ABSTRACT There is a global consensus on the urgent need to promote sustainable modes of mobility such as the bicycle. This realisation is accompanied by an increased promotion of the research on bicyclists’ behaviour to plan bicycle-friendly road infrastructure effectively. Some of the commonly used methods of data collection in this research include surveys and interviews, which help in capturing the cyclists’ perceptions and conscious behaviour. However, these data collection methods are susceptible to reporting and hypothetical biases. Hence, recent studies have started exploring the use of physiological measures, such as the Galvanic Skin Response (GSR), as the objective measures of cyclists’ subconscious behaviour. The GSR potentially captures the perceived stress of the cyclists and can be easily recorded using sensor-based wearables. For the processing and analysis of the data thus collected, it is necessary to identify the appropriate statistical tools. Previous studies have discussed the merits, demerits, scope, and limitations of using GSR in urban cyclists’ behaviour analysis at length. However, methodological advancements expected in future studies in this domain need a comprehensive understanding of the currently used statistical methods and tools to build upon. This review facilitates them by providing an overview of these methods and tools, and by categorising existing studies based on their purpose (Methodological Exploration or Stimuli Response Analysis), methodological approach (Inductive or Deductive), method of data processing (splitting GSR signal, counting GSR peaks, measuring GSR peaks, taking Log of GSR, or curve fitting), and method of data analysis (map matching, descriptive analysis, hypothesis testing or Generalized Linear Modelling). The review further concludes that future studies need to find the correct balance between the granularity of GSR data to be preserved (given higher granularity ensures a higher level of detail in the analysis) and the resource consumption owing to its procedural and computational complexity.
Read full abstract