Kwon, Eil

This study estimated and analyzed the travel-time reliability (TTR) and traffic-flow measures of effectiveness (MOE) for 74 directional corridors in the metro freeway network in Twin Cities, Minnesota, from January 2018 to December 2023, for both morning and afternoon peak periods. The network-wide trends for both TTR and MOE indicate that the traffic flows in the Twin Cities freeway network have not reached the pre-pandemic level as of December 2023. The TTR and MOE estimation results were applied to identify a set of the most vulnerable routes in the current network. Further, the preliminary resilience model, developed in the previous phase, was enhanced and applied to determine the operational resilience of 74 directional corridors in the network and a set of the low-resilient routes were identified. The effects of the route-wide geometric configuration on TTR, MOE and operational resilience on individual corridors were also analyzed. The results from this research could provide the basis for geometric and operational improvements of the metro freeway corridors.

This research developed an online procedure to estimate the weaving capacity through time for a simple ramp-weave section, the most common type of weaving areas in the Twin Cities' freeway network. The field observations and the analysis of the traffic data collected from a sample weaving section indicate that the freeway-to-ramp and ramp-to-freeway vehicles first merge and travel together at the beginning portion of the auxiliary lane before they split to the mainline or exit ramp. The length of the shared portion of the auxiliary lane, called an "effective weaving zone," varied depending on the length of an auxiliary lane and the amount of weaving volume. The above merge-split behavior and the resulting mixed flow on the auxiliary lane for a short time period explains the fact that the maximum possible weaving volume in a simple ramp-weave section equals the maximum through volume that the auxiliary lane can handle. Researchers used a Kalman Filter to obtain estimated weaving volume data from three weaving, which supported this observation. Based on the above findings, an online procedure was developed to estimate the maximum possible weaving volume for a given ramp-weave area through time using the volume and occupancy measurements from the loop detectors. The proposed procedure assumes that the maximum possible weaving volume for a given time interval is a function of downstream traffic conditions that can be quantified by estimating the time-variant merging and diverging capacities of a given weaving section. Test results with the five-minute data from a ramp-weave site indicate that the maximum possible weaving volume can be estimated with reasonable accuracy during congested peak periods.

This report summarizes the final results of the research effort to develop a freeway traffic simulator with the capability to evaluate freeway operational strategies, such as traffic-responsive ramp metering and high-occupancy vehicles (HOV) lanes. Researchers first developed an efficient software data structure by adopting a dynamic memory allocation scheme to use the available memory as efficiently as possible. That work also included modifying the existing macroscopic, segment-based modeling structure and developing new types of pipeline segments to facilitate detection modeling and further model enhancements. Based on the new segment-based modeling structure, researchers developed a new simulation module to handle HOV lane traffic flows and extended the simulation procedure for an exclusive HOV lane to handle a network of freeways. Further, the simulation model also incorporates a new module to emulate the traffic-responsive ramp metering algorithm implemented by the Traffic Management Center since the 1980s. The new software structure developed in this research allows the future addition of new metering algorithms without major difficulties. To facilitate the data input process for the expanded simulation features, a new Windows-based user interface was developed using the Delphi software development tool kit. With the new user interface, most of the data input process can be done without exiting the main menu screen. Note: The phase I report is available at https://hdl.handle.net/11299/156893.

Successful implementation of advanced traveler information systems over an entire urban network requires real-time measurement or estimation of arterial travel times ( or equivalently arterial journey speeds). This project develops an arterial journey speed model using data from inductive loop detectors and traffic controllers. This model incorporates the following key findings of traffic data analysis that researchers collected in Phase I. • Spot speeds are highly correlated with journey speeds when both speeds are low (0-15 mph) and uncorrelated with journey speeds when both speeds are high (greater than 25 mph). • Signal offsets or greenband width, traffic demand, green splits and capacity-reduction incidents are major factors that affect arterial travel time/journey speed. The model consists of two parts--the speed estimated from the volume and occupancy measured by detectors and the speed estimated based on critical volume/capacity ratio. Researchers tested and compared the model with a number of existing models, with promising results.

