Highway/Railroad Grade Crossing Studies: Improving Safety Through Information Technology By Ken Gerling, P.E., and Cleve crossings based on accident poten- DeVault, P.E. tial. The highest-ranked crossings become candidates to receive feder- al funds for roadway safety Introduction improvements. The Kansas railway Operation Lifesaver, a national, non- network consists of 17 rail carriers, profit public education program dedi- over 5,000 track miles and more cated to reducing crashes at high- than 6,200 public at-grade crossings way/railroad intersections, reports in the state (Figure 1). To receive the that nationwide in 1999, about 400 most benefit from limited federal Helping find ways to people died from train/vehicle colli- sions at grade crossings and another funds, KDOT strives to make safety improvements first to those cross- 1,360 people were injured. ings that have the most potential for reduce accidents at accidents. Helping find ways to reduce accidents grade crossings is the at grade crossings is the driving force Burns & McDonnell is assisting the behind a study Burns & McDonnell is state by updating the analytical driving force behind a conducting for the Kansas Department of Transportation tools and data supporting the cross- ing management program. The proj- (KDOT). KDOT is responsible for ect involves visiting all of Kansas’ study Burns & managing the federally funded public crossings to update the state (Section 130) grade crossing safety and federal inventory, updating the McDonnell is conduct- program for upgrading public road- existing ranking equation and way/rail crossings in Kansas. The developing a software application to ing for the Kansas safety program involves maintaining an updated inventory of crossing provide tools for viewing and ana- lyzing the multitude of data in the characteristics and prioritizing the crossing inventory. Department of Transportation (KDOT). Figure 1 Kansas Rail Network 6 Burns & McDonnell Field Collection Ranking Formula Development each variable, such as average daily KDOT was doing a re-inventory of To select locations that have the most traffic and accidents. Regression about 10 counties per year. With 103 potential for accidents, KDOT uses a analysis is used to determine what counties in Kansas that have rail serv- multi-phase safety improvement deci- significant linear or non-linear (e.g. ice, crossings were updated once sion process. The first phase involves logarithmic, exponential or quadratic) every 10 years. With the help of cre- the initial ranking of the crossings relationship would best predict future ative staffing techniques, Burns & using an empirically derived formula accidents. McDonnell was able to visit all public to create a list of candidate locations. highway/railroad crossings in the The formula takes into consideration Application Development state within 12 weeks. characteristics such as volume of The existing crossing database is highway traffic and train traffic, train stored as a flat file within an outdated To accomplish the field inventory, a speed, crossing angle and number of mainframe computer. KDOT is mak- study was conducted to determine tracks. Once a candidate list is gener- ing an organization-wide conversion what items were important to collect ated, a diagnostic team conducts a to an Oracle relational database plat- and the most efficient way to collect field review to verify the existing con- form. This presents an excellent them. Time and motion tests were also ditions and determine the appropriate opportunity for KDOT to take advan- conducted to forecast the amount of type of upgrade for the crossing. tage of recent developments in infor- labor required to complete collection mation technology to help make within the project schedule. Because The current empirically derived for- informed decisions about highway/ of the quantity of data being collected mula is being evaluated to verify that railroad grade crossing safety (over 400,000 field items) it was also it still reflects recent accident trends. improvements. Burns & McDonnell necessary to develop an efficient The evaluation consists of a two-step evaluated several alternatives before it means of storing and transferring the process involving an analysis of past was determined that a unique soft- data to the office for incorporation accidents and a statistical analysis. ware application was needed to meet into the database. all of KDOT’s functionality require- The past accident analysis combined ments, including the ability to: Burns & McDonnell recruited 32 col- the Kansas accident database with the • Generate queries (data requests) lege students majoring in engineering new inventory data. To help identify and reports from a graphical (map) and GIS technologies to complete the trends, graphs were generated to rep- interface field work over the summer. The resent the distribution of accidents • Automate the ranking equation crews collected more than 60 charac- with respect to the variables used in under different "what if" scenarios teristics about each location using GPS the rating formula and the breakdown • View digital photographs receivers, digital cameras and elec- of number of accidents by type of tronic data collectors. The data includ- warning device. The application, Crossing Inventory ed road/rail intersection coordinates, Information Management System, or 24-hour traffic counts, photographs, The statistical analysis began by CIIMS (Figure 2) will be developed warning signs and crossing device assembling the data set and splitting it using Visual Basic 6.0 to lace together inventories. into two using stratified randomiza- a wide variety of functions under a tion to minimize biases; one set is single program. One of those func- The majority of data was captured in used for formula development/testing tions includes spatial analysis. Spatial the field on pen-based computers that and the other for validation. Once the operated on a Windows CE environ- data sets are assembled, a Pearson ment. The crossing data and digital correlation matrix is developed show- photos were stored on 10 MB "flash ing the correlative strength between cards." At the end of each day, the crews uploaded the data files and photographs to the home office FTP (File Transfer Protocol) site. TechBriefs 7 analysis is carried out with geograph- Conclusion ic information system (GIS) controls KDOT realizes that without sufficient that are imbedded in the application. tools for analysis, it is difficult to take Spatial analysis will allow the user to advantage of the abundance of infor- analyze multiple layers of information mation that is available through the from both a geographic and tabular personal computer. Therefore, KDOT interface to help determine which is implementing recent advancements crossings have the most potential for in data collection, data storage, data accidents, which crossings are candi- management and the analysis of infor- dates for closure and which ones mation. The utilization of information should be upgraded. The application technology will greatly enhance their Ken Gerling is a project man- will also convert the data to the method for inventorying grade cross- ager in Burns & McDonnell’s Federal Railroad Administration’s ings and identifying those that have transportation department. He (FRA) required format so that it can the highest potential for accidents in has a bachelor’s degree in civil be electronically transferred to update an effort to improve safety for the national grade crossing inventory. motorists across Kansas. engineering from the University of Missouri-Kansas City and a master’s degree in engineering management from the University of Kansas. Cleve N. DeVault is manager of Burns & McDonnell’s GIS Services. He has a bachelor’s degree in history from Ohio Wesleyan, a bachelor’s degree in civil engineering and a master’s degree in geotechnical engineer- Figure 2 ing, both from Ohio State Crossing Inventory Information Management System University.
Pages to are hidden for
"HighwayRailroad Grade Crossing Studies"Please download to view full document