Data Collection

Subasish Das
UL Lafayette

http://subasish.github.io

• PhD candidate in Systems Engineering [expected Summer 2015]
• Research areas include:
  • traffic safety analysis
  • machine learning and knowledge discovery
  • data visualization

• Background
• Content
• Speed studies
  • Sight distance studies
  • Volume studies
  • Travel time and delay studies
• GPS data collection

• Data collection is an important part of transportation engineering. System performance evaluation requires data so does transportation planning.
• The development of Intelligent Transportation Systems (ITS) requires high quality traffic information in real-time. For several years, under growing pressure for improving traffic management, collecting traffic data methods have been evolving considerably and the access to real-time traffic information is becoming routine worldwide

Traffic studies may be grouped into three main categories:

• Inventories,
• Administrative studies, and
• Dynamic studies.

Provides a list or graphic display of existing information, such as:

• street widths,
• parking spaces,
• transit routes,
• traffic regulations.

• uses existing engineering records, available in government agencies and departments.
• includes the results of surveys, which may involve:
  • field measurements and/or
  • aerial photography.

• Examples of Traditional Methods
  • License plate
  • “in-situ”
  • Floating cars
• Examples of New Technology
  • Camera
  • GPS
  • Cell phone

• Speed Limit
• Establish Speed Trends
• Design Impacts (Safe Sight Distance)
• Traffic Control Placement and Timing
• Speed Related Collision Study
• Investigate Speed Complaints

• Collect during time period representative of desired outcome
• Usually cannot collect all vehicles
• Random sample

Systematic Errors and Solutions

• Error – looking for “fastest” vehicle
  • Solution – Sample every nth vehicle
• Error – too many heavy vehicle measurements
  • Solution – same as above – sample every nth vehicle
• Error – Inclusion of vehicle following platoon leader
  • Solution – Don’t include vehicles following too closely (200’ if < 40 mph, and 350’ otherwise)

• Appropriate study lengths
• Proper location and layout
• Record observations and calculate vehicle speeds
• Generate frequency distribution table and determine speed percentiles

• Spot speed studies are conducted to estimate the distribution of speeds of vehicles in a stream of traffic at a particular location on a highway.
• Carried out by recording the speeds of a sample of vehicles at a specified location.

• Represent different traffic conditions on a highway for basic data collection.
• Mid-blocks of urban highways and straight, level sections of rural highways for speed trend analyses.
• Any location may be used for solution of a specific traffic engineering problem.

Depends on the purpose of the study.

• Recommended when traffic is free-flowing,
• During off-peak hours.
• Typically:
  • Duration is at least 1 hour and
  • Sample size is at least 30 vehicles.

The larger the sample size, will give an estimated mean within acceptable error limits.

• Average Speed
• Median Speed
• Modal Speed
• The ith-percentile Spot Speed
• Pace
• Standard Deviation of Speeds

• Manual and automatic
• Manual method is seldom used
• Automatic devices
  • Road detectors,
  • Radar-based,
  • Principles of electronics.

Pneumatic road tubes & induction loops collect data on speeds & volume at the same time

Advantage

• Human errors are considerably reduced

Disadvantages

• Expensive
• May affect driver behavior,

• Laid across the lane in which data are to be collected.
• When moving vehicle passes over, an air impulse is transmitted to the counter.
• Two tubes are placed across the lane, 2 m apart.
• An impulse is recorded when the front wheels of a moving vehicle pass over the first tube;
• A second impulse is recorded when the front wheels pass over the second tube. The time elapsed between the two impulses and the distance between the tubes are used to compute the speed of the vehicle.

• Collect during time period representative of desired outcome
• Usually cannot collect all vehicles
• Random sample

• A rectangular wire loop buried under the roadway surface.
• Operates on the principle that a disturbance in the electrical field is created when a motor vehicle passes across it.

• Radar-Based Traffic Sensors
• Electronic-Principle Detectors
  • Traffic characteristics, such as speed, volume, queues, and headways are computed.
  • Using video image processing

• Statistical methods
• Analyzing data
• Frequency histogram
• Cumulative frequency distribution curve

• Generally assume spot speeds are normally distributed
• Assume sufficient samples are collected
• Should be able to calculate mean, standard deviation, etc. based on assumption of normality.

• Notes are posted in Moodle.
• Skim through the notes before next class.
• Will do some problems.

• Become popular very fast.
• rpubs1
• Shiny apps

require(maps)
data(canada.cities)
head(canada.cities, 2)

           name country.etc    pop   lat    long capital
1 Abbotsford BC          BC 157795 49.06 -122.30       0
2      Acton ON          ON   8308 43.63  -80.03       0