New Software Aims to Make Radar More Reliable and Safer for Commercial Vehicles
The system could have significant safety benefits for Class 8 trucks
Atomathic, a radar-software developer, said it has found a way to make radar more reliable in the kinds of edge cases that typically challenge driver-assist and automated-driving systems in commercial vehicles.
Large metal objects, such as tractor-trailers, can dominate radar returns, creating ghost targets or masking weaker signals, such as pedestrians, cyclists or debris, entirely, said Behrooz Rezvani, CEO and founder of Atomathic.
However, the reliability problem is primarily a software and inference challenge rather than an issue with the hardware, Rezvani explained. As a result, fixing the problem doesn’t require new sensors, just a new way of interpreting radar data.
Rezvani said Atomathic’s breakthrough came from reframing radar perception as a reasoning problem rather than a single-frame detection problem. Instead of making decisions based on one radar snapshot, the software generates multiple hypotheses from each incoming frame and evaluates them over time. A second, slower processing layer tests those hypotheses using physics-based models of how radar signals propagate and reflect, rejecting interpretations that violate physical constraints.
By generating multiple hypotheses per frame and validating them against physics-based expectations, Rezvani said Atomathic can isolate the weaker but real target without opening the door to noise and ghost objects.
Atomathic offered side-by-side demonstrations of its new radar technology compared to traditional systems.(Photo by Steve Fecht)
“It makes the invisible visible,” said Lawrence Burns, executive advisor to Atomathic.
For trucking, the implications could be significant. Large trucks create some of the most challenging radar environments, particularly in urban settings where pedestrians and cyclists may operate close to reflective trailers and cabs. Night driving and higher operating speeds further compress reaction time, increasing the importance of early and reliable detection.
If radar can more consistently identify weak targets near large vehicles, it could improve the performance of systems such as automatic emergency braking and reduce false alerts or disengagements that undermine driver trust. Broader validation, including independent testing, will be critical before the technology can be deployed in safety-critical applications.