THE CHALLENGE
The growing adoption of FMCW radar in modern vehicles—used for crucial functions like collision avoidance and autonomous driving faces a serious business and technical challenge: mutual radar interference. As more vehicles hit the road with similar radar systems, their signals often overlap, creating "noise" that hides real objects or creates false ones, undermining safety and reliability. Traditional solutions like blanking or filtering either erase important data or slow down response times, while advanced methods like compressed sensing demand too much computing power to be practical in real-time. For automakers and radar suppliers, this creates a pressing need for smarter, scalable interference mitigation solutions that deliver high accuracy without adding hardware costs or latency—especially as safety regulations tighten and consumers demand better performance in increasingly congested driving environments.
OUR SOLUTION
We deliver a real-time, software-based interference cancellation system for vehicle-mounted FMCW radars that stands out in both performance and practicality. Unlike traditional methods that often erase useful data or require costly hardware upgrades, our approach uses advanced signal processing and GPU acceleration to cleanly separate true target echoes from radar interference—even in dense traffic. By focusing on the unique patterns of interference in the time-frequency domain and applying a compressed sensing algorithm (CoSaMP) within a narrowed search space, we dramatically reduce computation time without sacrificing accuracy. The result is a fast, scalable, and cost-effective technology that integrates seamlessly into existing radar systems, enabling automakers and suppliers to meet the rising demands of autonomous driving and safety-critical functions without adding latency or complexity.
Figure: Overview of invention
Advantages:
Potential Application: