PHASE ONE PROJECT REPORT

MOORHEAD AREA INTEGRATED TRAIN DETECTION AND TRAFFIC CONTROL SYSTEM PHASE ONE PROJECT REPORT December 2000 Prepared for: Minnesota Department of Transportation Office of Advanced Transportation Systems 395 John Ireland Blvd, Mail Stop 320 St. Paul, MN 55155 (651) 296-8602 Prepared by: SRF Consulting Group, Inc. Suite 150 One Carlson Parkway Minneapolis, MN 55447 (763) 475-0010

TABLE OF CONTENTS Page 1 PROJECT OVERVIEW... 1 1.1 INTRODUCTION... 1 1.1.1 Problem Statement... 1 1.1.2 Proposed Solution... 3 1.2 PROJECT GOALS... 3 1.3 PROJECT TEAM DESCRIPTION... 4 2 PHASE ONE SYSTEM DESCRIPTION... 5 2.1 SYSTEM SCHEMATIC... 5 2.1.1 Component One: Detection... 5 2.1.2 Component Two: Communications... 5 2.1.3 Component Three: Host-End System... 5 2.2 DATA COLLECTION REQUIREMENTS... 9 3. PHASE ONE SYSTEM DEPLOYMENT... 12 3.1. HARDWARE DEPLOYMENT... 12 3.2. SOFTWARE DEPLOYMENT... 12 3.3. DEPLOYMENT PROBLEMS ENCOUNTERED... 12 4. FUTURE PROJECT PHASES... 14 4.1 GRAPHICAL USER INTERFACE (GUI) FOR EMERGENCY DISPATCH (PHASE 2)... 14 4.2 TRAFFIC SIGNAL PREEMPTION OUTPUTS (PHASE 2)... 14 4.3 VARIABLE MESSAGE SIGN OUTPUTS (PHASE 3)... 16 5. CONCLUSIONS... 17 5.1 LESSONS LEARNED... 17 5.2 NEXT STEPS... 17 6. REFERENCES... 18 i

1 PROJECT OVERVIEW 1.1 INTRODUCTION The Minnesota Department of Transportation (Mn/DOT) and the City of Moorhead are implementing an integrated train detection and traffic control system. The goal of this project is to develop a system to track train movements through the Fargo-Moorhead area and use this information to reduce motorist delay and improve emergency vehicle response. The project team is comprised of individuals from the following agencies/organizations: Federal Highway Administration (FHWA), Mn/DOT (Office of Advanced Transportation Systems, Office of Freight, Rail and Waterways, and District 4), the Cities of Moorhead and Fargo, the Fargo-Moorhead Council of Governments, Image Sensing Systems, Electronic Design Company, Burlington Northern Santa Fe (BNSF) Railroad and SRF Consulting Group, Inc. The project is divided into three phases. This report presents detailed information on the deployment of Phase 1. An overview of the project phases is provided below: Phase 1: Phase 2: Phase 3: Deploy demonstration test site consisting of a video-based train detector, wireless communication infrastructure, and host-end system. Expand system to include multiple detection sites. Interface with signal systems to implement alternative timing plans. Graphical User Interface (GUI) to provide emergency dispatchers with real time information. Expand system to include additional detection sites as warranted. Deploy Variable Message Signs (VMS) for motorist information. 1.1.1 Problem Statement Downtown Moorhead is divided by two BNSF lines that run east-west through the City. These tracks include a north line, which runs between Center Avenue and 1st Avenue, and a south main line, which runs between Center Avenue and Main Avenue. On a typical day, ten trains use the north line and 50 trains use the south line. The maximum train speed in the downtown area is 35 miles per hour. Refer to Figure 1 for the location of rail operations in the Fargo-Moorhead area. Train operations block the north-south roadways for an average of 4.5 minutes per train. This equates to a total blockage of 4 hours and 30 minutes (almost 20 percent of the day) at each roadway segment that crosses the railroad tracks on a typical day. The existing average daily traffic volume for 1st Avenue is 15,000 vehicles, for Center Avenue it is 10,000 and for Main Avenue it is 20,000 vehicles. 1

FIGURE 1: THE LOCATION OF RAIL OPERATIONS IN THE FARGO-MOORHEAD AREA 2

This train activity has several impacts to traffic in the cities of Moorhead and Fargo. First, the presence of trains degrades the operation of intersections controlled by traffic signals. Second, the presence of trains increases the response time of emergency vehicles. 1.1.2 Proposed Solution The Moorhead area integrated train detection and traffic control system will increase safety and reduce motorist delays related to the presence of trains in the corridor. This will be accomplished through implementation of alternative traffic signal timing plans, an interface that provides emergency dispatchers with real time information on train movements within the corridor, and variable message signs to disseminate train blockage information to motorists. 1.2 PROJECT GOALS The Moorhead Area Integrated Train Detection and Traffic Control System project establishes the following goals: A. Reduce delay for motorists and public transit through: Improved signal timing, specifically when trains are present Providing real-time information to motorists to enable them to select an alternate at-grade crossing or divert to a grade-separated crossing Providing real-time information to the transit operators to enable route diversions B. Improve Emergency Service delivery through: Providing real-time information to emergency vehicle operators for auto and train traffic C. Improve safety through: Reducing exposure at at-grade crossings Reducing conflicts with emergency vehicles for auto D. Develop, implement and test a system that may be beneficial for many cities throughout the country. 3

1.3 PROJECT TEAM DESCRIPTION The project team is comprised of individuals from the following agencies/organizations: FHWA, Mn/DOT, the Cities of Moorhead and Fargo, the Fargo-Moorhead Council of Governments, Image Sensing Systems, Electronic Design Company, BNSF Railroad and SRF Consulting Group, Inc. Project Team members include the following: James McCarthy, FHWA Farideh Amiri, Mn/DOT Office of Advanced Transportation Systems James Kranig, Mn/DOT Office of Advanced Transportation Systems Bob Martin, City of Moorhead Tom Sopp, City of Moorhead Rick Lane, City of Fargo Durga Panda, Image Sensing Systems Todd Grandt, Electronic Design Company Brian Scott, SRF Consulting Group, Inc. Erik Minge, SRF Consulting Group, Inc. Mark Gallagher, SRF Consulting Group, Inc. Bruce Shriver, SRF Consulting Group, Inc. 4

2 PHASE ONE SYSTEM DESCRIPTION 2.1 SYSTEM SCHEMATIC Phase 1 of the train detection and traffic control system consists of three main components: detection, communications and host-end processing. These components have been integrated into a comprehensive system that can estimate train arrival and clearance times at crossings within Fargo and Moorhead. Refer to Figure 2, System Schematic 2.1.1 Component One: Detection The train detection component utilizes the Autoscope Solo TM video-based sensor to detect the presence, speed, length and direction of trains. The sensor processes the video image inside the sensor itself, outputting both video and processed data. For Phase 1 an Autoscope Solo was placed on the 34th Street Bridge in Moorhead. Future phases will deploy sensors at additional locations throughout the Fargo-Moorhead corridor, including the primary east-west line and additional lines to the south and southeast of Moorhead. Refer to Figure 3, 34th Street Demonstration Site Photos. In addition to video detection sites, permission is being requested from BNSF to obtain preemption status at selected crossings in the Fargo-Moorhead area. This will provide a costeffective data point regarding the presence of a train at or near select crossings. 2.1.2 Component Two: Communications The communication component of the system connects the detection sites to the host-end system. A wireless communication link was used for the test site deployment. Future phases will likely require a combination of wireless and twisted pair communication. Spread spectrum radio equipment provides the wireless component. Direct line-of-sight is required with this equipment. Both data and video are transmitted from the sensor to the hostend system via RS-232. Similarly, the host-end transmits data commands to the video sensor. The video is transmitted via 2.4 GHz and the bi-directional data is transmitted via 900 MHz. Refer to Figures 2 through 4. 2.1.3 Component Three: Host-End System The host-end system records, stores and displays detection information. Train detection data is received from the field site and automatically written to the host system s database. The system s software monitors the real time data in the database in order to predict train movements within Fargo and Moorhead area. Future phases of the project will utilize the train operations information to implement alternative traffic signal timing plans. The Moorhead Wastewater Treatment Facility houses the host-end equipment. This location was selected to take advantage of the existing communication infrastructure. Refer to Figure 4. 5

