英语论文网

sified as “good,” “intermediate,” or “inappropriate.” The good sites were the ones
where data collection could be carried out safely, and the inappropriate sites were the sites where
data collection was not possible because of safety concerns. For some sites, although the safety
criteria were met, sufficient volumes were not observed during the visit. Since there was a
possibility of volumes being higher during peak hours, these sites were classified as
intermediate.
Data Collection at Signalized Intersections
The STV was used to collect data on traffic volume, queue length, speed and headways at
signalized intersections. The STV is a state-of–the-art van equipped with a computer and two
cameras on the mast that can capture the traffic data. The Autoscope program was installed in
the computer of the STV. Autoscope is a video detection package, which collects traffic
volumes, speeds, time headways, etc. The mast of the STV measures 45 feet when fully raised
and gives a bird’s-eye view of the intersection. This van was used because it could capture the
full view of the intersection including (in some cases) the signal heads. Data were collected for
approximately 2 hours at each intersection during either the morning or evening peak hours.
The Autoscope program was used to set up speed and count detectors. The count
detector did not have any length in the direction of the traffic flow and counted only the number
of vehicles that passed over it. It recorded the time at which the vehicle was activated and
deactivated the detector. These times were recorded to within a 1-second accuracy. There was
also a global positioning system (GPS) in the STV that recorded the times of each detector
activation and de-activation within a 1-millisecond accuracy. This was used to calculate the
headways in the opposing flow. The cameras of the STV covered the subject link approach so
that the end of the queue was visible (see Figure 1). From these tapes, the queue length (number
of cars) at the end of each red phase was calculated in the laboratory. This was used during the
validation of the simulation program. The approximate location of a typical count detector is
also shown on the subject link and opposing link of Figure 1.
Actuated signalized intersections do not maintain a constant cycle length since the phase
lengths change depending on arrival patterns. Furthermore, the camera on the STV did not
always capture the signal heads. Therefore a computer program was written to record signaltiming
changes. Whenever the user pressed a character in the keyboard, the program recorded the
character and time. This program was installed on the laptop that was carried to the field during
5
Figure 1. Layout of Data Collection Scheme
the data collection. Whenever a phase change occurred, a key (“r” for red, “g” for green, and “y”
for yellow) was pressed. The difference of the times between consecutive values in the text file
gives the time for which the signal showed green, yellow, or red. The changes in signal phase
times observed in the field were obtained using these procedures. All the signalized sites
recommended by the engineers were fully actuated or semi-actuated; hence the cycle lengths as
well as the green times were not constant.
6
Data Collection at Unsignalized