Predicting for Traffic Risk Degree: Novel Prediction Method and Samples

Abstract

To effectively predict the risk degree in both maritime and road traffic, a novel method is proposed in this study. First, the improved Dempster–Shafer evidence theory was derived to address multiple evidence based on the uncertain mass in the traffic environment. Further, the iterative combination equations reduced the computational complexity when computing the traffic risk degree in a given scan. Accordingly, the modified adaptive Kalman filter was explored to predict the traffic risk degree for the next scan. To maintain a positive definiteness in the estimation covariance during the whole filtering process, the Cholesky decomposition was applied to enhance reliability. By transmitting the lower triangular matrix from the Cholesky decomposition of estimation covariance, the computational complexity was reduced relatively. Finally, the experiment results indicated that the proposed method had satisfactory prediction performance for the traffic risk degree.