The envisioned operational tests of Advanced Traveler Information Systems (ATIS) and Advanced Traffic Management Systems (ATMS) in the Minneapolis/St. Paul area call for the provision of timely and reliable travel times over an entire road network. Unfortunately, travel time cannot be directly measured by certain new detection technologies, such as Automatic Vehicle Identification (A VI) or Automatic Vehicle Location (A VL) systems, these new technologies are not widely deployed and are much more costly than loop detectors. Finding an accurate way to estimate link travel time using loop detector data offers great economic benefits. This project examines the development if improved arterial travel time models. In the project's first phase, researchers reviewed existing travel time database. The project's second phase will seek to develop and evaluate new travel time estimation models.

The previous phases of this research reviewed and tested existing intersection control algorithms in a simulated environment. Further, a machine-vision detection system with four cameras was installed at the intersection of Franklin and Lyndale Avenues in Minneapolis, Minnesota, to develop a live intersection laboratory. Phase III enhanced the live laboratory with two additional cameras covering the intersection proper and the extended approach of southbound Lyndale Ave. A comprehensive operational plan for the laboratory was developed and a new microscopic simulator for the laboratory intersection was -also developed. Two types of new intersection control strategies, i.e., one with link-wide congestion measurements and the other based on neural-network approach, were developed and evaluated in the simulated environment. Further, using the data collected from the machine-vision detection system, an automatic procedure to estimate the intersection delay was also developed and applied to compare the performance of fixed-timing control with that of the actuated control strategy. The Phase I report is available at https://hdl.handle.net/11299/155938.

This report summarizes the final results of the current project to enhance the KRONOS program. The resulting version, V8.0, being operated under the MS-DOS environment, can handle two freeways merging/diverging with a common section with total length up to 20 miles and eight lanes wide. The input data preparation is performed interactively using pop-up menu screen with a mouse. For instance, the geometrics are entered by selecting the configuration of each segment from the available alternatives presented on the screen. The current version has a total of 26 available segment types, which can treat most of the freeways in the U.S. Further, a new incident simulation module that can handle up to six stages in terms of time-variant capacities is also added to the new version. Following definition of the geometrics of each segment, the entire freeway can be plotted for verification. The other input requirements include the arrival/departure demand pattern at the freeway boundaries and ramps in user specified time intervals (minimum 1 minute). The demand patterns are also plotted for verification and can be as complex as desired. The program allows employment of user-specified flow-density models which are entered interactively. The change to input already entered can be made at any stage during the data entry. Note: The phase I report is available at https://hdl.handle.net/11299/156893.

This project evaluates various intersection control strategies in a simulated environment and also helped establish a live laboratory for use in future testing of new control strategies. The report reviews major intersection control strategies, including the state-of-the-art strategies with adaptive and on-line timing generation features. In addition, it details simulation results for the OPAC control strategy. The NETSIM simulator created the simulation environment for a test network that included part of downtown Minneapolis. Comparison results indicate that OPAC performs best with low-traffic demands, and pretimed control was the most effective during peak periods when traffic demand reached capacity. In conjunction with this project, Minneapolis city traffic engineers installed a machine-vision video detection system at a live intersection laboratory. Located at Franklin and Lyndale Avenues, the test site will help researchers evaluate new control strategies before full-scale implementation in later phases of this research. The Phase I report is available at https://hdl.handle.net/11299/155938.

This report summarizes the final results of the enhancement and validation of the control-emulation method in the real freeway environment. A computer-based control-emulation method that can evaluate various automatic rate-selection strategies was developed by this research team in the previous phase of this research. Software was developed that operationalizes this method for field applications. A method was developed and tested to determine the best metering thresholds for a given section of a freeway, under normal weather conditions, using the control-emulation method and the downhill simplex optimization procedure. The method finds the optimal thresholds for each ramp in a given. section of the freeway by considering the system-wide traffic conditions. An independent procedure to determine the initial thresholds that can be used to initialize the optimization process was also developed. A preliminary study for developing an on-line predictor for freeway exit demand was performed and an adaptive prediction procedure was developed. The prediction model formulated in this research used historical demand and current day measurements. Further, the parameters in the prediction model are updated in real time using the Extended Kalman Filter, so that the propagation of prediction error can be minimized.

These appendices pertain to report 2022-01, Estimation of Metro Freeway System Reliability and Resilience.