Deep Learning-Based Transition Detection for Seamless Indoor-Outdoor Localization
Chanyeong Ju, Jaeho Jang, Jaehyun Yoo, IPIN LABS
Alternate Number 2
Indoor localization methods operate autonomously, independent of Global Navigation Satellite Systems (GNSS), which are primarily designed for outdoor positioning. However, the integration of these two distinct localization methodologies often yields erratic and delayed location estimates, particularly at transitional points such as building entrances. This paper proposes a transition detection algorithm leveraging deep learning to discern semi-indoor and semi-outdoor spaces, aiming to enhance the smoothness and accuracy of location estimation. The algorithm utilizes sensor data such as WiFi Received Signal Strength Indication (RSSI), Bluetooth Low Energy (BLE), and GNSS parameters like Dilution of Precision (DOP). During transitional phases, the positioning system employs Inertial Measurement Unit (IMU) tracking, seamlessly switching between indoor and outdoor positioning engines. The outdoor positioning engine relies on Kalman Filter (KF)-based tracking with IMU and GNSS data, while the indoor positioning algorithm utilizes WiFi and BLE fingerprinting techniques. Experimental validation involves a user transitioning from an outdoor to an indoor environment. Comparative analysis demonstrates the efficacy of the proposed transition detection algorithm in enhancing the continuity and accuracy of tracking performance.
For Attendees Call for Abstracts Registration Hotel Travel and Visas Exhibits Submit Kepler Nomination For Authors and Chairs Abstract Management Student Paper Awards Editorial Review Policies Publication Ethics Policies For Exhibitors Exhibitor Resource Center Marketing Resources Other Years Future Meetings Past Meetings