Multisource flight surveillance radar data fusion

摘要

Multisource flight surveillance data fusion based on radar data containing different errors, time stamps and time delays is considered in this work. The data fusion is done using simulated as well as real radar data, where it is assumed that the data have been corrected for the systematic errors, with only random errors remaining in the data. The systematic errors for real surveillance radar data are thus estimated and corrected for, using available DGPS data as the true track. A simple Kalman filter is then set up for the radar data, estimating the aircraft position based on the corrected position from each radar site. The covariance matrix of the estimation error and the distance between the radar site and the aircraft are then used in a weighting function towards the data fusion of the Kalman filtered aircraft position. The data fusion is tested on real and simulated data, resulting in a drastic improvement in positional accuracy, leading to errors below 210m. The resulting errors are less than the errors due to the data quantization in the radars, which range from 230-750m, thus further improvement is hardly achievable. The data fusion with and without using the simple Kalman filter showed that the benefit of using a simple Kalman filter is not significant.