5.1 Overview

The two major goals of the UTT system pilot test were to test the UTT system's functional requirements and determine ways that the system could improve safety, security, and efficiency. This section provides information about the pilot test.

5.2 General Test Procedures

Prior to initiating the pilot test, the test participants received training on the UTT system functions and testing procedures. An Implementation Guide was used in addition to various training tools and documentation. The training defined the test participants' roles and responsibilities, which included how to use the trailer-tracking host application to configure the system, investigate and document alerts, and maintain logs.

During the pilot test, automated test data were collected and reviewed weekly to check system performance and usage. For example, the UTT terminal battery was monitored throughout the pilot test to determine its performance against the requirements, which included an assessment of battery life and the low battery indicator.

Telephone interviews and nine on-site visits occurred from October 2004 through January 2005. These site visits and interviews with pilot test participants were conducted to verify and collect functional data and other information concerning UTT system usage. The initial telephone interviews and on-site participant interviews documented the following elements:

• User experiences.
• Perceptions of technologies that included appropriateness, efficacy, satisfaction, and willingness to invest.
• Possible integration with processes and procedures to enhance safety, security, and operational efficiency.
• Suggestions and recommendations.

Interviews were conducted with participant management personnel, dispatchers, maintenance personnel, and clerical/office employees. The questions were asked in a Likert-scale format, and the scale included levels of agreement and disagreement with statements for a broad-based compilation of interview results. The interview guide is located in Appendix F.

5.3 Data Collection Methods

The following data collection sources were used for information about the UTT system:

• Automated System Logs. A tool collected the UTT system data from the host application database on a weekly basis from each of the three participants. The data was compiled and archived. See Appendix D for a sample of the collected data and description of the data.
• Staged Tests. Limited on-site stage testing was conducted on a monthly basis.
• Manual Logs. In each scenario, the dispatchers maintained a manual data collection log in bound serialized books as they performed their daily dispatch activities. At least one dispatcher at each participant's location completed information in the logs about trailers with and without the UTT system. See Appendix E for a sample of a daily manual log.
• On-Site Interviews. On-site interviews were conducted to obtain qualitative data about the UTT system. See Appendix F for a sample of the on-site test participant guide used to collect qualitative data.

5.4 System-Generated Status and Location Reports during Daily Operations

During the 3-month field testing, system-generated incidences were collected relating to changes in cargo status, door status, connection status, and trailer positions, depending on the scenario.

Appendix C presents a sample of the automated data format from the data logs of all the incidences recorded during the pilot test. The data set contained 13,470 data points representing discrete trailer events and status updates for all trailers involved in the pilot test. The data logs from the daily operations provided information about:

• Cargo and Door Sensors
• Location and Mapping of Trailers Functionality
• Risk Area Geo-fencing
• Battery Life Performance

5.4.1 Cargo and Door Sensors

The automated event logs were reviewed to determine the numbers of door and cargo events and if the door and cargo sensor incidences provided information about "no sensor" indications in the daily field test operations. Table 7 presents the total number of door and cargo status events documented in the automated logs in each scenario.

For the cargo sensor and door sensor test, a "no sensor" status indicated when the cargo or door sensor failed to report a status for a specific event. For scenarios 2 and 3, the alert test configuration was set to provide a message notification to a carrier only when a "door open" event corresponded with a cargo status of "not empty." If the number of alerts for the "not empty" cargo status events did not equal the number of "door open" alerts, it was assumed that the cargo and/or door sensors failed to report a status. The results of the numbers of door and cargo status events and information on the numbers of "no sensor" indications are summarized in Table 8.

5.4.2 Location and Mapping of Trailers Functionality

The system-generated data of the UTT system provided position reports of both tethered and untethered trailers. The UTT system was configured to provide a position report with every event. Table 9 presents the number of trailer location reports when trailers were both tethered and untethered by scenario.

For all three scenarios, the test configuration was set at a maximum position reporting interval of 24 hours where the 24-hour cycle was configured to begin from the previous event or status report. It was assumed that exceeding a 24 hours and 30 minutes time interval would indicate a window of lost visibility, which could be the result of a lack in communications coverage [5], loss of power to the UTT system, damage to the UTT system, or a needed system configuration adjustment. First, the percentage of trailer location reports for both tethered and untethered trailers that fell within or near the test-configured maximum reporting interval of 24 hours and 30 minutes was determined. These results are summarized in Table 10.

Second, the position reports for only untethered trailers were reviewed when a trailer was disconnected from a tractor, as discerned in the data. The number of untethered occurrences was estimated to represent approximately 94 percent of all untethered trailer events during the field test. These results are summarized in Table 11.

