TY - JOUR TI - A vehicle-to-infrastructure based dynamic merge assistance method for mixed traffic with manual vehicle, connected vehicle, and automated vehicles in highway merging section DO - https://doi.org/doi:10.7282/t3-8kmk-z165 PY - 2019 AB - Merging activities on the highway can cause significant recurrent and non-recurrent bottleneck congestion and is a severe issue in traffic operations. The efficient enhancement of highway merging activities has been considered as a major task in highway management research and practice. Some macroscopic active traffic management (ATM) methods have been proposed to mitigate bottleneck congestion. In recent years, microscopic dynamic merging assistance (DMA) methods have been proposed as efficient methods to improve the mobility and safety in merging maneuver. These proposed methods have been serving their roles as effective merge control methods for decades. Despite some positive outcomes of these existing methods in the improvement of the highway merge traffic, there are still some gaps for new merge assistance methods to catch. Recent developments and deployments of Automated Vehicle (AV) technologies and Connected Vehicle (CV) technologies such as Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication provide new opportunities for developing more efficient merge assistance methods. This leads the new wave of Connected-Automated-Vehicle (CAV) based DMA methods. This dissertation proposes a microscopic CAV V2I-based Dynamic Merge Assistance (DMA) method for mixed types of vehicles including manual vehicle (MV), manually controlled connected vehicle (CV) and connected automated vehicle (CAV). The research starts from a comprehensive review of existing merge assistance methods and the evolution of microscopic DMA mechanisms and algorithms. Then integrated partial coordination merging control algorithm based on a pairing between mainline gaps and on-ramp merging vehicles (vehicle-gap pair) is proposed. The vehicle-gap pair is determined by a prediction of their merging potential, and the merging potential is predicted according to their instantaneous virtual trajectories (IVT) which are generated from the instantaneous lane speed profiles (ISP) of both mainline and onramp which consists of all detected vehicles’ location and speed at each time frame. The pairing scenario varies with different combinations of vehicle types (i.e., MV, CV, CAV) involved in a merging maneuver. These scenarios involve vehicles with different features in terms of control, sensing, and communication. Thus, a set of 4-level pairing criteria is proposed to fit the DMA method in the universal traffic environment with mixed vehicle types. The pairing process is then followed by a coordination car-following control specifically designed for different vehicle types. The coordination car-following control maintains the mainline gaps for the paired on-ramp vehicle and guarantees the on-ramp vehicle can catch up with the paired gap safely and smoothly. Different control mechanisms are proposed based on the observability, controllability, and the availability of communication in different vehicle combinations which is aligned to the 4-level pairing scenarios. A VISSIM simulation is built based on the traffic flow data collected from the I-35 corridor in Austin TX with multiple merging and weaving sections. The proposed DMA model is implemented through a VISSIM Application Programming Interface (API) named external drivers’ model (ETM). The safety performance of merging is evaluated by time-to-collision (TTC) and critical gap size (gap between mainline following vehicle and lane-changing on-ramp vehicle). The mobility performance is evaluated by average travel time and speed contour map in the whole simulation area including on-ramp, mainline merging area and mainline upstream. The proposed method is found to have a promising performance in both mobility and safety impact, especially on the on-ramp traffic during peak hours. KW - Connected and Automated Vehicle KW - Civil and Environmental Engineering KW - Traffic engineering KW - Traffic flow LA - eng ER -