Throughout the project, we successfully accomplished most of our objectives, which were divided into five main components. Initially, we conducted a qualitative study to establish a theoretical framework for exploring dynamic tracking and SA in football. Additionally, we undertook a systematic review of SA research across various fields to inform research on SA in sports, particularly football. Subsequently, we delved into investigating the eye movement characteristics of sports athletes in dynamic circumstances. Following this, we developed the VR tool integrated with eye tracking and investigated football players dynamic tracking and SA within a three-dimensional (3D) environment. Furthermore, we explored the SA transformation between 3D and 2D perspectives.
1) The qualitative research comprised explicitation interviews conducted with eight professional football players from two teams in Finnish Premier League. Data collected from the qualitative research were coded and analyzed through thematic analysis. The analysis process used a combination of deductive (i.e. theory-driven) and inductive approach. The outcomes provided profound insights into players' cognitive processes on the football field, corroborating existing SA theories while also uncovering novel insights, such as the impact of risk analysis, player characteristics, and emotions on SA in football. The research resulted in the development of a conceptual model elucidating SA and decision-making in football (Figure 1).
2) The systematic yielded a total of 2,547 empirical studies spanning the period from 1975 to 2022. The review categorized empirical studies on SA into 11 distinct fields. By analyzing the connections and distinctions between these fields, the review further grouped them into three clusters: ergonomics-centered, human-centered, and machine-centered groups. This clustering framework provided valuable insights into the classification of SA application fields, offering a comprehensive picture of SA research. Notably, the review identified sports as a human-centered field within which SA research holds significance. Such categorization offers practical guidance for conducting SA research, particularly in the context of football.
3) We examined the gaze behavior of handball players in multiple object tracking (MOT) and multiple identity tracking (MIT) tasks. Participants were tasked with tracking 3–5 targets among 10 moving objects, with objects being identical in MOT and differing in shape and color in MIT. The results revealed that athletes exhibited more target-oriented eye movements in both tracking tasks compared to non-athletes (Figure 3). These findings offer valuable insights into analyzing football players' eye movements within simulated football scenarios.
4) We developed the VR tool integrated with eye tracking, in which a virtual football field is created, allowing for the simulation of real football game scenarios with multiple players running on the field following trajectories extracted from real football games (Figure 2). Participants can adopt the perspective of any designated player on the field to see the scenario from the first-person view. Utilizing VR controllers, participants engage in a series of tasks designed to evaluate and potentially enhance their SA and decision-making skills within the scenarios.
We employed the VR tool to conduct quantitative research in both Finland and China. The study comprised testing 22 football players from three professional teams in Finland, 16 student athletes and 28 amateurs in Beijing Sport University, representing both genders and players from senior and junior levels. During these sessions, we recorded players' eye movements while they viewed the scenarios and captured their behavioral responses and verbal report as well. We are working on the data analysis and writing a paper to report the findings.
5) We also developed a version of the VR tool for 3D-2D transformation, allowing participants to view the same scenarios from a first-person view or a bird's-eye view and reproduce them on a virtual 2D plane (Figure 4), enabling a comparison of SA creation and updating from egocentric and allocentric views.
We conducted a study on players' SA on 2D display prior to investigate transformation between 3D and 2D. The results (see Figure 5) showed that from 2D perspective student football athletes could maintain the locations of about eight players, along with the ball, in accurate memory for complex football scenarios, showing a greater quantity of accurate memory than non-athletes, whereas the precision of memory did not differ between the two groups. On the other hand, non-athletes falsely memorized the locations of a larger number of attacking players and tended to make more guesses about the locations of defending players than football athletes. Moreover, football athletes displayed superior precision when guessing player locations compared to non-athletes. The 2D experiment has laid both a theoretical and an empirical foundation for the 3D-2D transformation experiment. We are now working on using the VR tool to test the transformation of participants' SA between 3D and 2D perspectives.
