A Comprehensive Comparative Analysis of the Basic Theory of the Short Term Bus Passenger Flow Prediction

In order to meet the real-time public travel demands, the bus operators need to adjust the timetables in time. Therefore, it is necessary to predict the variations of the short-term passenger flow. Under the help of the advanced public transportation systems, a large amount of real-time data about passenger flow is collected from the automatic passenger counters, automatic fare collection systems, etc. Using these data, different kinds of methods are proposed to predict future variations of the short-term bus passenger flow. Based on the properties and background knowledge, these methods are classified into three categories: linear, nonlinear and combined methods. Their performances are evaluated in detail in the major aspects of the prediction accuracy, the complexity of training data structure and modeling process. For comparison, some long-term prediction methods are also analyzed simply. At last, it points that, with the help of automatic technology, a large amount of data about passenger flow will be collected, and using the big data technology to speed up the data preprocessing and modeling process may be one of the directions worthy of study in the future.

Validation and Sampling of Automatic Rail Passenger Counters for National Transit Database and Internal Reporting at TriMet

This paper presents findings from an evaluation of the accuracy of automatic passenger counters (APCs) on the Tri-County Metropolitan Transportation District of Oregon's (TriMet's) light rail vehicles. APC boarding and alighting counts were compared with manually recorded counts. Overall, the APCs tended to undercount boardings and overcount alightings. Thus, correction factors will be needed for use of APC-recovered data for National Transit Database (NTD) and internal reporting. The paper also describes the sampling procedures employed by TriMet for NTD and internal reporting, in which sampled trips are linked to archived APC data records.

Making Automatic Passenger Counts Mainstream

Although automatic passenger counters (APCs) have been used for many years, significant obstacles have hindered their becoming a mainstream source of data for monitoring ridership and peak load, estimating passenger miles, and other measures of passenger use important for transit management. The key to APC usefulness is the automatic, routine conversion of the APC data stream into a database of accurate counts. On the basis of case studies of transit agencies, five issues important to achieving this goal are analyzed: data structures, data accuracy, accuracy need and sampling requirements, controlling drift, and balancing algorithms. Balancing algorithms deal with routes with loop ends, negative loads, and rounding. Sampling and accuracy requirements related to passenger miles estimates for National Transit Database (NTD) reporting are also analyzed. The analysis shows that, for most agencies, NTD precision requirements can be met with a small level of fleet penetration, provided that measurement, screening, parsing, and balancing methods keep bias in load measurement below 8%.

Generating Transit Performance Measures with Archived Data

Measuring the performance of a transit system is the first step toward efficient and proactive management. Since 1990, the use of performance measures for transportation planning and operations has gained a great deal of attention, particularly as transportation agencies are required to provide service with diminishing resources. In the past, it was very difficult and costly to collect comprehensive performance data. Thus, until recently, the transit industry has relied on limited, general, and aggregate measures for reporting performance to external funding and regulatory agencies. In Portland, Oregon, the local transit provider (TriMet) has developed a bus dispatch system (BDS) consisting of automatic vehicle location, communications, automatic passenger counters, and a central dispatch center. Most significantly, TriMet had the foresight to develop a system to archive all of its stop-level data, which are then available for conversion to performance indicators. It is demonstrated that there are powerful ways in which the data collected by the BDS can be converted into potentially valuable Transit Performance Measures (TPMs). These TPMs have been proposed in the past but were not implemented because of data limitations. It is envisioned that systematic use of TPMs can assist a transit agency in improving the quality and reliability of its service, leading to improvements for customers and operators alike.

Estimating Bus Boarding Matrix Using Boarding Counts in Individual Vehicles

Automatic passenger counters that perform continuous boarding and alighting counts for transit vehicles have become widely available. They provide valuable information on passenger travel patterns. The BUS-OD model, which uses the boarding counts to refine the information, is described. The BUS-OD model estimates boarding trip tables, giving the number of trips between all station pairs along a bus line. A technique is devised to use the information that is embedded in counts for individual vehicles, combined with other information in the table’s structure. The problem is formulated as a nonlinear optimization problem. The model was tested thoroughly and applied successfully to full-scale problems. It is now ready for use in operational applications. The model and its behavior are described.