7/28/2019 rr0710

1/4

U.S. DepartmentOf Transportation

Federal RailroadAdministration

RR07-10April 2007

Study of High-Speed Wireless Data Transmissions forRailroad Operation

SUMMARYWith the widespread deployment of high performance wireless data network technologies (such asInstitute of Electrical and Electronics Engineers (IEEE) 802.11a/b/g, 802.16e, etc.), it has becomefeasible to utilize these standard-based wireless technologies for applications desired by the railroad

industry for improving operations effectiveness, monitoring and control, and safety. Before any of thesewireless systems are employed, however, serious questions must be answered to assure that the railroadrequirements of mobility, throughput, coverage range, Doppler effects, and response times are met.

The University of Nebraska, under a grant from the Federal Railroad Administration (FRA), has beenstudying the feasibility of using 802.11a/b/g-based wireless networks for mobile railroad environments. A3.5-mile test bed has been designed and implemented with a series of access points in the BurlingtonNorthern Santa Fe Railroad (BNSF) Hastings Subdivision near Lincoln, NE, to support this study. Thetest bed has complemented the theoretical study supported with in-depth computer models andsimulations to measure and analyze network throughput of 802.11a/b/g for trains under differentvelocities. Simulation results have been compared and verified with test bed measurements to analyzeperformance characteristics of 802.11-based networks for mobile trains. Results and analysis show thateven though the system throughput of 802.11 standard decreases under high velocity, it can still supportrailroad applications if the coverage range is provided. Therefore, an 802.11-based wireless network cansupport real-time Internet accessibility for trains crew, passenger, and railway workers, improving railroadoperating safety and efficiency.

Figure 1. High-Speed Wireless Data Network for Railroad Operation

Internet

7/28/2019 rr0710

2/4

US Department Of Transportation

Federal Railroad Administration Research Results RR07-10

Page 2

INTRODUCTIONNear real-time communication between groundinfrastructure and moving locomotives and othervehicles provides rapid response capability tounexpected events. It may prevent injury or

minimize the impact of an incident. Other real-time communications, such as event recorderdownload, onboard health monitoring system forlocomotives and trains, and bettercommunication between the dispatching officeand trains, can improve the efficiency of railroadoperation. Newer broadband wirelesstechnologies offer the potential to support thistype of communication. Many major issues,however, remained unanswered for thesuitability and performance of these networks.

This study with field test data has investigatedthese issues and its feasibility for railroadapplications before deployment.

TECHNICAL DESCRIPTION

SimulationThis study investigated the 802.11b standardunder mobile scenarios for a variety of velocitieswith computer modeling. The indepth modelswere designed and implemented to incorporateRayleigh fading distribution and Doppler shiftinto the NS-2 simulation environment in order toprovide an approximation of bit error rates based

on the channel and environment characteristics.Support for multiple channels, channel scans,and switches, as well as data rate adaptation,were also added as enhancement of the models.

802.11a/b/g were designed and coded in themodel. Physical layer characteristics weresimulated to determine the impact of noise,fading, and Doppler shift, as well as to calculateeffects of shadowing, Rician fading, andRayleigh fading. Medium Access Control (MAC)and Data Link Layer protocols in the modelimplemented functions, such as fragmentationand de-fragmentation, data retransmission,

multi-rate support, multiple channel scanning,synchronization, power management,authentication, association, and handoff. Thesimulation results display the performanceevaluations of mobile scenarios under a varietyof velocity and transmission configurations.These results would demonstrate the low andhigh-speed impacts on system throughput fordifferent data rates under 802.11b. This

includes data rates of 1 to 11 Mbps for velocitiesup to 90 mph.

Test BedA test bed was designed and implemented to

provide field data based on industryimplementation of 802.11a/b/g. The test bedwas chosen carefully to include differentterrains. Measured data from the test bed wasused to complement analysis of theoreticalstudies and simulation results to understandnetwork throughput under the real operatingenvironment.

BNSF WAN

Microwave Link

VPN into

BNSF WAN

DSL Link

CC

BNSF Core Network

Router/Gateway

AP7 AP6 AP5 AP4 AP3 AP2

NC

NCNCNC CCCCCCCCNCNCNCNC CCCCCC

AP8

Crete DepotAP1

Berks East

Internet

Microwave Microwave

Figure 2. Test Bed with Planted 802.11Access Points in Test Bed

A 3.5-mile section at Crete, NE, in the BNSFHastings Subdivision was chosen for its closeproximity to the university and challengingenvironment with heavy foliage, curves, and

surrounding hills. BSNF contributed by installingthe hardware, including the 5.7-mile microwavetower extension and link. The test bed utilizesthe 802.11 technology to support wirelessconnectivity between moving trains and fixedaccess points. The test bed was also designedto support equipment from different vendors, aswell as upgrading to the upcoming technologiessuch as the 802.16e (mobile WiMAX) and802.20. Several Class I railroads and the

Association of American Railroads (AAR) wereinvolved in the design of the test bed and theselection of the vendor. Strix Systems

equipment was finally selected for the network.A total of eight access points was deployed inthis 3.5-mile segment. The farthest spacebetween access points is approximately 1.2miles where the track is tangent, but, in curves,they have to be deployed much closer togetherfor seamless coverage. Figure 3 shows a curvewhere access points have to be closer becauseof the curve obstruction. Figure 4 shows the

7/28/2019 rr0710

3/4

US Department Of Transportation

Federal Railroad Administration Research Results RR07-10

Page 3

antenna coverage for this 3.5-mile stretch oftrack.

