Davis, Gary A.

The main objectives of phase 2 of this project were to obtain relevant data to calculate the percent remaining service life interval (PRSI) and two additional metrics and to perform Markov chain analysis and dynamic programming to determine how much time and funding is required to bring the system to a stable configuration, which allows for more consistent planning. First, relevant pavement management data was obtained from MnDOT and preliminary data analyses were performed. The prediction models and optimization process currently used by MnDOT were investigated and summarized. Next, two additional metrics, Asset Sustainability Ratio and Deferred Preservation Liability, were calculated for MnDOT’s network. Then details of the estimation process of state-to-state transition probabilities to be used in the Markov chain model were presented. To allow for site-specific variation, ordinal logistic regression models were incorporated in the Markov chain model. The results were used in a dynamic programming optimization methodology to obtain baseline and optimal policies for different scenarios and a user-friendly excel spreadsheet tool was developed. Finally, a summary of the work performed followed by conclusions and recommendations was presented.

Stop lines are ubiquitous; but do they really impact intersection safety? Prior to this project; no long-term studies on intersection safety with stop lines had been completed. This project was developed with two parallel research efforts: a safety study and an observational study. The safety study was developed to address stop lines' effects over the long term and used crash data from five cities' stop-controlled intersections to perform regression and see if stop lines actually influenced safety. The observational study was developed to determine if stop lines have an effect on driver behavior at intersections and to look at where drivers were stopping. Video was collected at 16 different intersections before and after a stop line was painted. The safety study and observational study showed that stop lines did not have a significant impact on driver behavior or intersection safety; but other factors like speed limits and sight distance did. Implications for practice include carefully examining sight distance at the intended stopping point to ensure drivers have adequate sight distance in both directions. If sight distance is not adequate; moving the intended stop location or reconsidering whether the intersection should have signage -- stop or yield -- or be uncontrolled could yield better driver compliance and safety.

Pedestrian-Activated Crossing (PAC) systems have been shown to have a generally positive impact on driver yield rates. However; there has been insufficient research on the effect PAC treatments have on pedestrian crash rates; and there is little guidance as to when and where each treatment should be used. This study estimates the effects of PACs on pedestrian crash rates using Monte Carlo simulation and examines the relationships between driver yield rates and a variety of treatments and site designs by conducting an observational study using video data from 34 locations. The simulation outcomes suggests that while the percentage of yielding drivers might be a useful indicator of pedestrian level of service; it is less helpful as safety surrogate. This could be because a driver's yielding to a pedestrian; as observed in field studies; might not be the same behavior as a driver attempting to stop during a vehicle/pedestrian conflict. The observational study shows that the number of lanes to cross at a crossing is positively correlated with the rate at which pedestrians activate the system; but it is not correlated with the delay. Additionally; the study showed that the effect of PAC systems is most pronounced at sites with a higher number of movements conflicting with the crossing or poor visibility from upstream without signs warning drivers of an upcoming crosswalk.

As part of an Urban Partnership Agreement project; the Minnesota Department of Transportation added lanes and began operating a priced dynamic shoulder lane (PDSL) on parts of Interstate 35W. Following the opening of these improvements; the frequency of rear-end crashes increased in certain sections; especially in the PDSL region. The object of this study was to determine if these increases were direct effects of the improvements or were due to changes in traffic conditions. Logistic regression analyses which controlled for changes in traffic conditions indicated no direct effect on the likelihood of rear-end crashes due to operation of the PDSL; the observed change in crash frequency was explained by the change in traffic conditions. This study also found evidence for a nonlinear relationship between a proxy for traffic density; lane occupancy; and the probability of a rear-end crash occurring during an hour. In several sections crashes were most likely when lane occupancies were approximately 20%-30%; and crash likelihood tended to decrease for lane occupancies below and above this range.

This project explores the possibility of using high-resolution traffic signal data to evaluate intersection safety. Traditional methods using historical crash data collected from infrequently and randomly occurring vehicle collisions can require several years to identify potentially risky situations. By contrast; the proposed method estimates potential traffic conflicts using high-resolution traffic signal data collected from the SMART-Signal system. The potential conflicts estimated in this research include both red-light running events; when stop-bar detectors are available; and crossing (i.e. right-angle) conflicts. Preliminary testing based on limited data showed that estimated conflict frequencies were better than AADT for predicting frequencies of angle crashes. With additional validation this could provide a low-cost and easy-to-use tool for traffic engineers to evaluate traffic safety performance at signalized intersections.

