FIELD EXPERIMENT TO IDENTIFY POTENTIALS OF APPLYING BLUETOOTH TECHNOLOGY TO COLLECT PASSENGER VEHICLE CROSSING TIMES AT T H E U. S. -MEXICO BORDER by Rajat Rajbhandari Texas Transportation Institute Project performed by July 2009 Report prepared by Texas Transportation Institute 4050 Rio Bravo, Suite 151 El Paso, Texas 79902 TEXAS TRANSPORTATION INSTITUTE The Texas A&M University System College Station, Texas 77843-3135
TABLE OF CONTENTS List of Figures... iii List of Tables... iii Disclaimer and Acknowledgments... iv Executive Summary... 1 Chapter 1: Background and Problem Statement... 2 Chapter 2: Bluetooth Study at the U.S.-Mexico Border... 3 Chapter 3: Bluetooth Technology... 4 Chapter 4: Data Collection... 6 Field Equipment Setup... 6 Results... 12 Chapter 5: Conclusion... 19 Bibliography... 20 Page Texas Transportation Institute Page ii
LIST OF FIGURES Figure 1. Ports of Entry along the U.S.-Mexico Border.... 6 Figure 2. Typical Location of Bluetooth Readers at a Port of Entry.... 7 Figure 3. Location of Bluetooth Readers at U.S. and Mexican Sides of the Bridge of the Americas POE.... 8 Figure 4. Location of Bluetooth Readers at U.S. and Mexican Sides of the Ysleta POE.... 9 Figure 5. Location of Bluetooth Readers at U.S. and Mexican Sides of the Paso Del Norte POE.... 10 Figure 6. Installation of Bluetooth Reader on Top of a TTI Truck on the U.S. Side.... 11 Figure 7. Installation of Bluetooth Reader on Top of a TTI Vehicle on the Mexican Side.... 11 Figure 8. Undergraduate Student Conducting Intercept Survey of Motorists.... 12 Figure 9. Comparison of Measured and CBP-Reported Crossing Times at Paso Del Norte POE.... 14 Figure 10. Comparison of Measured and CBP-Reported Crossing Times at Ysleta POE.... 15 Figure 11. Comparison of Measured and CBP-Reported Crossing Times at BOTA POE... 15 Figure 12. Placement of Bluetooth Readers at BOTA POE.... 16 Figure 13. Result of Moving the Bluetooth Reader Further Upstream on the Mexican Side... 17 Figure 14. Intercept Survey of Motorists Entering the United States.... 18 Page LIST OF TABLES Table 1. Results from the Field Experiment, Including Number of Matching MAC Addresses. 13 Table 2. Results from the Intercept Survey.... 18 Page Texas Transportation Institute Page iii
DISCLAIMER AND ACKNOWLEDGMENTS This research was performed by the Center for International Intelligent Transportation Research, a part of the Texas Transportation Institute. The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the data presented herein. The research team thanks Turnpike Global Technologies for providing the Bluetooth readers. The research team also thanks undergraduate student interns Noel Chavez, Rodrigo Torres, and Daniel Arroniz for providing assistance with the field data collection. Texas Transportation Institute Page iv
EXECUTIVE SUMMARY Millions of motorists cross the U.S.-Mexico border every day. Currently, U.S. Customs and Border Protection (CBP) is the only agency that provides estimated wait times, updated hourly, to those motorists waiting to cross the border. CBP collects border crossing times by visual identification of the queue length and surveys of motorists waiting to cross the border. However, these methods often produce unreliable wait times. Local and regional transportation agencies are interested in gathering more accurate border crossing times, which are crucial to efficient operation at the border and are important to motorists. The Texas Transportation Institute in collaboration with Turnpike Global investigated the feasibility of using Bluetooth technology to gather more accurate, scientifically based border crossing times of passenger vehicles at the U.S.-Mexico border. Bluetooth technology has been used extensively at ports of entry (POEs) along the U.S.-Canada border to measure crossing times but has not yet been used along the U.S.-Mexico border. To determine the feasibility of using Bluetooth at the U.S.-Mexico border, researchers collected data at three POEs in the El Paso region: the Bridge of the Americas, Ysleta, and Paso Del Norte. At each POE, a Bluetooth reader on each side of the U.S.-Mexico border gathered data from drivers and passengers Bluetooth-enabled devices to determine border crossing times. Researchers manually counted the passenger vehicles entering the United States and performed an intercept survey to determine how many motorists had Bluetooth-capable mobile devices. The study confirmed that Bluetooth technology could be used to measure U.S.