Analysis And Modeling On Characteristics Of Through Vehicles' Movement At Signalized Intersections With Mixed Traffic

Most developing cities are characterized by mixed traffic. Mixed traffic itself always leads to inefficient utilization of network facilities and makes existing methods and models involved in researches on traffic and transportation not so suitable, as they were presented under conditions of pure motorized traffic. The intersection is one of the most significant facilities among the road network in a city, and it is also one of the most prevalent places that pools mixed traffic flows as well as problems induced by mixed traffic flows. Capacity analysis of signalized intersections under conditions of mixed traffic is a difficult but important research subject for developing cities. But existing capacity formulas either ignore effects of disturbances caused by pedestrians and non-motorized vehicles or merely introduce a fixed discount coefficient to meet the effects. Inevitably these formulas will result in big errors when mixed traffic occurs and varies among different intersections. This thesis develops a novel method for capacity analysis and estimation that based on the recognition and discrimination of mixed traffic conditions. It is relied on the analysis of impact of pedestrians and non-motorized vehicles on the advancement of motorized vehicles. Prior to the capacity method, a mathematic model is developed to describe moving features of motorized vehicles under conditions of mixed traffic, and based on the model disturbance effects of pedestrians and non-motorized vehicles are investigated and analyzed. Recognition of disturbance effects makes it possible to distinguish different mixed traffic conditions and this work helps to develop a novel sorting art for signalized intersections which is also presented before the capacity analysis. All works of this thesis are summarized as below:(1) Discussion on the spatio-temporal feature of disturbances caused by pedestrians and non-motorized vehicles at signalized intersections. As the production of the discussion, disturbance time and disturbance zone are defined. The spatio-temporal analysis on disturbances is groundwork in the thesis.(2) Development of a mathematic model (called MASPA model in this thesis) to describe through vehicles advancing during the start part of green time. As green indication starting, through vehicles in the involved approach often experience a particular course observed as that a closely-spaced platoon comes into being and then turns to disperse. The model is based on this course and the movement of each through vehicle is divided into four stages:initially accelerating from rest, shortly after which reducing acceleration rate to join a moving platoon, then increasing acceleration rate to disperse from the platoon and eventually stepping into a free flow. Constraints of time headway and speed are introduced to link adjacent stages. This model produces the velocity-time curve for each through vehicle at each stage and at each junction of adjacent stages. And from the velocity-time curve, two important variables of stop-line passing time and intersection passing time are defined for each vehicle, which are used in the following analysis. Case studies show a good fit for this model to describe moving behavior of through vehicles during the early time of green phase. Especially, for performance measures of stop-line passing time and intersection passing time, the relative error between the empirical value and the estimated value by the model is usually less than 1%.(3) Quantitative analysis on the effect of disturbances derived from pedestrians and non-motorized vehicles. This work leads to discrimination of mixed traffic conditions at signalized intersections and consequently a novel sorting art for signalized intersections. The concept of relative increment (RI) of the through vehicle's intersection passing time is introduced and calculated by the MASPA model. Here it is confirmed by the technique of traffic survey that the RI of intersection passing time is mainly caused by disturbances of pedestrians and non-motorized vehicles, so the RI is adopted to stand for the effect of disturbances and used to recognize and distinguish different mixed traffic status at different intersections. This part of study makes it clear that according to mixed traffic status (in fact disturbances effect from pedestrians and non-motorized vehicles) signalized intersections in developing cities can be classified into three interweaving grades:the lowest grade (also named grade A), the moderate grade (grade B) and the highest grade (grade C).(4) Development of a novel method for capacity analysis and estimation. This method uses the data of stop-line passing time produced by the MASPA model to found a capacity formula, meanwhile by revising the one of stop-line method which was presented by Beijing Municipal Design Institute, Beijing city, China. In the case study of theoretic calculations at practical intersections, the new method is proved more fit for all kinds of mixed traffic conditions compared to other existed methods.(5) Simulating analysis on the capacity methods. A VISSIM-based simulator is designed to evaluate the capacity method developed in this thesis, comparing with other existed methods. In the simulator varied but feasible motorized flows are preset as the input and the fluctuant outputs of flow and delay are recorded. The analysis and thus the evaluation are derived from the relationship between the output and the input. The result of simulating analysis is quite close to that of theoretic calculations, which confirms the validity of the new method.