Collaborative Research: Decision Processes in Human Navigation

成功地从一个地方导航到另一个地方可能需要做出艰难的决定。我们经常需要考虑可能路线的成本和收益。例如，市中心建筑物之间的最佳步行路径可能是天气宜人时的短户外路径，或者在酷热或寒冷的月份通过室内通道的较长路径。我们还利用我们的知识来决定在哪里搜索我们需要的东西。例如，经验丰富的司机知道，与住宅区相比，购物中心是更适合寻找加油站的地方。我们可能还需要决定，我们是否足够了解一个环境，依靠我们的记忆和方向感，或者是否应该使用手机上的地图应用程序。为了做出好的选择并避免迷路，我们需要依赖几个信息来源。一个重要的来源是我们所看到的，例如道路、小径和熟悉的地方。另一个重要的来源来自我们的身体：当我们走路和转身时，即使闭着眼睛，我们也能感觉到我们已经走了多远，我们面对的是什么方向。这些信息来源告诉我们，我们在哪里，我们要去哪里，以及到达那里有多难(例如，爬陡峭的山与绕着它走)。这项研究调查了人们如何做出这类决定，如何处理相互冲突的信息来源(例如，我们的方向感表明我们应该左转，但一个熟悉的地标指示我们应该右转)，以及如何使用导航辅助工具(例如，环境的头顶地图)。研究人员将使用人们选择和行动的数学模型来理解人脑如何存储和使用空间知识进行导航。结果可以为技术的使用提供参考，从视频游戏的移动界面到支持GPS的地图。研究人员探索导航决策和行动受以下因素影响的方式：(A)关于导航员位置和目标位置的空间线索(例如，环境中的地标、行走和转弯时的基于身体的线索)；(B)与可能的选择相关的成本(例如，努力、时间)；以及(C)导航员的个人特征(例如，空间能力、风险承受能力)。实验使用身临其境的虚拟现实来保持对视觉场景的严格控制，同时允许在导航过程中进行完全的物理运动；这项技术允许导航员在虚拟环境中行走和转弯，就像他们在现实世界中所做的那样。这些实验考察了1)当这些线索彼此不一致时，导航员如何使用来自多个空间线索的信息来找到目标；2)导航员如何考虑导航成本，比如在深沙中的短路和坚固地面上的较长路径之间进行选择；3)导航员如何利用先验知识做出导航决策，比如在一个不熟悉的城市中选择最有可能搜索餐厅的地方，因为他们知道餐馆更有可能位于商业区而不是居民区；以及4)导航员如何将科技提供的空间信息与视觉和身体运动提供的自然线索结合起来。人类导航认知过程的计算模型将被用来扩展人类决策和行动的解释性理论。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Navigating successfully from one place to another can require difficult decisions. We often need to consider the costs and the benefits of possible routes. For example, the best walking path between downtown buildings may be a short outdoor path when the weather is pleasant or a longer path through indoor passageways during overly hot or cold months. We also use our knowledge to make decisions about where to search for something that we need. Experienced drivers know, for example, that a strip mall is a better place to find a gas station than is a residential neighborhood. We also may need to decide whether we know an environment well enough to rely on our memories and sense of direction or should use the mapping app on our cell phones. To make good choices and to keep from getting lost, we need to rely on several sources of information. One important source is what we see, such as roads, trails, and familiar places. Another important source is from our bodies: As we walk and turn, even with our eyes closed, we have a sense of how far we have traveled and which directions we are facing. These sources of information tell us where we are, where we are headed, and how hard it will be to get there (e.g., climbing a steep hill vs. walking around it). This research investigates how people make these sorts of decisions, deal with conflicting sources of information (e.g., our sense of direction indicates that we should turn left but a familiar landmark indicates that we should turn right), and use navigation aids (e.g., an overhead map of the environment). The investigators will use mathematical models of people’s choices and actions to understand how the human brain stores and uses spatial knowledge for navigation. The results can inform the use of technology, ranging from movement interfaces for video games to GPS-enabled maps. The investigators explore the ways in which navigational decisions and actions are affected by (a) spatial cues about the navigator’s location and the goal location (e.g., landmarks in the environment, body-based cues from walking and turning), (b) costs associated with possible choices (e.g., effort, time), and (c) individual characteristics of the navigator (e.g., spatial ability, risk tolerance). Experiments use immersive virtual reality to maintain tight control over the visual scene while allowing for full physical movement during navigation; this technology allows navigators to walk and turn in virtual environments just as they do in the real world. The experiments examine 1) how navigators use information from multiple spatial cues to find a goal when those cues are inconsistent with one another; 2) how navigators account for navigational costs, as when choosing between a short path through deep sand or a longer path on firm ground; 3) how navigators use prior knowledge to make navigational decisions, as when selecting the most likely place to search for a restaurant within an unfamiliar city, knowing that a restaurant is more likely to be located in a business district than a residential neighborhood; and 4) how navigators combine spatial information from technology, such as that provided by a GPS-enabled map, with natural cues provided by vision and bodily movement. Computational models of the cognitive processes involved in human navigation will be used to expand explanatory theories of human decisions and actions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.