-Mexico border crossing times. An advantage of using this technology is that it does not require any special software to run on motorists devices. The disadvantages of using this technology are that the physical geometry and queues at a crossing may affect the number of readers that can be installed and that Bluetooth technology requires complex algorithms to determine average crossing times. The researchers recommend that a pilot project be conducted before full-fledged implementation to minimize potential problems. Though the Bluetooth readers do not gather any private information, a transportation agency considering using this technology should inform the public of how the technology works and how crossing times are estimated and used. Texas Transportation Institute Page 1
CHAPTER 1: BACKGROUND AND PROBLEM STATEMENT Millions of motorists cross the U.S. and Mexico border every day for work, business, and social activities. The El Paso region has three ports of entry, through which 13 million passenger vehicles crossed into the United States in 2008. The crossing time of northbound passenger vehicles crossing into the United States easily exceeds hours during peak periods in the morning. Currently, only U.S. Customs and Border Protection (CBP) provides border wait times at all ports of entry across the U.S.-Mexico border. The wait times are updated every hour. CBP collects the border wait times using visual identification of the queue length and survey of motorists waiting to cross the border. However, it is widely acknowledged that the wait times provided by CBP are underestimated, are not reflective of the real crossing time, and are not scientifically collected. For motorists, CBP is the only source of information regarding wait times at the border. Providing reliable and accurate border crossing times for passenger vehicles and trucks is crucial to efficient operation at the border and to motorists. Hence, in addition to CBP, local and regional transportation agencies are interested in deploying intelligent transportation system (ITS) solutions at the border to provide such information to motorists planning to enter the United States. Texas Transportation Institute Page 2
CHAPTER 2: BLUETOOTH STUDY AT THE U.S.-MEXICO BORDER Bluetooth technology is widely used to establish point-to-point communication between electronic devices. Bluetooth technology has been used extensively at international ports of entry along the U.S.-Canada border to measure crossing times of passenger vehicles. However, this technology has not been used at the U.S.-Mexico border to measure crossing times of passenger and commercial vehicles. The Texas Transportation Institute (TTI) collaborated with Turnpike Global (TG) to perform a study to apply Bluetooth technology for measurement of border crossing times of passenger vehicles at international ports of entry in the El Paso region. This study is the first of its kind along the U.S.-Mexico border and is part of a broader research project being undertaken by TTI to apply innovative technologies to measure passenger and commercial vehicle crossing times along U.S.-Mexico border regions. The overall objective of the research project is to study the feasibility of using Bluetooth technology to measure border crossing times at the U.S.-Mexico border. The project will quantify the number of Bluetooth signals (identified by media access control [MAC] addresses) transmitted by mobile phones from drivers and passengers who are crossing the international border, and the ability of Bluetooth readers to measure the crossing time of passenger vehicles. Bluetooth readers used in the project read the identification of Bluetooth signals originating from Bluetooth-enabled devices. By matching the identification of signals at readers located on the Mexican and U.S. sides, the travel times of vehicles crossing the border can be estimated. Texas Transportation Institute Page 3