FIGURE 2: SYSTEM SCHEMATIC Moorhead Area Train Detection - Component Overview Wastewater Treatment Facility Trango 2.4 GHz Video Receiver United Signal 900 MHz Data transceiver 34th St. Bridge Autoscope Trango Video Transmitter (2.4 GHz) United Signal 900 MHz Data transceiver Data Processing Computer Mounting pole Detection Zone System Components Detail - Host Site System Components Detail - Field Site +24dB Parabolic Antenna Autoscope MVP Data Processing Computer Trango 2.4 GHz Video Receiver Yaggi Antenna United Signal 900 MHz Data Transceiver Yaggi Antenna United Signal 900 MHz Data Transceiver Autoscope Interface Cable Autoscope Interface Panel RS-485 Data Patch Antenna RS-485 to RS-232 Converter Wastewater Treatment Facility Communications Tower Trango 2.4 GHz Video Transmitter Coaxial Video 6

FIGURE 3: 34TH STREET DEMONSTRATION SITE 7

FIGURE 4: WASTEWATER HOST-END SYSTEM 8

2.2 DATA COLLECTION REQUIREMENTS The train detection system utilizes real-time information on train speed, length, and location. Accurate data collection is essential to the successful operation of the system. During the preliminary design process several detection approaches were considered. BNSF personnel indicated that the detector had to be non-intrusive to the track environment. A clearance of 25 feet was required between the centerline of the track and any obstacles. Various detection technologies were analyzed with these requirements in mind. Two detection approaches were identified as leading candidates, video and ultrasonic, refer to Table 1. Table 1: Train Detection Technologies Review Technology Speed Sidefire Bi-directional Cost Sensors Total Cost Recommendation Capability Capability Capability Per Sensor Per Site Passive Varies Yes Varies $1,000 2 $2,000 Infrared Pulse Ultrasonic Yes (Paired) Yes Both $1200 4 $4,800 Yes (Second Choice) Magnetic Yes No Both $800 4 $3,200 No (Paired) Doppler Yes No Both $800 2 $1,600 No Microwave Video Yes Yes Both $8,000 1 $8,000 Yes (First Choice) Passive Yes Yes Both $2,500 2 $5,000 Acoustic Active Yes Yes Both $6,500 2 $13,000 Infrared Radar Yes Yes Both $3,300 2 $6,600 Radar/PIR Yes Yes Oncoming Only $2,400 2 $4,800 Video-based detection technology was selected for the system s detection. This choice was made for several reasons. First, video detection allows for detection in multiple detection zones from one camera. This has the advantage of using one camera to detect train activity on two different tracks. This reduces the installation and communication costs. Second, a video image of the detection zone is available for viewing at the host-end site. A real-time image provides visual confirmation of train activity. This is useful during system setup and during normal system operations. The video-based detection sites utilize the Autoscope Solo TM video detector developed by Image Sensing Systems. This commercially available video detection system uses machine logic to process video images in order to generate detector outputs. The Solo TM was selected because the sensor has been used for train detection in previous applications. Additionally, Image Sensing Systems, located in St. Paul, Minnesota, has agreed to modify the sensor for optimal train 9

detection performance and provide on-site support. The video detection system provides the following outputs: 1. Train Presence 2. Train Speed 3. Train Direction 4. Identification of Track being used 5. Detector Location In addition to direct observations of train activity, the status of various crossing gates will be used to provide train presence information at selected at-grade crossings. The status of the crossing gates is received in the form of a high or low signal. A high signal indicates that the crossing gate is activated due to train activity at the crossing. There is no way to know when the train is actually at the crossing, only that the crossing gate has been activated. The crossing gate information is received either directly from the BNSF track-side bungalow, or from traffic signal controllers that are already interfaced with the railroad crossing. A video detector was installed on the 34th Street Bridge in Moorhead during the summer and fall of 2000. This site served as a technology demonstration site, further deployments are planned along the corridor. Refer to Figure 5 for the location of the demonstration site and for the future video and crossing gate status detection locations. 10

