Atomathic, a radar technology company is making a bold claim: that long-standing safety blind spots around large vehicles aren’t a hardware problem — they’re a software one.

It says it stumbled across this discovery, practically by accident, in recent weeks by taking a physics approach to solving radar’s shortcomings.

At a briefing and live demo for truck and automotive press at the House of Journalists ahead of CES in Las Vegas, the company said it has developed a physics-based radar processing system that can detect weak objects, such as pedestrians. That includes when they’re standing beside large metal vehicles like trucks — a scenario where conventional automotive radar often fails.

Radar has been used in vehicles for decades and performs well in rain, snow, and darkness. However, it’s limitations have become more apparent as automakers pushed it into advanced driver assistance systems and automated driving roles, according to Dr. Behrooz Rezvani, CEO and founder of the company.

Large reflective objects can overwhelm the sensor, creating ghost targets or masking smaller ones entirely. The company argues that today’s radar systems rely too heavily on single “snapshots” of data and aggressive filtering to suppress false positives — a trade-off that can also erase real but weaker targets.

Its approach instead generates multiple hypotheses from each radar frame and tests them over time using physics-based models of how radar signals propagate and reflect. The result, the company says, is a radar system that can see a pedestrian walking beside a truck at night — even when the pedestrian’s signal is tens of thousands of times weaker than the truck’s return.

In side-by-side demos, the company showed a conventional radar system failing to detect a pedestrian until it was too late to be useful, while its software maintained a stable track despite the nearby truck dominating the scene.

The technology is software-only and is designed to run on existing radar hardware, provided automakers allow access to raw sensor data. While it currently requires more computing power than today’s radar processing, the company said the load remains modest compared to camera-based vision systems and should fit easily into next-generation centralized vehicle architectures.

For trucking, the implications could be significant. Large trucks represent some of the most challenging environments for radar, particularly in urban settings where pedestrians and cyclists operate close to reflective trailers and cabs. It also address the complexities for radar in recognizing things like e-bikes, Rezvani added.

The company cautioned that the technology is still early and has not yet been optimized for production use. But if its claims hold up in broader testing, radar — long viewed as a supporting sensor — could regain relevance as a primary safety tool around commercial vehicles.

“It makes the invisible visible,” summed up Dr. Lawrence Burns, executive advisor to the company.