CHAPTER 3: BLUETOOTH TECHNOLOGY Bluetooth wireless technology is a short-range communications technology intended to replace the cables connecting portable and/or fixed devices while maintaining high levels of security. The key features of Bluetooth technology are robustness, low power, and low cost. The Bluetooth specification defines a uniform structure for a wide range of devices to connect and communicate with each other. Bluetooth technology has achieved global acceptance such that any Bluetooth-enabled device, almost anywhere in the world, can connect to other Bluetoothenabled devices in proximity. Bluetooth technology operates in the unlicensed industrial, scientific, and medical (ISM) band at 2.4 to 2.485 GHz, using a spread spectrum, frequency-hopping, full-duplex signal at a nominal rate of 1600 hops/second. The 2.4 GHz ISM band is available and unlicensed in most countries. Bluetooth technology s adaptive frequency hopping (AFH) capability was designed to reduce interference between wireless technologies sharing the 2.4 GHz spectrum. AFH works within the spectrum to take advantage of the available frequency. This is done by detecting other devices in the spectrum and avoiding the frequencies that are being used. This adaptive hopping allows for more efficient transmission within the spectrum, providing users with greater performance even if using other technologies along with Bluetooth. The signal hops among 79 frequencies at 1-MHz intervals to give a high degree of interference immunity. The operating range depends on the device class: Class 3 radios have a range of up to 1 meter, or 3 feet. Class 2 radios are most commonly found in mobile devices and have a range of 10 meters, or 33 feet. Class 1 radios are primarily for industrial use and have a range of 100 meters, or 300 feet. Bluetooth specifications are developed and licensed by the Bluetooth Special Interest Group (SIG). The Bluetooth SIG consists of companies in the areas of telecommunication, computing, networking, and consumer electronics. Bluetooth-enabled devices can communicate with other Bluetooth-enabled devices from a range of 1 meter to about 100 meters, depending on the class of radios attached to the device. The Bluetooth protocol uses an electronic identifier in each device called a media access control address. Bluetooth readers are able to search for nearby devices using a refresh rate defined by the software running inside the reader and can obtain the MAC addresses of Bluetooth-enabled devices along with a timestamp. Because each MAC address is unique, traditional matching algorithms analogous to those used for license plate, cellular, or toll tag tracking can be used to estimate travel time along a freeway or arterial. MAC addresses are not directly associated with any of the users personal information. The MAC address of mobile phones or other electronic devices is not linked to a specific person through any type of central database, thus minimizing privacy concerns. Additionally, users with privacy concerns can turn off the Bluetooth discovery mode in their device. Texas Transportation Institute Page 4
The Bluetooth reader software is rather basic: each reader constantly issues a Bluetooth discovery request. According to the standard Bluetooth protocol, a Bluetooth device set to discoverable mode must respond to the discovery request by transmitting its unique Bluetooth identifier (12 hex digits) and device class (6 hex digits). The reader constantly issues the same discovery request and records the presence of the various devices it encounters (along with the date and time of each distinct instance a device was discovered). Using this approach gives the additional benefit of not requiring any special software to run on passengers devices. The only requirement is that passengers set their devices Bluetooth adapter to discoverable mode. Texas Transportation Institute Page 5
