Dynamic Scene Adjustment for Player Engagement in VR Game

Design of the Dynamic Scene Adjustment (DSA) mechanism. Users are not only unilaterally affected by the environment; the decisions made at every moment and their status, performance, and experience can also affect the user’s state at the next moment. Thus, the DSA mechanism can be described with the following equations.

where $\rho$ is the world parameters, the function $DSA^{*}\left(\right)$ is the best DSA intervention strategy, $A_{t}$ is users’ status, $r_{t}$ is users’ current performance, $r_{t+1}^{*}$ is the user’s performance after enhanced by the DSA mechanism at the next moment, $A_{t+1}$ is the status when the user interacts with the scene adjusted by the DSA mechanism, the function $G\left(\right)$ is the game-play rules. $R^{*}$ is the best user overall interaction performance, which is also the adjustment goal of the DSA mechanism. $D$ is the user’s action, which satisfies the mapping relationship $f$ with $\rho$. The intervention strategy $DSA$ calculates the world parameters with users’ status and performance as input. Many intervention strategies exist, and the best strategy $DSA^{*}$ will obtain the most suitable parameter $\rho^{*}$. $\rho^{*}$ will cause the user to take the best action $D^{*}$. After the user’s state and action interact with the gameplay, their new status $A_{t}$ and performance $r_{t+1}$ will occur. This is the logic of the operation of the DSA mechanism. Therefore, the DSA mechanism mainly includes three modules. a) World parameters are the environmental variables that mainly affect the user’s performance and experience. b) Adjustment goal is the expected state or desired experience of the users (or the anticipated state the examiner hopes participants can complete). c) Intervention strategy is the calculation function of world parameters, which is the logic about how users achieve the adjustment goal.

Gameplay design with DSA mechanism. In this part, we introduce our DSA game to readers according to the three main modules required by DSA, as illustrated in Figure 1. We selected the background color in the VR environment as the world parameter $\rho$ and set it to either red or blue. Because the selected game is a kind of musical rhythm VR game, it has been proven that color can influence users’ attention, Red can stimulate users’ attention, and Blue will keep users’ attention stable [5]. Users’ attention will influence users’ experience [1] and performance [4]. We set the users’ game score (user performance) in the VR environment as the adjustment goal. Regarding the intervention strategy, we input the user’s real-time attention and instant performance into the DSA strategy module, dynamically adjusting the world parameters selected (color) in this game. Users’ instant performance is calculated by the user’s score changes within an adjacent time window. The adjustment strategies of the game are illustrated in Table 1. We assume that the user’s score ratio in the first time window is $S_{t}$, while in the second time window is $S_{t+1}$. Users’ instant performance $r_{t+1}=S_{t+1}-S_{t}$. The time window size is 2.5 seconds which is the hyper-parameters obtained after decoupling the game.

Experiment Setup The experiment was supervised and guided by the Ethics Committee of UESTC. 13 participants, including 4 female and 9 male students between 20 and 28 years old, completed the comparative experiment. The game with (or without) the DSA mechanism is used as the experimental (control) group. Participants were asked to be involved twice in each experimental and control group experiment in random order. To avoid the impact of proficiency, all participants must practice several times before the formal experiment. Participants are given enough rest time after each game to avoid the impact of fatigue. The user’s attention and task performance (game score) are collected and saved during the experiments. A short user experience interview is conducted after the experiment.

Experiment Results With the DSA mechanism, 83.3$\%$ of the participants improved their performance. The paired sample t-test results show that the test subjects performed better in the experimental group (M=86.13278, SD=8.79289) than in the control group (M=84.47861, SD=9.7518). The experimental results show that the DSA mechanism can improve the user’s task performance in the VR environment. In addition, according to interview feedback, almost all participants reported that the experimental group was more excited and they experienced better than the control group.

In summary, this paper proposes the Dynamic Scene Adjustment (DSA) mechanism for dynamically improved user interaction performance by adjusting the world parameters in the VR environment following the user’s real-time status and instant performance. In addition, we combined the DSA mechanism with a musical rhythm VR game and applied it to VR environment tasks. The experimental conclusions support the rationality of the mechanism. This work can help researchers think about dynamic regulation in VR environments from a new perspective. It can inspire the design of VR therapy, VR education, VR games, and other fields. Future research will expand the experimental data based on the existing DSA mechanism, adding more world parameters. In addition, we will want to establish a universal DSA mechanism through a data-driven approach.