Optimization and Evaluation of Multi-Detector Deep Neural Network for High Accuracy Wi-Fi Fingerprint Positioning
Journal
IEEE Internet of Things Journal
Date Issued
2022
Author(s)
Abstract
To fulfill the need for high accuracy indoor positioning in many location-based services (LBS) and the emerging Internet of Things (IoT) applications, in this paper we propose a novel scene-analysis positioning solution of Multi-Detector Deep Neural Network (DNN) architecture, with preprocessing steps, model optimization techniques, and variance estimation methods. During the off-line site-surveying phase in our approach, fingerprint databases are created by purposely built robotic surveying devices traversing the target site to gather perceivable Wi-Fi and other signals including to create spatial positioning models for further use in the online positioning phase. The intricate non-linear relationship between fingerprints and spatial positions are thus resolved by the Multi-Detector DNN in our approach. Hyper-parameter analyses were conducted to further optimize our proposed Multi-Detector model in terms of complexity, achieving at least 6.7 times of parameter complexity reduction while retaining <1% degradation of 0.9m (3ft) positioning accuracy level. IEEE
Subjects
Deep Neural Network; Detectors; Estimation; Feature extraction; Fingerprint recognition; Indoor positioning; Internet of Things; Internet of Things; Model Optimization.; Neural networks; Wi-Fi Fingerprinting; Wireless fidelity
Other Subjects
Complex networks; Deep neural networks; Feature extraction; Indoor positioning systems; Internet of things; Location based services; Surveys; Telecommunication services; Wireless local area networks (WLAN); Features extraction; Fingerprint Recognition; High-accuracy; Indoor positioning; Model optimization; Model optimization.; Multi-detectors; Neural-networks; Wi-fi fingerprinting; Wireless fidelities; Wi-Fi
Type
journal article