CHAPTER 4: DATA COLLECTION Data collection was conducted at all three ports of entry (POEs) in the El Paso area, including Bridge of the America (BOTA), Ysleta, and Paso Del Norte. The location of the ports of entry is shown in Figure 1. The data collection started on July 29, 2008, and ended on August 20, 2008. At each port of entry, data were collected for five consecutive weekdays from 8 to 11 AM. At each POE, one Bluetooth reader was placed on each side of the U.S.-Mexico border and placed on top of vehicles as shown in Figures 2 and 3 to reduce line of sight and signal obstructions. In addition to the Bluetooth readers, the volume of passenger vehicles entering the United States was manually counted. Bridge of the Americas Paso Del Norte Ysleta Figure 1. Ports of Entry along the U.S.-Mexico Border. FIELD EQUIPMENT SETUP Potential locations of readers were determined based on conversations with U.S. and Mexican bridge operators and U.S. Customs and Border Protection. The locations of readers were also selected to capture the maximum number of vehicles entering and exiting the ports of entry, considering the fact that the omnidirectional range of readers was 300 feet. The typical locations of readers on the U.S. and Mexican sides of individual ports of entry are shown in Figure 2. The locations of readers at all three ports of entry are shown in Figure 3 through Figure 5. Texas Transportation Institute Page 6
Mexico United States Figure 2. Typical Location of Bluetooth Readers at a Port of Entry. Texas Transportation Institute Page 7
Location of Bluetooth Reader on the U.S. Side of the POE Location of Bluetooth Reader and Surveyor on the Mexican Side of the POE Figure 3. Location of Bluetooth Readers at U.S. and Mexican Sides of the Bridge of the Americas POE. Location of Bluetooth Reader and Surveyor on the U.S. Side of the POE Texas Transportation Institute Page 8
Location of Bluetooth Reader and Surveyor on the Mexican Side of the POE Figure 4. Location of Bluetooth Readers at U.S. and Mexican Sides of the Ysleta POE. Texas Transportation Institute Page 9
Location of Bluetooth Reader and Surveyor on the U.S. Side of the POE Location of Bluetooth Reader and Surveyor on the Mexican Side of the POE Figure 5. Location of Bluetooth Readers at U.S. and Mexican Sides of the Paso Del Norte POE. Texas Transportation Institute Page 10
A TTI student was deployed at each port of entry with a Bluetooth reader, which was mounted on the roof of a vehicle, as shown in Figure 6 and Figure 7. Figure 6. Installation of Bluetooth Reader on Top of a TTI Truck on the U.S. Side. Figure 7. Installation of Bluetooth Reader on Top of a TTI Vehicle on the Mexican Side. Texas Transportation Institute Page 11
In addition to the collection of Bluetooth signals, an intercept survey was performed (Figure 8). The survey included questionnaires related to the availability of Bluetooth-capable mobile devices among motorists entering the United States. The objective of the survey was to obtain a sample distribution of the number of mobile phones that may or may not be Bluetooth capable. Figure 8. Undergraduate Student Conducting Intercept Survey of Motorists. RESULTS The Bluetooth readers were successfully able to read MAC addresses of wireless devices on both sides of the border. The raw data collected by the readers consisted of a list of MAC addresses, the identification number of the reader, and the timestamp. The MAC addresses were matched between readers in the United States and Mexico. A summary of the total number of unique addresses and matching addresses read at each port of entry on different dates is provided in Table 1. Texas Transportation Institute Page 12
Table 1. Results from the Field Experiment, Including Number of Matching MAC Addresses. Port of Entry Date Total Passenger Vehicles Entering U.S. Number of Unique MAC Addresses (on Mexican Side) Number of Unique MAC Addresses (on U.S. Side) Zaragoza 07/29 1184 185 (16%) 77 4% 07/30 1219 157 (13%) 76 3% 07/31 1086 252 (23%) 78 2% 08/01 1154 213 (18%) 95 5% % of Total Entering Passenger Vehicles with Matching MAC Addresses Paso Del Norte 08/04 1064 408 (38%) 281 5% 08/05 955 211 (22%) 265 4% 08/06 940 365 (39%) 250 6% 08/07 914 301 (33%) 221 5% 08/08 917 386 (42%) 187 6% BOTA 08/11 2156 187 (9%) 199 3% 08/12 2354 210 (9%) 330 2% 08/13 2373 219 (9%) 177 2% 08/14 2628 285 (11%) 277 2% 08/15 2748 139 (5%) 285 1.5% The readers read significantly fewer MAC addresses on the U.S. side than on the Mexican side. Even though the exact reason for this disparity is not known, this could be attributed to the fact that vehicles are moving at stop-and-go conditions on the Mexican side and move much faster after exiting the inspection station on the U.S. side of a port of entry. On average, 5 percent of the total entering vehicles were matched during a 4-hour period of data collection, which translated to 15 matching addresses every hour. Texas Transportation Institute Page 13
