Language
Country/Area
No result found
Surprising Findings from Chicago Traffic Study: How Accurate Are Early Traffic Pattern Analyses?
In traffic planning and prediction models, mode choice analysis is the third step of the four-step traffic prediction model, the previous step is trip allocation and the next step is path assignment. Using the origin-destination map provided by trip allocation, mode choice analysis can determine the probability that a traveler will use a certain mode of transportation. These probabilities are called mode shares and can be used to estimate the number of trips made using each feasible mode.
Historical BackgroundEarly transportation planning models developed by the Chicago Area Transportation Study (CATS) focused on public transit, primarily to understand how many trips would continue to be made using that mode of transport. CATS divides public transport trips into two categories: to the Central Business District (CBD) and other trips. CBD trips rely mainly on the subway, express buses and commuter trains, while other trips mostly use the local bus system. As car ownership and use increase, the use of public buses decreases. Analyze CBD travel using historical data and CBD land use forecasts.
These analytical methods have been used extensively in many studies, for example the London study adopted a similar procedure to explore the association between income and mode choice by differentiating trips into inner and outer city trips.
Diversion Curve Technology
CATS uses diversion curve technology, which was originally developed to study how car traffic is transferred from streets and arterial roads to proposed expressways, for some of its missions. This technology can quantify how much car traffic can be attracted by building detour roads around cities. The split curve analysis of mode choice is usually performed by forming a ratio that expresses the travel time and other factors of different transportation modes.
These diversion curves are based on empirical observations, increase in accuracy as data improves, and can be applied to the analysis of traffic patterns in many markets.
Non-aggregate travel demand model
The introduction of travel demand theory has made non-aggregate travel demand models gradually become mainstream. These models are based on Stan Warner's research in 1962, which established the analysis of choice behavior under individual observation through biological and psychological models. Such models usually combine psychological concepts of consumer behavior and choice behavior and generate a set of parameters that describe the choice behavior of the population as a whole.
In research at the University of California, Berkeley and the Massachusetts Institute of Technology, researchers developed a variety of models known as choice models, which are widely praised because they help compare more than two options.
Psychological roots
Early psychological research often tested the choice of physical objects through experiments, and found that the greater the difference in gravity of the objects, the higher the probability of correct selection. This has led to the gradual attention paid to the application of psychology in traffic pattern analysis. The "perceived weight" model proposed by Louis Leon Thurstone in the 1920s further introduced this concept into the mathematical description of transportation choices.
Econometric Model
Economists typically focus on utility rather than specific gravity, which provides an alternative perspective for predictive analysis of traffic patterns. Starting from the diversion model, the econometric approach allows us to take into account the characteristics of each choice and incorporate these characteristics into the utility function, at which point the choice model can be described as a probabilistic forecast of a specific transportation mode.
However, these models also face many challenges, especially limitations in accuracy and diversity of crowd behavior.
With the progress of the times, these traffic pattern analysis tools have become quite mature and continue to influence the formulation of transportation policies. However, in the face of diverse traffic demands and rapidly changing environments, the accuracy and effectiveness of these analysis methods remains a question worth reflecting on. Can we really rely solely on these model predictions to formulate future transportation policies?
