Influence of GPS and Self-reported Data in Travel Demand Models

• Published in: Procedia, social and behavioral sciences Vol. 162; pp. 467 - 476
• Main Authors: Ribeiro, Marcelle D., Larrañaga, Ana M., Arellana, Julian, Cybis, Helena B.B.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 19.12.2014
• Subjects: GPS-data; Self-reported data; Travel demand models; Self-reported data; GPS-data; Travel demand models
• ISSN: 1877-0428; 1877-0428
• DOI: 10.1016/j.sbspro.2014.12.228

Data on household travel patterns represent key information to the development of travel demand models. The technology of Global Positioning Systems (GPS) may substitute or be used in association with traditional data collection approaches. However, it is important to know how the quality of this information influences the results for planning purposes, such as in travel demand analysis. The objective of this study is to evaluate the influence of different sources of travel information - GPS-recorded compared to self-reported - in travel demand models. Several structures of discrete choice models were tested to represent choice behavior: multinomial logit, mixed logit with random coefficients and nested logit, trying to include possible correlations between alternatives and heterogeneity of individuals. Subjects were recruited from a list of contacts of the Transport Laboratory at the Federal University of Rio Grande do Sul, Brazil. The results showed that GPS technology collects the travel patterns more precisely reducing the bias by collecting data from short trips not reported in traditional surveys. The models estimated with GPS data showed greater significance due to less measurement error. The cost of processing GPS information must be considered. An adequate modeling with self-reported data, by more complex models incorporating heterogeneity and correlation among alternatives, allowed an equivalent adjustment to those estimated with GPS data. The self-reported data is less precise due to respondents under/overestimation of travel times. More complex models allow capturing measurement errors inherent to self-reported travel surveys.