelsamid, Vice President, Market Research, Telemetry and report author. “Traditional radar processing discards large portions of uncertain data, resulting in sparse environmental models. By applying physics-grounded reasoning and evaluating multiple hypotheses across frames, it’s possible to more accurately confirm genuine reflections, reduce ambiguity, and generate a much more robust model of the world even with today’s mainstream automotive radar hardware.”
Radar’s Expanding Role in Physical AI
The white paper places enhanced radar processing in the broader context of accelerating adoption of advanced driver assistance systems (ADAS) and automated driving systems (ADS). According to Telemetry’s 2025 ADAS and ADS Global Forecast, nearly 65% of new light-duty vehicles sold by 2030 are expected to include hands-off/eyes-on Level 2+ functionality, with additional growth in Level 3 systems.
Vehicles becoming larger and heavier are increasing risk to vulnerable road users and the need for reliable detection and classification becomes more urgent. Radar plays a critical role because it operates reliably in low light, glare, fog, rain, and other atmospheric conditions that degrade cameras and lidar.
However, conventional radar processing pipelines rely on filtering and confidence thresholds that often discard ambiguous reflections. The result is low-resolution, sparse point clouds that struggle to distinguish pedestrians, cyclists, and vehicles in close proximity.
The white paper explains how AI reasoning models can address these deficiencies by:
- Generating multiple physics-constrained hypotheses for radar reflections
- Comparing patterns across sequential frames to confirm consistent targets
- Leveraging material reflectivity characteristics to distinguish objects
- Preserving more valid data instead of discarding uncertain returns
Through this approach, even widely deployed 16-channel automotive radar sensors can achieve closer to lidar-like environmental modeling at approximately 0.3-degree angular resolution — a substantial improvement over traditional processing.
Abuelsamid also analyzes the compute implications of reasoning-enhanced radar. While AI-based radar processing may require five to six times the compute resources of traditional radar signal processing, it remains two to three orders of magnitude lower than the requirements of a single automotive camera stream.
While the report focuses heavily on automotive applications, it also outlines implications for:
- e-VTOL aircraft and automated aviation systems
- Mobile industrial robotics operating in lights-out environments
- Indoor and outdoor security systems
- Broader physical AI deployments requiring reliable sensing in adverse conditions
- Intelligent Transportation Systems (ITS)
The white paper concludes: “Atomathic makes radar compute-complete. By pairing AIDAR (a physics-informed sparse autoencoder) with AISIR (a generative reasoning engine), the company delivers the ‘Seventh Sense’ for autonomy: a perception stack that is as precise as lidar, as robust as a camera, and stable enough to serve as a key safety layer for ADAS and Level 4 autonomous vehicles.”
About Atomathic
Atomathic is a pioneering physical AI-sensing technology company transforming how machines perceive and interpret complex signals in the real world. Leveraging deep expertise in advanced mathematics and proprietary AI platforms—including AIDAR™ for detection and ranging, and AISIR for Radar™ for signal intelligence reasoning—Atomathic delivers hyper-resolution sensing that enables sensors and systems to detect, interpret, and visualize ultra-high-resolution signals in real time. Atomathic’s technology is hardware-agnostic and applicable across automotive, aviation, defense, robotics, and semiconductor markets. By grounding inference in physical principles and scalable compute, Atomathic helps enable safer autonomous decision-making and intelligent machines. Atomathic can be found on the Web and LinkedIn.
###
All product and company names may be trademarks or registered trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.
Contact Info:
Name: Gary Bird
Email: Send Email
Organization: 43PR
Address: 1537 Pacific Ave., Ste 300-C, Santa Cruz, CA 95060
Phone: 831.888.9011
Website: https://www.43pr.com/
Release ID: 89186164
In the event of encountering any errors, concerns, or inconsistencies within the content shared in this press release, we kindly request that you immediately contact us at error@releasecontact.com (it is important to note that this email is the authorized channel for such matters, sending multiple emails to multiple addresses does not necessarily help expedite your request). Our dedicated team will be readily accessible to address your feedback within 8 hours and take appropriate measures to rectify any identified issues or facilitate press release takedowns. Ensuring accuracy and reliability are central to our commitment.
