Study On Pedestrian Level Of Service At Signalized Intersection Under Mixed-traffic Conditions

In the urban traffic, signalized intersections are the bottleneck of network capacity and the places where traffic accidents take place frequently. Motor vehicles are being taken as the main consideration while vulnerable groups of transport system-pedestrians are being neglected in ongoing intersection research. How to create a relaxing natural walking environment and how to improve pedestrian Level-of-Service (LOS) has become one of the problems to be solved urgently.In view of this, supported by National Natural Science Foundation Project "Study on Several Issues of Pedestrian Level-of-Service at Signalized Intersection", pedestrian LOS was studied under mixed traffic conditions as following:(1) Both the signal delay during red time and pedestrians' conflicts with turning vehicles during green time were considered, and pedestrian signal delay and disturbing delay were selected as evaluation indexes.(2) Pedestrian LOS classification was attributed to pattern recognition problem. Due to respective advantages and disadvantages of supervised and unsupervised classification for pattern recognition, a mixed classification method combining K-means clustering and Support Vector Machines (SVM) was used to achieve LOS classifier results.(3) Further, Orthogonal Design Method was used to determine the sensitivity degree of affecting factors on LOS.Finally, an actual assessment of typical signalized intersections' pedestrian LOS in Beijing was taken. The results show that the method is more practical than HCM2000.