For most field tests, a hy-rail vehicle was used torepresent a train. For the high-speed test, the

vehicle traveled on a highway adjacent andparallel to the railroad track. Wayside computerequipment connected to this network was usedat one end of the test bed to receive from andtransmit data to the hy-rail vehicle. A series oftests was performed, including operationalverification, various stationary, and mobilitytests, between October 2005 and December2006. The throughputs in terms of Mbps weremeasured and compared to the maximumtransmission rate of 11 Mbps.

Figure 3. Access Point Antenna Pole onthe Farther Side of Curve

Figure 4. Coverage in the Test Bed

RESULTSFigure 5 depicts the throughput measured under55 mph in the test bed compared with thesimulation results for a similar scenario. Itclearly shows that the simulation results match

measured results from test bed. This is one ofthe sample results in verification of thedeveloped simulation model.

Figure 6, as a part of the actual field test results,illustrates the round trip delay time of a

message. As the vehicle was moving away fromAccess Point (AP) 8 (Crete Depot), the roundtrip time for a message was measured as afunction of the distance from this access point.Generally, as expected, as the distanceincreases, the delay increases, but notsignificantly.

Figure 5. Sample Simulation Results

Figure 6. Received Signal StrengthIndication and Round Trip Time

Figure 7 shows the throughputs as the vehicle isfarther from AP 2. The throughput was measuredwith a stationary vehicle. Figure 8 shows thethroughput results with velocities of 45 and 90 mph.

The impact of handoff is clearly shown, and it alsoillustrates the insignificant impact of higher velocity onthroughput. Additionally, audio and videoapplications were tested with this network. Out of apossible Mean Opinion Score of 5, the quality of theaudio streaming was rated around 4.5 consistently.The video streaming sent from the caboose or thehighway vehicle was recorded with good quality andwith little break in the picture composition.

Intel Measured RSSI vs. Distance from Crete Depot (AP 8)

-100

-80

-60

-40

0 1000 2000 3000 4000 5000 6000 7000

Distance(m)

RSSI(dBm

)

Roundtrip Delay vs. Distance from Crete Depot (AP 8)

0

20

40

60

80

100

0 1000 2000 3000 4000 5000 6000 7000

Distance (m)

RoundtripDelay

(ms)

7/28/2019 rr0710

4/4

US Department Of Transportation

Federal Railroad Administration Research Results RR07-10

Page 4

Throughput vs.Time

0

1

2

3

4

5

6

7

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Time (second)

Throughput(Mbps) Throughput

Distance vs. Time

0

200

400

600

800

1000

1200

1400

1600

1800

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Time (second)

Distance(m)

Distance

Figure 7. Stationary Test of Throughput

as a Function of Distance from AP

Figure 8. Throughput Comparisonbetween 45 and 90 mph

CONCLUSIONSBoth the modeling and field tests showed thatthe broadband communication technology basedon IEEE 802.11a/b/g protocols can be applied torailroad operation for high-speed mobile datatransmission. Future protocols, such as 802.16and 802.20, will enable significantly longerspacing between access points, making thistype of infrastructure buildup a viable datacommunication network for applicationsenhancing safety and efficiency of railroadoperations.

ACKNOWLEDGMENTSThis research effort at the AdvancedTelecommunications Engineering Laboratory(TEL) at the University of Nebraska-Lincoln is

funded by FRA. Extensive assistance and costcontributions are provided by BNSF Railway andUnion Pacific Railroad, as well as support from

AAR, CSXT, Norfolk Southern Railway,Canadian National Railway, and CanadianPacific Railway. Dr. Sharif is the director of TELand the Principal Researcher for this project.

REFERENCES[1] Performance of IEEE 802.11b under MobileRailroad Environments, IEEE 62nd VehicularTechnology Conference (VTC2005), Dallas,Texas.

[2] A Wireless Test Bed for Mobile 802.11 andBeyond, International WirelessCommunications and Mobile ComputingConference (IWCMC2006), Vancouver, Canada.

[3] BER Performance of 802.11b Systems overMobile Wireless Channels, Accepted for IEEEInternational Conference on Communications(ICC-2007), Glasgow, Scotland.

[4] Comparison of Throughput Performance forthe IEEE 802.11a and 802.11g Networks,

Accepted for IEEE 21st International Conferenceon Advanced Information Networking &

Applications (AINA-07), Niagara Falls, Canada.

CONTACTTerry TseFederal Railroad AdministrationOffice of Research and Development1120 Vermont Ave. NWMail Stop 20Washington, DC 20590Email: Terry.Tse@dot.gov

KEYWORDSIEEE 802.11a/b/g, mobility, throughput, railwaycommunications, performance