The new highway noise database currently being collected by federal and state transportation agencies requires data from more than 1,000 individual vehicle pass-bys. However, the standard data collection method–tape-recording in the field and analysis in the laboratory–prohibits the collection of a large amount of information in a reasonably short time. Therefore, a new method of data collection has been proposed. This method would acquire, analyze, present, and store data automatically on site. Robert Coulson discussed its benefits and limitations in “Method for Measuring Vehicle Noise Source Heights and Subsource Spectra” (Transportation Research Record 1559). The new method is quick, easy to set up and operate, and can be entirely powered by a small generator. The system is limited at certain frequencies, but this and other limitations are expected to improve with further study.

SYSTEM DESCRIPTION / OPERATION

Figure 1 illustrates the measurement instrumentation for this new method of measuring vehicle noise, and Figures 2 and 3 show the equipment arrangement.

FIGURE 3
Plan view of test site setup

The axes of eight microphones are arranged vertically between ground level and 1 meter (39 inches) and are interfaced with the data acquisition and analysis system. The apparatus is placed 7.5 meters from the center of the traffic lane being measured, with the computer equipment and operator located in a support van downstream. Data acquisition begins with a signal triggered by a vehicle’s front tires breaking the first infrared light beam, “which crossed the highway 2.5 m[eters] before the array’s axis to the closest point of approach (CPA).” Data acquisition ends when the same vehicle’s front tires trigger a second signal as they cross a second light beam located 2.5 meters beyond the CPA. Software written in National Instruments LABVIEW 3.1.1 Graphical Programming for Instrumentation platform controls the data acquisition, analysis, presentation, and storage, and also processes the signals. Calibration occurs at the beginning and end of each measurement session.

Set up and operation require only two people, and the system requires minimal operator input. After running a microphone calibration program, “the operator is then responsible only for arming the trigger at the approach of a possible good vehicle event.” After acquiring microphone signals as the vehicle passes through the triggers, the program takes about 30 seconds to compute the source heights and spectra, and the operator selects the vehicle classification for that pass-by. The program computes the subsource spectra and displays all data for that vehicle. The program then prompts the operator to indicate if the result should be stored to file. Finally, the program returns to the “arm trigger” window in readiness for the next vehicle. Data collection, reduction, and storage take less than one minute, which allows for “almost real-time, on-site evaluation of the data.”

SYSTEM LIMITATIONS

Despite its speed and ease of operation, the system does have limitations. Its primary drawback involves the signal-to-noise ratio. Noise in the microphone signals can introduce errors and limit the accurate region for each pair of microphones. As a result, “the source heights for frequencies below 500 Hertz (Hz) cannot be accurately measured.” In addition, problems have occurred above about 2,500 Hz. This is probably the result of “turbulence driven by the vehicle pass-by or . . . dropouts in the signal-to-noise ratio caused by destructive interference between the direct and reflected noise paths.” Although “the system is designed to work at these frequencies and is performing well for a known stationary noise source, it is apparent that something is producing a significant increase in the standard deviation of the data at these frequencies.”

In addition, few sites allow equipment to be placed the required 7.5 meters from traffic, and moving the apparatus farther from the road would decrease the signal-to-noise ratio. Furthermore, the surface between the vehicle and microphones must be hard, and the size of the road shoulder is limited. Sites with low background noise and little traffic provide the best situations, since the system functions optimally when signal-to-noise levels are high. However, the remote areas where these conditions are usually found are not conducive to collecting low-speed (less than 64.38 kilometers per hour or 40 miles per hour) data, especially for heavier vehicles.

TEST RESULTS

In excess of 1,000 individual vehicle pass-bys were recorded at about a dozen different sites in Florida. Vehicles were classified in the following groups: small vehicles (automobiles), medium trucks, heavy trucks, motorcycles, and buses. The site types included level asphalt and concrete combined, level asphalt, level concrete, and graded asphalt. Results from these sites showed “that at lower frequencies there is little difference in source heights among different vehicle types.” Source heights at about 1,000 Hz are relatively low because of the effect of tire noise. When the system approaches its low-frequency limit, the measured source height increases, “resulting in an increasing overestimate of the vehicle’s actual source height.” Aerodynamic noise seems to dominate at higher frequencies, where “the differences in source heights among different vehicle types are more pronounced.”

CONCLUSIONS