For all three scenarios, 121 location reports for tethered and untethered trailers were received after 24 hours and 30 minutes, and 81 reports were received after 25 hours. For Scenarios 1, 2, and 3, the number of reports that exceeded 25 hours was 25, 14, and 42, respectively. Twelve location reports from different trailers were not received within 3 days. In general, reporting intervals exceeding 3 days occurred during the first 2 weeks of the carrier test period. This change in reporting intervals coincided with the brief period of time following the initiation of the field test when technical issues, such as faulty wiring of the fuse kits, broken cables on the tractors, loose wiring on the door sensor, and configuration issues regarding the battery pack were resolved.

For all three scenarios, 57 location reports for untethered trailers were received after 24 hours and 30 minutes. For the 37 reports that were received after 25 hours, all of the trailers were found in less than 2 days during the 2 weeks following the initiation of the test.

Trailer Movements Within United States and Into/Out of Mexico and Canada

Throughout the pilot test, 13,470 position reports were generated. Approximately 63 percent of these position reports showed no trailer movement from the previous report. By scenario, the percentages of position reports showing no trailer movement from the previous reports were 55, 58, and 73 percent for Scenarios 1, 2 and 3.

When the trailers were moving en-route, the average observed haul lengths were defined as trip miles loaded to next loaded cargo status for the three scenarios. The haul lengths were reported as 564 miles (standard deviation [s.d.] 431 miles), 789 miles (s.d. 327 miles), and 928 miles (s.d. 401 miles). Trailer movement locations are illustrated in Figures 19, 20, and 21.

The trailer position reports showed movement into and out of Mexico and Canada, as presented in Table 12.

In Scenario 3, seven individual trailers made seven separate trips to and from Canada near the Canadian border. Each trip was approximately 50 to 80 miles between participant locations in New York State and customer sites in Quebec. During these trips, 84 position reports were generated.

Trailer movements into Mexico and near the Mexican border were limited for Scenarios 1 and 3, with one and nine trailer position reports, respectively. Scenario 2 did show a number of movements into Mexico – 53 separate trips from drop lots in Texas to 47 separate locations in Mexico (all locations in northern Mexico near the U.S.-Mexican border). Of the 285 trailer position reports in Mexico, performance reporting intervals were approximately the same as observed within the United States. Two reports showed reporting intervals greater than 2 days, and less than 4 days.

Anecdotally, Scenario 2 participants believed that partner carriers were using their trailers in Mexico in an unauthorized manner, although they could not prove it. During the test, the Scenario 2 participants successfully used the UTT system to identify a trailer that was used to move other unauthorized goods within northern Mexico for 2 days prior to picking up the proper load.

5.4.3 Risk Area Geo-fencing

Risk area geo-fencing was used most frequently by participants in Scenarios 1 and 3. Geo-fences were created around a trailer at the carrier's terminal, customer site, or other location when it was disconnected and expected to remain untethered at one location for more than 24 hours. In Scenario 1, a trailer with an advertisement for hiring new drivers was placed alongside the freeway and geo-fenced. For over one month, no unauthorized geo-fence breaks occurred.

In Scenario 3, 7 trailers with UTT systems were stored at a DOD facility and geo-fenced. Scenario 3 involved a carrier of munitions and other sensitive loads for the DOD. At the beginning of the pilot test, DOD revised the Hazards of Electromagnetic Radiation to Ordnance (HERO) standards for UTT devices. Due to an advisory released by DOD regarding the use of UTT devices, the Scenario 3 participant was unable to fully utilize these trailers with UTT systems to haul DOD munitions shipments; therefore, the trailers were geofenced at the bases. No unauthorized geo-fence breaks occurred on the DOD facility.

5.4.4 Battery Life Performance

When the trailers were untethered, the UTT system's battery was the sole power source for the sensors, positioning functionality, and transmitting telemetry. When the trailer was tethered to a tractor, the battery was recharged from the tractor's power supply via the 7-pin connector interface on the connecting wiring harness.

During daily operations, a "charge battery" alert was generated four times on four separate trailers: two each in Scenarios 1 and 3. Due to the short time frame of the pilot test, these alerts were infrequent. Three alert notifications were generated at the end of the pilot test data collection period. In Scenario 1, the UTT systems generating the "charge battery" alerts were configured for 2-hour wakeup intervals, and the batteries lost their charges in 32 and 56 days. In Scenario 3, which included door status checks, the UTT systems were configured to wakeup every 6 hours, and their batteries lost their charges in 71 and 90 days. The more frequent wakeup intervals drained the battery at least two to three times as fast as the less frequent wakeup intervals.

In addition, a bench test was conducted to determine the expected life of the battery charge while the trailer is untethered and configured to report positions at high frequency rates (2-minute intervals). This test showed that the battery charge would last approximately 12.5 hours, starting with a fully charged battery.