Figure 9 through Figure 12 show measured crossing times of passenger vehicles compared with CBP-reported crossing times at all three ports of entry. 80 70 60 50 40 30 20 10 0 7:14 AM 7:20 AM 7:33 AM 7:46 AM 8:02 AM 8:08 AM 8:17 AM 8:21 AM 8:27 AM 8:34 AM 8:41 AM 8:48 AM 8:53 AM 8:58 AM 9:04 AM 9:08 AM 9:13 AM 9:18 AM 9:22 AM 9:26 AM 9:44 AM 10:02 AM 10:08 AM 10:12 AM 10:14 AM Measured Crossing Time CBP-Reported Waiting Time Figure 9. Comparison of Measured and CBP-Reported Crossing Times at Paso Del Norte POE. Texas Transportation Institute Page 14
60 50 40 30 20 10 0 7:45 AM 7:56 AM 8:02 AM 8:07 AM 8:11 AM 8:50 AM 8:55 AM 8:56 AM 9:06 AM 9:16 AM 9:42 AM 9:47 AM 10:04 AM 10:14 AM Measured Crossing Time CBP-Reported Wait Time Figure 10. Comparison of Measured and CBP-Reported Crossing Times at Ysleta POE. 100 90 80 70 60 50 40 30 20 10 0 7:12 AM 7:12 AM 7:32 AM 7:40 AM 7:41 AM 7:43 AM 7:45 AM 7:51 AM 8:09 AM 8:26 AM 8:34 AM 8:44 AM 8:50 AM 9:05 AM 9:21 AM 9:26 AM 9:28 AM 9:29 AM 9:43 AM 9:44 AM 9:45 AM 9:46 AM 9:56 AM Measured Crossing Time CBP-Reported Wait Time 10:07 AM 10:09 AM 10:18 AM 10:23 AM Figure 11. Comparison of Measured and CBP-Reported Crossing Times at BOTA POE. Texas Transportation Institute Page 15
Figure 12. Placement of Bluetooth Readers at BOTA POE. Texas Transportation Institute Page 16
Figure 13. Result of Moving the Bluetooth Reader Further Upstream on the Mexican Side. Texas Transportation Institute Page 17
Total Surveyed = 40 Have Mobile Phones = 78% Do not Have Mobile Phones = 22% Bluetooth Capable = 80% Bluetooth Not Capable = 10% Do not Know = 10% Bluetooth Disabled = 60% Bluetooth Enabled= 40% Figure 14. Intercept Survey of Motorists Entering the United States. Table 2. Results from the Intercept Survey. POE Survey Date Total Drivers Surveyed Mobile Phone Bluetooth Capable Yes No Yes No Do Not Know Bluetooth Enabled Enabled Disabled Market Penetration Ysleta 7/29/2008 40 31 9 25 3 3 10 16 25% Ysleta 7/30/2008 40 36 4 23 8 5 9 14 23% Ysleta 7/31/2008 50 44 6 29 14 1 12 17 24% Ysleta 8/1/2008 51 45 6 34 11 0 15 19 29% Paso Del Norte 8/4/2008 49 38 11 27 11 0 13 16 27% Paso Del Norte 8/5/2008 40 33 7 21 11 1 12 9 30% Paso Del Norte 8/6/2008 40 34 6 28 5 1 16 13 40% Paso Del Norte 8/7/2008 49 39 10 26 11 2 13 13 27% Paso Del Norte 8/8/2008 40 36 4 24 11 1 9 15 23% BOTA 8/11/2008 47 39 8 32 7 0 12 20 26% BOTA 8/12/2008 37 35 2 28 7 0 12 16 32% Note: At BOTA, the intercept survey was performed for only 2 days. Texas Transportation Institute Page 18
CHAPTER 5: CONCLUSION The experiment and the results confirmed that Bluetooth technology could be used to measure crossing times of passenger vehicles at ports of entry in the El Paso region. The study found more than adequate penetration rates of matching Bluetooth-enabled devices to estimate average crossing times of passenger vehicles. No private information was obtained during the experiment. However, before implementation of a full-fledged project, the public should be made aware of how Bluetooth technology works and how crossing times are estimated. Also, a feasibility study should be performed to determine if there are enough Bluetooth-enabled mobile devices among motorists crossing the border. With large numbers of Bluetooth readers placed at strategic locations along the roadways entering the ports of entry, crossing times can be estimate more accurately. However, the physical geometry and queues at POEs govern the number of Bluetooth readers that can be installed. Even though Bluetooth technology is cheaper than radio-frequency identification (RFID), it requires more complex algorithms to determine average crossing times. This is mostly because Bluetooth readers read signals from devices that may not represent motorists waiting to cross. The researchers recommend that a pilot project be conducted before full-fledged implementation to minimize problems that could arise. A controlled experiment such as this one still does not reveal many technical constraints, such as power supply requirements, establishing communication between devices and central systems, etc. Texas Transportation Institute Page 19
BIBLIOGRAPHY Bluetooth Basics. Online Bluetooth Special Interest Group. http://www.bluetooth.com/bluetooth/technology/basics.htm. Texas Transportation Institute Page 20