A work zone in Arkansas was equipped with an automated work zone information system (AWIS). This study evaluated performance of the system to assess the impacts of AWIS on safety.

The site was a 6.3 mile segment Interstate 40 located in Lonoke County. This segment of roadway is considered rural and has an average daily traffic volume of 36,350 vehicles, with 43 percent of traffic being trucks. The AWIS deployed at this work zone site included: a Central System Controller, two highway advisory radios, five traffic sensors, five changeable message signs (CMS), and two supplemental speed stations per lane closure. This system was primarily a queue detection system designed to calculate and report delay times to travelers via changeable message signs at the roadside. To determine delay, traffic sensors were installed upstream of the lane closure or taper. If the difference in vehicle speeds between sensors was greater than 10 mi/h, a variable message sign located upstream would display the message "REDUCE SPEED TO XX MPH," followed by the message "YY MINUTE DELAY." If the difference in speed was less than 10 mph, only the delay message was displayed. In addition, The HAR system at the site provided the public with general work zone information, and informed travelers of expected delays.

The objective was to reduce the number of rear-end and fatal crashes at the site. The effectiveness of the system was determined by an evaluation between the Lonoke County site and two comparable construction sites not using AWIS (Brinkley-Goodwin and Goodwin-East). The analysis found that the fatal crash rate in Lonoke County was lower than both comparison sites. Lonoke County had a lower rear-end crash rate than Brinkley-Goodwin but higher rear-end crash rate than Goodwin-East. These results are summarized in the following table.

Description	Period of Crash Data used in Analysis	Fatal Crash Rate Per 100 Million Vehicle Miles Traveled	Rear-End Crash Rate Per 100 Million Vehicle Miles Traveled
Lonoke County Site (with AWIS)	*July 2000 thru June 2001	2.2	33.7
Brinkley – Goodwin (without AWIS)	*July 2001 thru December 2001	3.4	43.2
Goodwin – East (without AWIS)	*July 2000 thru Sept. 2001	3.2	29.5

*The period of the crash data for these projects corresponds with the time the work zone was active. Crash data was only available through December 2001. Source: Table-3 from Tudor, et al.


The accuracy of the work zone delay information (predicted travel times) was compared to actual travel time. This evaluation found that of the 144 trial runs, 14 travel times were outside the acceptable time difference of 5 minutes (greater or lower). The system was 89.71 percent accurate by these measures.

An interview with the engineer responsible for overseeing work zone construction in Lonoke indicated the system appeared to prevent/reduce rear-end collisions as long as traffic was not backed up past the message boards. The engineer also had the opinion that the system did not improve incident response time.

Notes

The study period for this project was limited and this conference paper did not document a statistical analysis of the safety data. In addition, Arkansas initiated a statewide safety program to increase driver planning and awareness, merging and speed limit compliance, and work-site safety. It is uncertain how these programs may have impacted driver behavior in work zones during the evaluation project.