The objective of this project was to develop guidelines for time-of-day use of permitted left-turn phasing, which can then be implemented using flashing yellow arrows (FYA). This required determining how the risk for left-turn crashes varied as traffic-flow conditions varied during the course of a representative day. This was accomplished by developing statistical models, which expressed the risk of occurrence of a left-turn crash during a given hour as a function of the left-turn demand, the opposing traffic volume, and a classification of the approach with respect to the opposing traffic speed limit, the type of left-turn protection, and whether or not opposing left-turn traffic could obstruct sight distance. The models were embedded in a spreadsheet tool which will allow operations personnel to enter, for a candidate intersection approach, existing turning movement counts, and a classification of the approach with respect to speed limit, turn protection, and sight distance issues and receive a prediction of how the risk of left-turn crash occurrence varies throughout the day, relative to a user-specified reference condition. NOTE: The spreadsheet tool is an Excel macro file and may be accessed from https://edocs-public.dot.state.mn.us/edocs_public/DMResultSet/download?docId=29248594

Weigh-In-Motion (WIM) system tends to go out of calibration from time to time, as a result generate biased and inaccurate measurements. Several external factors such as vehicle speed, weather, pavement conditions, etc. can be attributed to such anomaly. To overcome this problem, a statistical quality control technique is warranted that would provide the WIM operator with some guidelines whenever the system tends to go out of calibration. A mixture modeling technique using Expectation Maximization (EM) algorithm was implemented to divide the Gross Vehicle Weight (GVW) measurements of vehicle class 9 into three components, (unloaded, partially loaded, and fully loaded). Cumulative Sum (CUSUM) statistical process technique was used to identify any abrupt change in mean level of GVW measurements. Special attention was given to the presence of auto-correlation in the data by fitting an auto-regressive time series model and then performing CUSUM analysis on the fitted residuals. A data analysis software tool was developed to perform EM Fitting and CUSUM analyses. The EM analysis takes monthly WIM raw data and estimates the mean and deviations of GVW of class 9 fully loaded trucks. Results of the EM analyses are stored in a file directory for CUSUM analysis. Output from the CUSUM analysis will indicate whether there is any sensor drift during the analysis period. Results from the analysis suggest that the proposed methodology is able to estimate a shift in the WIM sensor accurately and also indicate the time point when the WIM system went out-of-calibration. A data analysis software tool, WIM Data Analyst, was developed using the Microsoft Visual Studio software development package based on the Microsoft Windows® .NET framework. An open source software tool called R.NET was integrated into the Microsoft .NET framework to interface with the R software which is another open source software package for statistical computing and analysis.

A flashing LED stop sign is essentially a normal octagonal stop sign with light emitted diodes (LED) on the stop sign's corners. A hierarchical Bayes observational before/after study found an estimated reduction of about 41.5% in right-angle crashes, but with 95% confidence this reduction could be anywhere between 0% and 70.8%. In a field study, portable video equipment was used to record vehicle approaches at an intersection before and after installation of flashing LED stop signs. After installing the flashing stop signs, there was no change in the relative proportion of clear stops to clear non-stops when minor approach drivers did not face opposing traffic, but the relative proportion of clear stops increased for drivers who did encounter opposing traffic.

The main objective of this study was to estimate the speed impact of occasional hazard (playground) warning (OHPW) signs along residential streets. Three types of data were collected at each of three study sites approximately one month before and one week to one month after the installation of a pair of OHPW signs. Vehicle speed data were collected with a pneumatic tube device. Manual observations were recorded, and focused on the magnitude and location of the on-street parking and park and/or playground activities occurring at the study sites. Linear regression analysis was used to estimate the change in mean vehicle speed associated with the presence of the OHPW signs, while controlling for the effects due to activity levels on the streets and the playgrounds. At one site the OHPW sign had no discernible effect on mean vehicle speeds, while at the other two sites mean vehicle speeds decreased by 1.5 mph and 0.9 mph following installation of the OHPW signs.

Report #10 in the series: Access to Destinations Study. The objectives of this project were to (a) produce historic estimates of travel times on Twin-Cities arterials for 1995 and 2005, and (b) develop an initial architecture and database that could, in the future, produce timely estimates of arterial traffic volumes and travel times. Our Phase I field study indicated that on arterial links where both the demand traffic volume and the signal timing are known, model-based estimates of travel time that are on average within 10% of measured values can be obtained. Phase II of this project then focused on applying this approach to the entire Twin Cities arterial system. The Phase II effort divided into three main subtasks: (1) updating estimates of demand traffic volume obtained from a transportation planning model to make them consistent with available volume measurements, (2) collecting information on traffic signal locations in the Twin Cities and compiling this into a geographic database, and (3) combining the updated traffic volumes and signal information to produce link-by-link peak-period travel time estimates. The traffic volume update took as inputs the predicted volumes generated by a traffic assignment model and measured average annual daily traffic from automatic traffic recorders, and gave as output updated estimates of the traffic volumes for links lacking automatic traffic recorders. A request to state, county and municipal agencies in the seven-county metro area produced Information on approximately 2,900 traffic signals. Estimated arterial travel times for the morning and afternoon peak periods for 1995 and 2005 were then computed and sent to other components of the Access to Destinations effort.