FIGURE 5: THE LOCATIONS OF THE DEMONSTRATION SITE, FUTURE VIDEO AND CROSSING GATE STATUS DETECTION 11

3. PHASE ONE SYSTEM DEPLOYMENT 3.1. HARDWARE DEPLOYMENT The train detection system hardware was initially tested in the Twin Cities Metropolitan Area to ensure the hardware components operated together as a complete system before they were installed in Moorhead. Next, the communication system, detection system, and host end equipment were installed in Moorhead during the summer and fall of 2000. A bucket truck was needed to install the communication equipment on the communications tower at the Moorhead Wastewater Treatment Plant. Traffic control was needed to close a lane on the 34th Street Bridge in order to work on the shoulder and install the detection and communication hardware. Refer to Figures 3 and 4 for pictures of the hardware deployment at the 34th Street Bridge and at the wastewater plant, respectively. 3.2. SOFTWARE DEPLOYMENT The train detection software was developed to provide communication between the train detection software and the host end computer s database. All of these software components reside on one computer located at the Wastewater Treatment Plant. 3.3. DEPLOYMENT PROBLEMS ENCOUNTERED As is common with any ITS operational test, unforeseen circumstances can create problems in system deployment. The Moorhead Train Detection Project was no exception. By far the most significant problems were encountered with the wireless communication equipment. The system requires full-duplex data and video communication at 9600 baud. Achieving this level of performance proved to be a time-consuming and complicated task. The Trango TM communication equipment was initially bench tested by EDC in June 2000. The equipment provided basic functionality during this inspection. The equipment was then installed at SRF for testing in a mock system setup. During this period several communication problems surfaced. These problems generally involved intermittent loss of data communication. Numerous modifications were made until the system was felt to perform satisfactorily. Next, the system was installed in the field in Moorhead. Once again, however, problems surfaced in the data communication component of the system. A bit error rate tester revealed a very poor level of performance with the equipment. Project team members made several attempts to make the system operational but failed. Concern about the line-of-sight between the 34th Street Bridge and the Wastewater Treatment Plant and interference from other sources were suspected, but could not be confirmed. The team decided to bring the equipment back to the Twin Cities for further evaluation. Back in the Twin Cities the project team members evaluated the system in greater detail, setting up the equipment alongside Interstate 35W at a distance of 1.5 miles. This evaluation revealed a problem with the RS 232 to RS 485 converter. Believing the system was now fixed; the 12

communication equipment was re-installed in Moorhead in August 2000. However, the equipment still did not meet the system s communication requirements. Further investigation revealed that the communication equipment was not capable of supporting the 9600 baud rate required by the data portion of the system. The communication equipment vendor was asked to investigate this problem and, after conducting some tests, indicated that the equipment was not capable of meeting this requirement, even though it was listed in their product s specifications.. At this point the project team began evaluating different communication vendors and selected United Signal to provide the data portion of the communication system. The original Trango equipment was retained to provide the video portion of the communication. Ultimately, several months were spent configuring the communication equipment. 13

4. FUTURE PROJECT PHASES Future phases of the project will build on the deployment at 34th Street to include multiple detection sites, a GUI for emergency dispatch, interface with signal systems, and variable message signs. 4.1 GRAPHICAL USER INTERFACE (GUI) FOR EMERGENCY DISPATCH (PHASE 2) Currently, emergency vehicle operators and dispatchers have little information regarding train locations and speeds they need to select the best route for crossing the railroad during train movements. Detailed information on train operations in the Fargo-Moorhead area will provide a valuable tool in dispatching emergency vehicles. The host-end computer will record, store and display the projected train operations. The interface will provide dispatchers with the train s estimated time of arrival at each at-grade crossing. The interface will also indicate the estimated time for the crossing to clear. Armed with this information, dispatchers will be able to deploy emergency vehicles more quickly and more efficiently. 4.2 TRAFFIC SIGNAL PREEMPTION OUTPUTS (PHASE 2) Extensive motorist delays are experienced where heavily traveled roadways cross the railroad tracks in downtown Moorhead. Given their proximity to at-grade crossings, five signalized intersections in the area are currently railroad preempted. While this preemption does reduce delay for some motorists, the preemption plan does not go into effect until the train is in the immediate vicinity of the crossing. Also, there are 14 additional signalized intersections that are not preempted. The current traffic signal system does not adequately respond to the presence of trains in the area. Advanced train detection provides the opportunity to implement a set of alternative signal timing plans. The existing individual intersection signal systems would be interconnected to arterial closed loop style master controllers. The detection of a critical length train would then trigger the master controllers to select the alternative timing plan. This plan allows north-south movements to clear prior to a train s arrival, and then favors east-west movements parallel to the tracks until the train has cleared the area. A flow chart indicating this operation is provided in Figure 6. The detection of a train will result in the implementation of an alternative timing plan. A different timing plan will be evoked depending on the length of the train and the train s direction. As a starting point, the presence of a train must result in a crossing closure that is equal to or greater than two signal cycle lengths. At 35 mph, this translates to a train that is at least 1,500 feet long. The new timing plan will result in a change to the signal system s splits. The cycle length and offset will not change. 14

