Lane change behavior is a common driving behavior on expressway. The duration of lane change is a fundamental parameter for microscopic traffic simulation model. The aim of this paper is to analysis the effects of different variables on the duration of lane change behavior. In order to get the accurate status information of each vehicle on the expressway, two UAVs were used at the same time. The high-precision map was built to support the data process from the videos. 8 types of different parameters such as speed, acceleration, lane position, coordinates, and 538 complete lane change behavior were extracted. 14 different variables related to the lane change duration were examined. It was found that the main reason for the lane change maneuver was the low speed of the leading vehicle and the average relative speed was 2.26m/s (8.14km/h). Drivers took risk to change the lane even when the relative distance between the lane change vehicle and the lag vehicle in the destination lane was small. Through the Spearman correlation analysis, it is found that the duration of lane change time was only related to three different variables which is the speed of the subject vehicle, the relative speed with the leading vehicle and the relative speed with the lag vehicle. The other variables were of no significant in this paper. The conclusion get from this paper can help us have a better understanding of the lane change behavior and also can be used to develop a more reasonable mode for the lane change behavior in China.

CICTP