5.5 On-Site Staged Tests

5.5.1 On-Site Staged Testing Process

During the pilot test, nine on-site staged tests were conducted to collect more information about the UTT system. The steps listed below were followed for the on-site staged testing process.

1. Located a properly equipped tractor and trailer for testing at participant facility.
2. Configured testing parameters and forwarded them to the trailer, such as setting the positioning interval at 2 minutes; selecting the geo-fence; entering email addresses to receive alert notifications; and inputting time settings for door and cargo sensors. (To facilitate the staged testing, the test configurations were set at a shorter duration than those set for the field test.)
3. Monitored testing progress for alert notifications, functional events, and trailer positioning via a Web-based user interface on a laptop computer.
4. Tested UTT system functionalities.

The staged technology events and alert notifications were displayed to participating dispatchers on the Web-based user interface screen and sent to the participant's designated email addresses to simulate the events and alert notifications that could normally occur during daily operations. Alert notifications were forwarded to the carrier participants' and test team members' email addresses, which displayed the test trailer ID and the time of the event. The following conditions were set for the staged testing:

• Trailer positioning was set for every 2 minutes.
• Cargo sensors were configured to check the cargo status every 5 to 20 minutes.
  • To simulate a loaded or empty trailer, a piece of paper or cardboard was placed in front of the cargo sensor to create a "cargo loaded" event and removed for a "not empty" event when necessary. The cardboard obstruction was used as an alternative for sensing cargo directly in front of the sensor due to the time constraints of actually loading and unloading cargo.
• Door open events were configured to generate an alert notification immediately upon the trailer door being opened with cargo present in the trailer.
• Trailer connections and disconnections were configured to generate near real-time alert notifications after the tethering and untethering of the trailer.
  • During this process, the connections indicating the trailer serial number and the tractor ID or mobile terminal number were checked in the UTT system Web-based user interface. Also, a trailer was disconnected from a tractor to check that the serial number did not change whether the trailer was tethered or untethered.
• Geo-fence violations were configured to be triggered in near-real time upon the violation of an inclusion zone or inside of an exclusion zone.
  • The driver was directed to pull a trailer from the geo-fenced terminal area.
  • When the driver left the geo-fenced area, an exit geo-fence alert notification was immediately generated and displayed on the Web-based user application.
  • As the trailer traveled down the road, its position was monitored at 2-minute position updates.
  • When the trailer was moved outside the geo-fenced area, the driver disconnected the 7-pin connection between the tractor and trailer to simulate a trailer drop in an "unknown location" (a location not designated by the carrier as a valid drop location). The disconnection generated an alert notification via the Web-based user interface and designated email addresses.

Figure 22 shows an example of the alert screen that would display events and alerts to a dispatcher. Each alert is displayed and assigned to a particular trailer. By clicking on the event, the dispatcher can see the details of a specific security event including the location of an alert.

5.5.2 Staged Testing Results

The staged tests procedures and results are described below for the previously listed functions.

Scenario 1

Testing was conducted at the participant's facilities on October 28, 2004; December 2, 2004; and January 13, 2005:

• Cargo Sensor (remote sensing of empty or loaded trailers)
  • Cargo loaded and unloaded functionalities were both tested successfully.
• Connections and Disconnections with Alerts
  • Connections and disconnections were each tested with alert notifications generated and displayed on the alerts Web page display as well as to the participant email accounts.
• Trailer ID
  • The trailer ID and tractor ID were successfully displayed on the TrailerTRACS program. A specific truck ID was not tied to a trailer ID; therefore, the mobile communications terminal (MCT) number was substituted instead.
• Location and Mapping of Trailers
  • The test was set with positioning updates at 2 minutes. The trailer was tracked and mapped following the geo-fence exit event.
• Unscheduled Movement Notification (geo-fencing) with Alerts including the function of detecting unauthorized trailer drops
  • A geo-fence was set up around the motor carrier's terminal and the driver pulled away with the trailer. The geo-fence exit alert was received. A drop at an unknown location alert notification was generated following the geo-fence exit event by having a disconnect event occur at a non-land marked location. During one test, the driver returned to the lot several minutes later to cause the geo-fence entry event. During another test, there was no geo-fence break as the tractor broke down with a blown turbo charger right on the edge of the geo-fence perimeter.

Scenario 2

Testing was conducted at the participant's facilities on October 26, 2004; November 30, 2004; and January 11, 2005.

