Daniel B Fambro

APPLICATION OF SUBSET AUTOREGRESSIVE INTEGRATED MOVING AVERAGE MODEL FOR SHORT-TERM FREEWAY TRAFFIC VOLUME FORECASTING

Traffic volume is one of the fundamental types of data that have been used for the traffic control and planning process. Forecasting needs and efforts for various applications will be increased with the deployment of advanced traffic management systems. With the importance of the short-term traffic forecasting task, numerous techniques have been utilized to improve its accuracy. The use of the subset autoregressive integrated moving average (ARIMA) model for short-term traffic volume forecasting is investigated. A typical time-series modeling procedure was employed for this study. Model identification was carried out with Akaike’s information criterion. The conditional maximum likelihood method was used for the parameter estimation process. Two white noise tests were applied for model verification. From the analysis results, four time-series models in different categories were identified and used for the one-step-ahead forecasting task. The performance of each model was evaluated using two statistical error estimates. Results showed that all time-series models performed well with reasonable accuracy. However, it was observed that the subset ARIMA model gave more stable and accurate results than other time-series models, especially a full ARIMA model. It is believed that the use of a subset ARIMA model could increase the accuracy of the short-term forecasting task within time-series models.

PREDICTING OPERATING SPEEDS ON TANGENT SECTIONS OF TWO-LANE RURAL HIGHWAYS

Prediction and estimation of speeds on two-lane rural highways are of enormous significance to planners and designers. Estimation of speeds on curves may be easier than prediction of speeds on tangent sections because of the strong correlation of speeds with a few defined and limiting variables, such as curvature, superelevation, and the side-friction coefficients between road surface and tires. On tangent sections, however, the speed of vehicles is dependent on a wide array of roadway characteristics, such as the length of the tangent section, the radius of the curve before and after the section, cross-section elements, vertical alignment, general terrain, and available sight distance. Few studies have dealt with this issue because a considerable database is necessary to identify any significant trends and substantial modeling effort is required. Research analyzed the variability of the operating speeds on 162 tangent sections of two-lane rural highways, and models were developed for prediction of operating speed based on the geometric characteristics available. A one-model approach was used initially. Because of the low R2 values, a family of models was developed that better predicted operating speeds.

Driver Perception-Brake Response in Stopping Sight Distance Situations

One of the most important requirements in highway design is the provision of adequate stopping sight distance at every point along the roadway. At a minimum, this sight distance should be long enough to enable a vehicle traveling at or near the design speed to stop before reaching a stationary object in its path. Stopping sight distance is the sum of two components–brake reaction distance and braking distance. Brake reaction distance is based on the vehicle’s speed and the driver’s perception–brake reaction time (PBRT). Four separate, but coordinated, driver braking performance studies measured driver perception–brake response to several different stopping sight distance situations. The results from the driver braking performance studies suggest that the mean perception–brake response time to an unexpected object scenario under controlled and open road conditions is about 1.1 s. The 95th percentile perception–brake response times for these same conditions was 2.0 s. The findings from these studies are consistent with those in the literature: that is, most drivers are capable of responding to an unexpected hazard in the roadway in 2.0 s or less. Thus, the American Association of State Highway and Transportation Officials’ perception–brake response time of 2.5 s encompasses most of the driving population and is an appropriate value for highway design.

DRIVER BRAKING PERFORMANCE IN STOPPING SIGHT DISTANCE SITUATIONS

Assumed driver braking performance in emergency situations is not consistent in the published literature. A 1955 study stated that in an emergency situation “it is suspected that drivers apply their brakes as hard as possible.” This idea differs from a 1984 report that states drivers will “modulate”their braking to maintain directional control. Thus, additional information is needed about driver braking performance when an unexpected object is in the roadway. In this research driver braking distances and decelerations to both unexpected and anticipated stops were measured. The study design allowed for differences in vehicle handling and driver capabilities associated with antilock braking systems (ABS), wet and dry pavement conditions, and the effects of roadway geometry. Vehicle speeds, braking distances, and deceleration profiles were determined for each braking maneuver. The research results show that ABS result in shorter braking distances by as much as 30 m at 90 km/h. These differences were most noticeable on wet pavements where ABS resulted in better control and shorter braking distances. Braking distances on horizontal curves were slightly longer than on tangent sections; however, they were not large enough to be of practical significance. Maximum deceleration during braking is independent of initial velocity, at least in the range of speeds tested. Differences were noted in individual driver performance in terms of maximum deceleration. Although maximum deceleration was equal to the pavement’s coefficient of friction for some drivers, the average maximum deceleration was about 75 percent of that level. Overall, drivers generated maximum decelerations from 6.9 to 9.1 m/s2. The equivalent constant deceleration also varied among drivers. Based on the 90-km/h data, 90 percent of all drivers without ABS chose equivalent constant decelerations of at least 3.4 m/s2 under wet conditions, and 90 percent of all drivers with ABS chose equivalent constant deceleration of at least 4.7 m/s2 on dry pavements.

Traffic characteristics of protected/permitted left-turn signal displays

At least four variations of the permitted indication in protected/permitted left-turn (PPLT) control have been developed in an attempt to improve the level of driver understanding and safety. These variations replace the green ball permitted indication with a flashing red ball, a flashing yellow ball, a flashing red arrow, or a flashing yellow arrow indication. In addition, the Manual on Uniform Traffic Control Devices allows several PPLT signal display arrangements. The variability in indication and arrangement has led to a myriad of PPLT displays throughout the United States. The level of driver understanding related to each PPLT display type, and the associated impact on traffic operations and safety, has not been quantified. A study was conducted to evaluate the operational characteristics associated with different PPLT signal displays. Specifically, the study quantified saturation flow rate, start-up lost time, response time, and follow-up headway associated with selected PPLT displays. No differences in saturation flow rate and start-up lost time were found due to the type of PPLT signal display. Saturation flow rates ranged from 1,770 to 2,400 vehicles per hour of green per lane and were related to differences in driver behavior between geographic locations. The variation in start-up lost time and response time between locations was primarily related to differences in phase sequence. The flashing red permitted indications were associated with the longest follow-up headway times, since drivers are required to stop before turning left with a flashing red permitted indication. The shortest follow-up headway was associated with the five-section cluster display using a green ball indication.

