Travel mode imputation using GPS and accelerometer data from a multi-day travel survey

Over the past decade, interest has grown in using Global Positioning System (GPS) data to augment or even replace traditional travel survey diaries. If the full potential of this new class of data is to be realized, processing techniques will need to be standardized and automated to some degree. This paper develops a multinomial logit (MNL) model to impute travel mode from GPS and accelerometer data. The MNL model is the workhorse of travel demand modeling, but this is its first application to GPS data processing. A web-based recall survey was used to create an estimation dataset of over 900 trip stages from a larger multi-day GPS travel survey in Portland, Oregon. Special attention is given to the imputation of bicycle travel, which has been rare in the US. The authors find that the MNL model performs well overall. Transit network data and accelerometer data significantly improve model fit and prediction success for specific modes. Accelerometer data is found to be particularly helpful in predicting walk and bike modes. No benefit is found to segmenting by age. The MNL model shows strong potential for automated GPS processing and should be relatively easy to implement elsewhere.