地图导航行为模式研究是现代地图学与地理信息科学领域的一个重要研究课题。当前相关研究主要是基于虚拟环境下开展导航行为模拟实验，但没有考虑真实环境与虚拟环境下用户对空间距离、方位、空间参考感知的差异性及其对眼动特征、导航行为的影响。为此，本项目基于人脑视觉注意机制与眼动跟踪技术，通过建立真实环境下用户视觉眼动行为与地图导航行为之间的关联，实现行人地图导航行为实时预测。主要包括：（1）行人地图导航行为多粒度建模方法；（2）行人导航眼动时空特征提取与轨迹建模方法；（3）行人地图导航行为模式离线识别方法；（4）行人地图导航行为实时预测模型与评价方法。本项目一方面将建立更准确、合理的地图导航行为识别与预测模型，揭示真实环境下人的地图导航行为模式；另一方面，提出的模型与方法可延伸应用于群体与个体的地理空间行为预测、个性化推荐与智慧导航等方面。

The study of map navigation behavior pattern is an important research topic in the field of modern mapping and geographic information science. At present, the related research is mainly based on the simulation experiment of navigation behavior in virtual environment, but it does not consider the difference of user's perception of spatial reference, orientation and space in real environment and virtual environment and its influence on eye movement feature and navigation behavior. Therefore, based on human brain visual attention mechanism and eye tracking technology, this project realizes real-time prediction of user navigation behavior by establishing the association between user's visual behavior and map navigation behavior under real environment. Mainly include: (1) multi-granularity modeling method of pedestrian map navigation behavior; (2) pedestrians eye movement time-space feature extraction and trajectory modeling method; (3) pedestrian map navigation behavior pattern offline recognition method; (4) pedestrian map navigation Real-time behavior prediction model and evaluation method. On the one hand, this project will establish a more accurate and reasonable model of map navigation behavior identification and prediction and reveal the human map navigation behavior under real environment; on the other hand, the proposed models and methods can be extended to the geospatial behavior of groups and individuals Forecast, personalized recommendation and smart navigation and so on.