Automotive radar image processing to predict vehicle trajectory

Millimetre wave radar sensors are being actively developed for automotive applications. Accurate knowledge of the road path ahead of the vehicle is essential to establish whether objects are a threat to the vehicle. An algorithm to predict vehicle trajectory, based on image processing techniques of the radar return signals, is discussed in this paper. The reliability of the algorithm was evaluated by comparison of its calculated radius of curvature with digital map database, GPS position, and yaw rate data. This approach produces an image which may be processed in a similar way to video data but has the benefit of all weather performance and additional high resolution range and velocity data. The process of road edge detection plays an important role in vehicle path prediction and will be discussed in this paper.