CMStatistics 2019: Start Registration
View Submission - CMStatistics
B1612
Title: The estimation of the traffic density and vehicle speed for detecting anomalies Authors:  Kai Kasugai - Chuo University (Japan) [presenting]
Toshinari Kamakura - Chuo University (Japan)
Abstract: It is an important issue to predict traffic density by using highway driving history data in order to ensure road safety on the highway. However, vehicles position and speed data often fail to represent the actual driving situation due to observation errors. It is often judged as abnormal even though an anomaly has not occurred. The purpose is to predict the state of the traffic density and vehicles speed at each point from the observed traffic density and vehicles speed at each point. With the Ensemble Kalman Filter, we could estimate precisely the traffic density and vehicles speed at each location following the observed values well. In case of calculating the Kalman gain, the stabilized results were obtained by applying a square root filter to the variance-covariance matrix. The proposed method illustrates easiness of detecting change points of traffic density and vehicles speed which results in findings of anomalies in the traffic highway system.