Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 43
43 a robust manner. They work independently to interpret events heavy commercial trucks or light cars) and study goals. Details and record additional variables into the database. Examples of of the trigger threshold values for VTTI studies are listed in variables obtained through data reduction include weather, Tables 4.3, 4.6, and 4.9 in Chapter 4. traffic conditions, and roadway type. Input from reduc- The second step is to check the validity of the potential tionists should be monitored and supervised by a reduction event epochs generated in step 1 through visual inspection. manager, and measures of inter-rater and intra-rater relia- As described in Figure 4.7 (1), invalid events and valid events bility should be used to ensure that quality control is main- (including conflicts and nonconflicts) are differentiated. tained. The contracting institution should have experience Invalid events are false alarms in which sensor readings are with data reduction, as well as facilities within which data spurious because of a transient spike or other anomaly. Valid reduction can be performed. events are events in which recorded dynamic-motion values The reduced data provide a rich source of information to occur and are verifiable in the video and other sensor data relate driver behavior to surrounding roadway, weather, and from the event. Valid events are further parsed as conflicts traffic conditions. Driver status information (e.g., secondary and nonconflicts. Conflicts are occasions in which a crash, tasks, driver distraction, driver apparent impairment, and near crash, incident, or conflict happens. Nonconflicts are driving habits) is available from the video data. Vehicle status occasions in which the trigger values satisfy the thresholds but information (e.g., vehicle speed, acceleration and decelera- the driver makes a maneuver and is in complete control of the tion, and lane orientation) is available as part of the in-vehicle vehicle. Examples can be a hard brake without any obvious kinematic data. A proper linkage of driver and vehicle vari- reasons or a high swerve value from a lane change. In most ables to external variables (traffic, weather, work zone, and cases, nonconflicts reflect aggressive driving habits or styles. accident data) through location and time information will In step 2, valid events are labeled for the next step. enable an in-depth study of the relationship between driver The final step is to apply the data dictionary to the vali- behavior and travel time reliability. dated events. In this step, a data dictionary created by the researcher in advance is used to extract useful information from the raw data. Data reductionists code results while they General Guidelines for watch the video by choosing correct answers from pull-down Video Data Reduction menus for each research question in the data dictionary. Vari- According to the general guidelines discussed earlier, the ables included in the data dictionary should reflect the inter- most challenging effort related to data reduction involves ests of investigators and the research objectives. This is the video data. Having conducted multiple naturalistic driving step in which the identification of driver behavior that affects studies involving video data, VTTI has developed a mature congestion is extracted. Reductionists can plot real-time procedure to organize, reduce, and analyze large-scale natu- charts of any selected variable in the DART--such as range, ralistic driving data. According to different research objec- velocity, and acceleration--simultaneously with the play of tives, the threshold values selected may vary from one data the associated video clips. When studying the relationship reduction to another, but the basic principle is consistent and between driver behavior and nonrecurring congestion, the the procedure is uniform. The step-by-step description of LOSs and traffic density before and after the event, as well as video data reduction in each project has been detailed indi- the contributing factors, are important. The LOS or traffic vidually in Chapter 4. Following is a summary of the general density before and after can be used in a simulation envi- guidelines for video data reduction. ronment or in statistical modeling (details of travel time There are three primary steps in video data reduction. The modeling are included in the previously submitted Task 5 first is to scan the raw data to identify potential safety events. Report) so that the events that generate congestion can be iden- VTTI developed its own software (DART) that reads and cal- tified. Contributing factors, especially those that are driver- culates numerical data to detect if events of interest have and vehicle-related, are judged by reductionists as correctable occurred. Therefore, the first step in data reduction is to run and avoidable or otherwise. For example, data reductionists the event trigger program. The variables examined include decide if a crash or near crash was caused by one of the fol- velocity, acceleration (longitudinal or lateral), swerve, TTC lowing reasons: (1) the driver was conducting a secondary task, (calculated from range and range rate between the subject such as talking on a cell phone, dining, adjusting the radio, or vehicle and other vehicles detected by radar), and yaw rate. talking to passengers; (2) the driver was inattentive to the road, The software reads through the raw numeric data and extracts either distracted by something inside or outside the vehicle or an epoch of data when it detects a parameter exceeding the simply daydreaming; (3) the driver was driving under the trigger threshold set by researchers in advance. Different influence (such as drugs, alcohol, or medicine); (4) the driver lengths of the epoch and threshold values can be adopted made a wrong decision; or (5) the driver has questionable according to different attributes of the subject vehicle (e.g., driving habits or driving skills. Simultaneously, other possible
OCR for page 43
44 Figure 5.1. Quality assurance and quality control flowchart at VTTI.