• Door Sensor
  • During the first site visit, the test trailer's door sensor was damaged and continuously indicated an open condition. Another trailer's door sensor also continuously indicated an open door status. One driver stated that that the trailer doors are frequently hit when trucks back up to loading docks. During the remaining site visits, the test trailers had functioning door sensors. In an improved post-test redesign, the door sensors will be recessed into the trailer door.
• Cargo Sensor (remote sensing of empty or loaded trailers)
  • Cargo loaded and unloaded functionalities were both tested successfully. During one site visit, the trailer was loaded with empty chemical barrels, and the cargo sensor correctly relayed the "not empty" message.
• Connections and Disconnections with Alerts
  • Connections and disconnections were tested with alert notifications displayed on the alerts Web page display as well as to the participant email accounts. During one site visit, there was no power on the truck's pin 7 of 7-way terminal, which did not allow visibility regarding the connections and disconnections. The driver brought the truck back to the yard for troubleshooting and repair. In disassembling the 7-way terminal, the wire feeding the UTT system had been severed. The damage may have been caused by a previous dent to the front of the trailer in the area between the 7-way connector interface and the UTT terminal. During another site visit, the truck was wired with the fuse kit that enabled this functionality. This truck was set up for manual connection and disconnection, but not with auto connection. Unsuccessful attempts were made to get the auto connection enabled from the hub. Instead, successful manual connections and disconnections were conducted via the tractor keypad.
• Trailer ID
  • The trailer ID was successfully shown with the corresponding tractor ID on the TrailerTRACS program. The truck was wired with the fuse kit that enabled this functionality.
• Location and Mapping of Trailers
  • The test was set with positioning updates at 2 minutes. The trailer was tracked and mapped following the geo-fence exit event.
• Unscheduled Movement Notification (geo-fencing) with Alerts, including the function of detecting unauthorized trailer drops
  • A geo-fence was set up around the motor carrier's terminal, and the driver pulled away with the trailer. The geo-fence exit and entry alerts were received and displayed as expected. A drop at an unknown location alert notification was generated following the geo-fence exit event with a disconnection event occur at a non-land marked location.

Scenario 3

Testing was conducted at the participant's facilities on November 2, 2004; December 7, 2004; and January 18, 2005.

• Door Sensor
  • In two of the three tests, the test participants did not receive the email alerts due to Anti-SPAM software on their email systems, which deleted the alerts. Following the staged tests, the software problem was remedied by changing the email header. In one test, the "door open" alert was not generated, since the cargo sensor incorrectly relayed an "empty" event due to a testing procedure failure. During the remaining site visits, the test trailer had a functioning door sensor.
• Cargo Sensor (remote sensing of empty or loaded trailers)
  • Cargo loaded and unloaded functionalities were both tested successfully. During one test, the sensor did not indicate the cargo loaded status. This failure may have been caused by the size of the piece of cardboard used in the test. On-site diagnostics verified that the sensor was working properly. The sensor was finally tested with a person standing in front of the sensor, which resulted in the system indicating a cargo loaded status.
• Connections and Disconnections with Alerts
  • Connections and disconnections were each tested with alert notifications generated and displayed on the alerts Web page display as well as to the participant email accounts.
• Near Real-Time Trailer ID
  • The trailer ID was successfully shown with the corresponding tractor ID on the TrailerTRACS program.
• Location and Mapping of Trailers
  • The test was set with positioning updates at 2 minutes. The trailer was tracked and mapped as it traveled eastbound following the geo-fence exit event. The trailer turned around several minutes later, returning prior to the geo-fence entry event.
• Unscheduled Movement Notification (geo-fencing) with alerts, including the function of detecting unauthorized trailer drops
  • During the tests, a geo-fence was set up around the motor carrier's terminal. The driver pulled away from lot with the trailer to cause the geo-fence exit event and returned to the lot several minutes later to cause the geo-fence entry event.

5.5.3 Manual Logs

The three participating motor carriers in the pilot test utilized the UTT system TrailerTRACS Web application to identify the location and cargo status of their 25 respective trailers with the UTT system. Simultaneously, the motor carriers also continued to use their normal methods of phone and manual searches for trailers without the UTT system. During the pilot test, the dispatchers or route planners at each motor carrier were asked to complete logs for trailers with and without UTT systems. A sample log with carrier entries is presented in Appendix D along with the instructions to the dispatchers.

Due to the different dispatchers in each scenario, the logs provided varying amounts of information. However, the logs showed that trailers without the UTT system were not found where they were expected to be 41, 11, and 38 percent of the time (percent of manual log entries showing "not where expected" versus all log entries for the particular participant carrier) for Scenarios 1, 2 and 3, respectively.

After using the UTT System, the participants in all three scenarios stated that it allowed for better and easier tracking of loads, drivers, and vehicles. Trailers can be located by querying the UTT system host application for a specific trailer, trailers near a specified location or landmark, trailers with a certain cargo status of "empty" or "not empty," or a listing of all assigned trailers.

Previous | Next