Online experiments as a supplement of automated driving simulator studies: a methodological insight

Online experiments offer several advantages compared to lab studies. We propose that reaction times in takeover requests in automated driving research can be evaluated using online experiments. Furthermore, we assume that online studies can complement simulator studies to some extent. However, online studies have several pitfalls (e.g. an uncontrolled environment). Affecting factors can occur in hardware, software, and human behavior. We provide a detailed methodology and technology section to reduce possible confounding factors and to increase the validity of automated driving online studies. Furthermore, to test the effect of the proposed methodology, we conducted a user study with a takeover scenario as online (n = 42) and simulator (n = 17) experiment. First results indicate that reaction times did not differ between the simulator and online experiment. Thus, we conclude that with some methodical refinements and practices, similar results can be achieved in online studies compared to simulator studies.