OPERATING SPEEDS ON CURVES WITH AND WITHOUT SPIRAL TRANSITIONS

Although the design policies of many countries encourage the use of spiral transition curves, questions exist about whether they have any benefit in roadway design. In addition, United States speed prediction models do not account for the presence of spiral transitions. To determine if spiral transitions affect the speed at which vehicles traverse horizontal curves on rural two-lane roadways, speeds of free-flow passenger cars on spiral transition curves were compared with speeds of free-flow passenger cars on circular curves that had similar geometric characteristics. Operating speeds and geometric data were collected at 12 spiral transition curves and 39 circular curves in six states. Using regression techniques, it was concluded that spiral transitions did not significantly affect the 85th percentile speed of drivers on horizontal curves. Analysis of the cumulative speed distributions for different curve radii showed that speeds were higher for spiral transition curves than for circular curves for a 145-m radius (smallest radii studied); however, no data were available to test the hypothesis further. It is believed that spiral transition curves may affect vehicle speeds as the curve radius decreases. It was concluded that spiral transitions did not produce significant differences in passenger-car operating speeds.

Driver eye and vehicle heights for use in geometric design

Driver eye, headlight, taillight, and vehicle heights are important elements for determining passing and intersection sight distances and horizontal and vertical curve lengths to provide required stopping sight distance. Driver eye and object heights have varied significantly since their inception in the 1920s, when their values were suggested as 1676 mm. The objective of this study was to determine appropriate driver eye, headlight, taillight, and vehicle heights for use in developing geometric design criteria. The results of this research were used to recommend a driver eye height of 1080 mm for design purposes. This value represents 90 percent of the passenger car driver eye height values and an even higher percentage of the total vehicle fleet, because passenger cars have the lowest driver eye height values and represent fewer than two-thirds of the total vehicle fleet. Headlight and taillight heights of 600 mm are recommended for design. These values represent over 90 and 95 percent of the passenger cars observed in this study, respectively. The vehicle height recommendation for sight distance was 1315 mm, which represents the 10th percentile passenger car height values measured in the research.

Operating speed on suburban arterial curves

Free-flow speeds were collected at both a control section and a curve section at 14 surburban sites with horizontal curves and 10 suburban sites with vertical curves. The scatter plots of the 85th percentile speed versus approach density indicate that when the approach density is between 3 and 15 approaches per km, approach density does not influence speed. Regression analysis indicated that the curve radius for horizontal curves and the inferred design speed for vertical curves can be used to predict the 85th percentile speed on curves for vehicles on the outside lane of a four-lane divided suburban arterial. For horizontal-curve sites, a curvilinear relationship exists between curve radius and the 85th percentile speed. A linear relationship provided the best fit between the inferred design speed and the 85th percentile speed for the vertical curve sites. For the horizontal and vertical curve sites, the speed at which 85th percentile speed becomes less than the inferred design speed is lower for suburban arterials than for rural highways. Drivers on suburban horizontal curves operate at speeds greater than the inferred design speed for curves designed for speeds of 70 kph or less, whereas on rural, two-lane roadways, drivers operate at speeds greater than the inferred design speed for curves designed for speeds of 90 kph or less. For vertical curves, the speeds at which drivers operate greater than the inferred design speed are 90 kph for suburban arterials and 105 kph for rural highways. These results are within 12 kph of the observed 85th percentile speeds on nearby control sections (approximately 80 kph for suburban arterials and 100 kph on rural highways).

Measuring Driver Performance in Braking Maneuvers

A simple, reliable instrumentation package with an on-board computer for installation in a test vehicle or the test drivers own vehicle is described. This package was used in a research project recently completed, an empirical investigation of stopping sight distance. Selected data on perception-response time (PRT) and braking performance under artificial and simulated on-the-road emergency conditions are presented. PRTs were less than the AASHTO assumed constant of 2.5 sec even at the 95th percentile. Braking performance in terms of steady deceleration was greater than -0.32 g at the 95th percentile.

Enhanced traffic control devices at passive highway-railroad grade crossings

More than 2,000 crashes and 239 fatalities were reported at public passive highway-railroad grade crossings in 1994. Driver error, often due to a breakdown in communication between traffic control devices and the driver, is commonly cited as a factor in passive grade crossing crashes. The objective of this study was to evaluate an improved method for communicating with drivers in an effort to improve safety at passive grade crossings. Specifically, this study evaluated the effectiveness of a vehicle-activated strobe light and supplemental sign as enhancements to the railroad advance (W10-1) warning sign at a passive highway-railroad grade crossing near Temple, Texas. Three study methods were used to evaluate this enhanced sign system including a before and after speed study, a driver survey, and a driver observation study. The results indicated that average speeds on the approaches to the grade crossing were lower after the installation of the enhanced sign system. Drivers responded favorably to the enhanced sign system, and no adverse driver reactions were observed at the onset of the flashing strobe light. The strobe light was effective in directing drivers’ attention to the railroad advance warning and supplemental signs. The enhanced sign system appears to increase driver awareness of the passive grade crossing, cause some drivers to approach the grade crossing with additional caution, and reduce the average speed near the nonrecovery zone on both approaches.