Travel Time Prediction Based on Pattern Matching Method

The microwave detection technology has become an effective tool for monitoring highway traffic flow in China. The cross section traffic data collected provides models opportunity for travel time prediction. However, the sparseness of data somewhat constrains the prediction accuracy. To tackle this problem, the paper presents a highway travel time prediction algorithm based on pattern matching method. First, a pattern library is established by choosing traffic volume and speed as its certain state components and time as its uncertain state component. Then, the space-time two-dimensional linear interpolation method is used to calculate the mean speed and subsequently the travel time. Finally, similar patterns are obtained using K Nearest Neighbor approach and predicted travel time is calculated by the Weighted Average method. The case study shows that the pattern matching method for travel time prediction based on microwave detection data produces sufficient accuracy, which solves the problem of sparse detectors effectively.