| 摘要: | 近年互動式的虛擬環境（virtual environment, VE）有許多以探索虛擬空間為主要使用者體驗（user experience）的應用，但有限於目前VE技術大多提供視覺與聽覺兩種刺激，在人機介面上仍仰賴鍵盤、滑鼠或搖桿，無法提供完整感官刺激與自然的體感操作，容易造成使用者在其間感到定位困難甚至迷路，以致無法流暢體驗內容而喪失繼續使用的意願。因此，如何導引使用者在虛擬環境中不迷失方向，設計出可輔助尋路（wayfinding）效率的VE是人機互動領域重要的研究議題。本研究目的為探討使用者在3D虛擬空間中的尋路行為會受到哪些因素的影響，以及這些因素影響了尋路行為的哪些面向。為了具體瞭解各因素如何影響VE中的尋路行為，本研究設計一個模擬3D迷宮尋物的受試者作業並執行一個尋路實驗，實驗的兩個受試者間獨變項為迷宮空間尺寸（大、中、小）與輔助地標類型（無地標、彩色地標、形狀地標）。實驗流程為受試者接受眼動追蹤儀器校正後，觀看以反方向行進拍攝的路線提示影片、以習得找到目標物的正確路徑，而後進行尋路實驗，完成後進行回溯訪談和填寫量表及簡易的背景調查，統計受試者的遊戲經驗、性別、與空間能力評量。實驗的依變項為尋路表現測量、尋路過程中的眼動測量、以及回溯訪談。尋路表現測量包括：受試者任務的完成時間、試誤次數、猶豫次數與折返次數。眼動測量為觀看路線提示影片與進行尋路實驗時觀看彩色與形狀地標區域的凝視次數。透過雙因子因素分析以及回歸分析實驗結果、並搭配訪談資料之後，本研究發現：熟悉第一人稱遊戲的受試者尋路效率較佳；男性受試者較常以試誤作為尋路策略（表現出較多次的試誤和折返行為）、而女性較常以蒐集和整合回想空間線索作為尋路策略（表現出較多的猶豫行為）；常玩遊戲的受試者注意力、尋路效率較差（較少凝視地標）；自認空間能力好的受試者較少猶豫行為；而透過眼動資料也發現對於輔助地標投注較多視覺注意力的受試者可較快找到目標。空間大小與地標類型也明顯影響尋路行為：受試者在不同尺寸的迷宮中表現出不同的尋路策略，在中等複雜度的空間中最專注於搜索與回憶空間訊息、但尋路效率最低。受試者在色彩地標的輔助下表現出最好的尋路效率以及對地標投注最少的凝視行為，反之形狀地標需要受試者投注較多注意力、但尋路效率卻最低，甚至低於沒有任何輔助地標的狀況。本研究以實驗方式探討了影響以迷宮尋路為應用情境的使用者行為，並發現VE空間複雜度、輔助地標設計、使用者本身經驗與背景等因素對於尋路行為與效率的影響，研究結果有助於作為改善虛擬空間的使用者經驗之依據。
The techniques of virtual environments (VE) have been used in many user interfaces providing user experience involving spatial interaction and exploration. However, the limited devices of human-computer interaction such as mouse, joystick, keyboards and flat displays provide incomplete sensational and perceptual information while users of the applications presented with VE. This could constrain the availability of spatial information to users and lead to disorientation and lost in VEs. Due to the importance of improving navigation design and wayfinding efficiency in common VEs, this study aims at investigating the plausible factors affecting the wayfinding behavior in VE. The researcher designed a pack of 3D virtual mazes and a target finding task to carry out a wayfinding experiment. We manipulated two between-subject factors, the spatial complexity of virtual mazes (three levels, 12, 8, and 4 intersections) and the types of navigation aids (three levels, color landmarks, shape landmarks, and no aids). During the experiment, the participants received eye-tracker calibration at first, then watched a video presenting the route of finding the target which filmed with inverse direction. After viewing this cueing video the participants conducted the wayfinding trials. The wayfinding behavioral metrics include the complete time, try-error count, stop count, and return count. When completing the task, the participants filled several questionnaires for quantifying their spatial ability, wayfinding styles, and experience about using VEs. The participants eye movements during watching cueing video and finding the target were also recorded. Incorporating with the interview feedbacks the wayfinding strategies could be revealed. These data and the analysis of Two-way ANOVA and regression suggest: the spatial complexity and types of navigation aids significantly affect the wayfinding behaviors in multiple metrics; the participants with more gaming experience result in better wayfinding efficiency but deploying less attention on landmarks; male participants tend to adopt try-error strategy, while female participants appear spending more time to collect and integrate spatial information. The eye movement data also reveal a positive correlation between intensive deployment of visual attention and wayfinding efficiency. In general, this empirical study examined multiple factors affecting the user behaviors in simulated mazes and explored the impact of spatial complexity, design of navigation aids, and users’ background on wayfinding efficiency and strategy. These findings regarding would provide valuable information for improving future VE design. |
