A domestic research team develops a technology that doubles the performance of radar used in Autonomous Driving vehicles using software alone, without adding hardware./Courtesy of Shutterstock

Daegu Gyeongbuk Institute of Science and Technology (DGIST) researchers, together with the Korea Naval Academy, developed a technology that greatly improves radar speed-measurement accuracy. Because performance can be boosted by improving software alone without additional hardware, it is expected to be used in a range of next-generation vehicles, including self-driving cars, unmanned aerial vehicles, and unmanned ships.

Researchers Kim Sang-dong and Kim Bong-seok of the DGIST Future Mobility Research Division said on the 4th that, in collaboration with Professor Choi Young-du's team at the Korea Naval Academy, they developed an "extrapolation-based Doppler resolution enhancement algorithm" that can be applied to FMCW radar.

Radar determines speed by analyzing changes in radio waves from a target object, known as the Doppler effect. Until now, analysis has relied on the fast Fourier transform (FFT) method, but it has limits in performance because it is difficult to distinguish multiple objects with similar speeds.

A team from Daegu Gyeongbuk Institute of Science and Technology (DGIST), together with the Korea Naval Academy, develops an extrapolation-based algorithm to enhance Doppler resolution for FMCW radar. The illustration shows an environment where an unmanned vehicle's radar must detect and distinguish multiple targets./Courtesy of DGIST

The researchers introduced an extrapolation technique. Extrapolation is a method used to predict values that lie beyond the original observation range based on the relationships among known variables. This allows for higher resolution in speed measurements without extending radar observation time.

As a result, applying the new algorithm reduced speed estimation error by up to 33% compared with the conventional method and decreased the rate of missed targets by up to 68%. It also effectively resolved signal overlap when objects with similar speeds were present at the same time, enabling more accurate distinction.

In addition, it does not require extra hardware and has computational complexity similar to the existing FFT method, making it well suited for immediate application to real-time radar systems.

Principal researcher Kim Sang-dong said, "This technology is a way to make radar performance more accurate and efficient," and added, "It will develop into a core technology in defense, Autonomous Driving, and unmanned systems."

The findings were published on the 1st in the international journal "Journal of Electrical Engineering & Technology."

References

Journal of Electrical Engineering & Technology (2025), DOI: https://doi.org/10.1007/s42835-025-02453-6

※ This article has been translated by AI. Share your feedback here.