A Discussion on the Consistency of Driving Behavior across Laboratory and Real Situational Studies

A Discussion on the Consistency of Driving Behavior across Laboratory and Real Situational Studies ∗1 Graduate Hitoshi Terai ∗1,2 (terai@is.nagoya-u.ac.jp) Kazuhisa Miwa ∗1 (miwa@is.nagoya-u.ac.jp) Hiroyuki Okuda ∗3 (h okuda@nuem.nagoya-u.ac.jp) Yuichi Tazaki ∗4 (tazaki@nuem.nagoya-u.ac.jp) Tatsuya Suzuki ∗4 (t suzuki@nuem.nagoya-u.ac.jp) Kazuaki Kojima ∗5 (koj@aoni.waseda.jp) Junya Morita ∗6 (j-morita@jaist.ac.jp) Akihiro Maehigashi ∗1 (mhigashi@cog.human.nagoya-u.ac.jp) Kazuya Takeda ∗1 (kazuya.takeda@nagoya-u.jp) School of Information Science, Nagoya University, Nagoya, Aichi, 464–8601 Japan Japan Science and Technology Agency, Chiyoda, Tokyo, 102–8666 Japan ∗3 Green Mobility Collaborative Research Center, Nagoya University, Nagoya, Aichi, 464–8601 Japan ∗4 Graduate School of Engineering, Nagoya University, Nagoya, Aichi, 464–8601 Japan ∗5 Faculty of Human Sciences, Waseda University, Tokorozawa, Saitama, 359–1192 Japan ∗6 School of Knowledge Science, Japan Advanced Institute of Science and Technology, Nomi, Ishikawa, 923–1211 Japan ∗2 CREST, Abstract This study investigated the degrees of consistencies in driving behavior when operating a real system (real car), a virtual sys- tem (high fidelity driving simulator), and a laboratory system (computer driving game). The same tendency of behavioral consistencies was confirmed among the three systems: i.e., the steering operation demonstrated the highest behavioral consis- tencies, followed by the acceleration and braking operations, respectively. The individuality of driving behavior emerged more in the braking and acceleration operations than in the steering operation. The same tendency for behavioral consis- tencies of braking, acceleration, and steering operations was confirmed in each of the three systems. Keywords: behavioral consistency; driving behavior; individ- ual differences; virtual environments Introduction In studies of human factors, analyses of human behavior are usually conducted in actual environments using observational methods. However, advances in computer technology are now facilitating experiments on human factors by using var- ious simulators because they provide a convenient and safe method for assessing human behavior. Thus many studies about human behavior in serious situations that may lead to accidents have been performed, such as people driving cars, operating airplanes, controlling industrial plants (e.g., dos Santos et al., 2008; Kemeny, 2003; J. D. Lee et al., 2002; Met- zger & Parasuraman, 2001; Parasuraman et al., 1996; Wick- ens & Alexander, 2009). Driving simulators in particular have played an important role in automobile human factors research for more than three decades. Various studies using driving simulators have examined not only basic character- istics of driving behavior but also applied investigations of those effects of drinking and aging that relate to social prob- lems because using automobiles is a major part of our daily lives (e.g., H. C. Lee et al., 2003; Mets et al., 2011; Pradhan et al., 2005; Rizzo et al., 1997). However, virtual systems cannot simulate real systems completely. Therefore, many researchers agree that an exam- ination of their validity is a crucial component in any study. The validity of driving simulators has previously been evalu- ated through a comparison of behavior when driving real cars and simulators (e.g., T¨ornros, 1998; Godley et al., 2002; Un- derwood et al., 2011; Shechtman et al., 2009; Mayhew et al., 2011). Previous studies have discussed both commonalities and specificities in the distributions of specific errors or char- acteristics of specific behaviors when operating real and vir- tual systems. Such discussions have an essential assumption of the consistency of behavioral characteristics when driving vehicles. However, we do not know to what human driving behavior is consistent. In the present study, we examined be- havioral consistency (BC) when driving vehicles on road and using simulators. The purpose of this study is to reveal the degree of BC by analyzing three basic operations of driving behavior: braking, acceleration, and steering operations. First, we investigate the BCs for the three operations when driving a real car. Then, we study the BCs in two other types of systems: a virtual system as a high fidelity driving simulator and a laboratory system as a low fidelity driving simulator (similar to a computer driving game). The following outlines our basic strategies for the investigation. Imagine a situation in which drivers repeatedly drive on a specific course. The BC within each participant shows the degree of consistency in individual behavior when repeat- edly driving on the same course. We also calculate the BCs across participants, demonstrating the degree of consistency in the general characteristics of human behavior independent of each participant’s individuality. We refer to the former as the intrapersonal BC and the later as the interpersonal BC. In our analyses, the interpersonal BC is treated as the base- line because it reflects the generality of BCs across partici-