An empirical comparison of four technology-mediated travel survey methods

The increasing demana for advanced modelling methods, which can reflect complex travel activities of individuals, requires enhanced travel data collection methods. The introduction of GPS-assisted data collection methods has provided an alternative to the conventional methods of travel data collection. GPS-assisted data collection methods improve the accu- racy of data collection and enable capturing more details of individuals＇ travel behaviour. Recent technological advancements in smartphone-based positioning technologies and communication facilities have opened up new opportunities to apply smartphones as the media of GPS-assisted data collection. Although, different GPS-assisted methods have been employed recently, their performance has not been widely evaluated in real-world experi- ments compared to traditional data collection methods. Accordingly, this paper evaluates the performance of three GPS-assisted methods, namely handheld GPS tracking, smart- phone-based GPS tracking and smartphone-based prompted-recall data collection methods, in conjunction with the web-based data collection to shed light on different aspects of GPS- assisted data collection methods. These methods are compared in terms of the quality and accuracy of the collected data, the demographic attributes of participants and the specifi- cations of labelled trips. The results show that an appropriate employment of smartphones enhances the accuracy of data collection. It is also found that putting an extra burden on participants during a travel data collection survey results in lower trip-rates and poor data quality. Finally, it is found that the application of smartphone-assisted data collection methods help reporting non-motorised trips more accurately.

"Weather" transit is reliable? Using AVL data to explore tram performance in Melbourne,Australia

This paper uses automatic vehicle location （AVL） records to investigate the effect of weather conditions on the travel time reliability of on-road rail transit, through a case study of the Melbourne streetcar （tram） network. The datasets available were an extensive historica; AVL dataset as well as weather observations. The sample size used in the analysis included all trips made over a period of five years （2006-2010 inclusive）, during the morning peak （7 am-9 am） for fifteen randomly selected radial tram routes, all traveling to the Melbourne CBD create a linear model Ordinary least square （OLS） regression analysis was conducted to with tram travel time being the dependent variable. An alternative formulation of the model is also compared. Travel time was regressed on various weather effects including precipitation, air temperature, sea level pressure and wind speed; as well as indicator variables for weekends, public holidays and route numbers to investigate a correlation between weather condition and the on-time performance of the trams. The results indicate that only precipitation and air temperature are significant in their effect on tram travel time. The model demonstrates that on average, an additional millimeter of precipitation during the peak period adversely affects the average travel time during that period by approximately 8 s, that is, rainfall tends to increase the travel time. The effect of air temperature is less intuitive, with the model indicating that trams adhere more closely to schedule when the temperature is different in absolute terms to the mean operating conditions （taken as 15 ℃）.