FIGURE 6: PROPOSED SIGNAL PLAN FLOW CHART Presence of Train at 34th St T > Tb at 34th St.? No Yes Special Timing Plan No Train Leave End Point (4th St.)? Yes Normal Timing Plan End Note: T - Detected Train Blocking Time Tb - Critical Train Blocking Time (Threshold) 15

4.3 VARIABLE MESSAGE SIGN OUTPUTS (PHASE 3) A 1998 survey of drivers in the Fargo-Moorhead area assessed the public perception of using ITS technologies to address delays at highway-rail intersections. One of the survey questions asked, How would the message sign help you? Seventy-four percent of the respondents said they might turn or would definitely turn if given a message that an approaching train would cause a ten-minute delay. A set of strategically placed VMSs can inform motorists of projected train delays in advance of railroad crossings. This timely information can allow motorists to select alternative routes, thereby avoiding delays. The system s host computer would control all of the signs. Wireless or twisted pair communication would be used, depending on the signs location. While detailed design of the signs has not been completed, the sign would most likely indicate the number of minutes that the nearest railway crossing will be blocked. 16

5. CONCLUSIONS Travel in downtown Moorhead is severely affected by the amount of rail traffic. Drivers routinely experience four-minute delays when one of dozens of trains block at-grade crossings. The train detection and traffic control system will reduce motorist delay and improve emergency vehicle response. This project utilizes video-based detection of train movements in the rail corridor. The detection information is used to implement alternative traffic signal timing plans, provide emergency dispatchers with real time information on train movements within the corridor, and activate message signs to disseminate train blockage information to motorists. 5.1 LESSONS LEARNED As with any its operational test, there are, fortunately, lessons to learned. In the Moorhead Area Integrated Train Detection And Traffic Control System significant problems were encountered with the system s communication system. Following are some of the lessons learned from the system deployment: A detailed survey should be made of the proposed communication sites. This site survey should examine all factors that could affect the performance of wireless communication equipment. A sound communication system design should specify hardware that has been used successfully in similar applications. The communication system should be fully tested in its final configuration before deploying in the field. 5.2 NEXT STEPS The next step in the Moorhead Train Detection Project is to expand the system to include additional detection sites and additional system functionality. The following steps are identified: Phase 2 Select additional video detection site(s). Identify at-grade crossings at which to receive information from the railroad on the status of crossing gates. This will include both crossings that are interconnected with traffic signals for railroad preemption, and crossings that are not preempt. Communication system design between detection sites, host site, traffic signal system, and emergency dispatch personnel. Develop GUI for emergency dispatchers. Integrate train detection system with traffic signal system. Phase 3 Select additional video detection and/or crossing gate preemption sites. Deploy VMS system for motorist information. 17

6. REFERENCES Analysis of Motorist Attitudes to ITS Application to Rail-Highway Crossings in the Fargo- Moorhead Metropolitan Area, Upper Great Plains Transportation Institute, North Dakota State University, April 1998. Moorhead Area Integrated Train Detection and Traffic Control System: Scoping Study, Minnesota Department of Transportation, June 1998. 18