Generation of coherent trip chains for travel behaviour analysis and modelling

Abstract

The generation of trip chains from National Household Travel Surveys (NHTS), frequently used both directly for data analysis as well as input for transport modeling, usually demands substantial post-processing. Transport models are highly dependent on information on the location type associated with each trip for the proper assignment and consequently estimation of transport flows between origins and destinations. Unfortunately, many NHTS only provide trip purposes without information about specific location types, especially in case of purposes such as ‘pick up / drop off someone’ or ‘accompanying adults’. Based on the German NHTS ‘MiD 2008 – Mobilitaet in Deutschland’, a step-wise method of data processing is introduced to specify reported locations by investigating the complex household context in which activity and trip planning took place. The step-wise identification of specific location sites considering the household context allows for the generation of additional 11% of qualified trip chains compared to the original data set, thus substantially enhancing the explanatory power of chain-based empirical analysis. Comparing the effect of different recoding schemes, it can be demonstrated that a systematic consideration of the household context proved most effective for joint trips of adults and children. Detailed recommendations are made to overcome the mainly design-based shortcomings of non-located trip purposes in future NHTS.