Recommendation-based Smart Indoor Navigation: Poster Abstract
Recommendation-based Smart Indoor Navigation: Poster Abstract
复制标题
基于推荐的智能室内导航：海报摘要
DOI：
10.1145/3054977.3057288
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Ku, Wei-Shinn
中科院分区：
文献类型：
--
作者：
Wang, Wenlu;Ku, Wei-Shinn
室内空间的定位必须依赖于传感设备（例如，射频识别（RFID）阅读器、WiFi路由器、蓝牙信标），而不是GPS设备。另一方面，我们可以构建一个智能的室内环境，通过各种传感设备促进各种空间服务。本文以空间导航为研究对象。由于室内环境的复杂性，我们认为室内导航策略不应该局限于最短路径。以购物中心为例，导航路径不仅应该是最短的路径，而且应该是对购物者有吸引力的路径。我们的目标是建立一个智能室内导航系统，它不仅可以通过以前的传感数据学习用户的行为，而且可以与异构设备一起工作。为此，我们提出了一种基于递归神经网络（RNN）的基于推荐的智能导航策略。该策略通过以下方式提供最佳用户体验：1)记忆用户历史数据；2)将导航与用户的室内行为模型重叠；3)根据传感器设备的实时检测结果提出建议。
Localization in indoor spaces has to rely on sensing devices (e.g., Radio Frequency Identification (RFID) readers, WiFi routers, bluetooth beacons) rather than GPS devices. On the other side, we could build a smart indoor environment that facilitates all types of spatial services with various sensing devices. In this paper, we focus on the topic of spatial navigation.Due to the complexity of indoor environments, we believe the indoor navigation strategy should not be limited to the shortest path. Taking shopping centers for example, a navigation path should be not only the shortest path, but also an attractive route to the shopper. We aim to build a smart indoor navigation system, which not only learns the user's behavior through previous sensing data, but also enjoys working with heterogeneous devices. Therefore, we propose a novel recommendation based smart navigation strategy with Recurrent Neural Network (RNN). This strategy provides optimal user experience by: 1) Memorizing the user's historical data; 2) Overlapping the navigation with the user's indoor behavior model; 3) Making recommendations based on real-time detections from sensor devices.