Survey of Indoor Localization Based on Deep Learning  

作　　者：Khaldon Azzam Kordi Mardeni Roslee Mohamad Yusoff Alias Abdulraqeb Alhammadi Athar Waseem Anwar Faizd Osman 

机构地区：[1]Centre for Wireless Technology,Faculty of Engineering,Multimedia University,Cyberjaya,63100,Malaysia [2]Communication Systems&Networks Research Lab,Malaysia-Japan International Institute of Technology,Universiti Teknologi Malaysia,Kuala Lumpur,54100,Malaysia [3]Department of Electrical Engineering(DEE),International Islamic University,Islamabad,44000,Pakistan [4]Rohde&Schwarz(M)Sdn Bhd,Shah Alam,40150,Malaysia

出　　处：《Computers, Materials & Continua》2024年第5期3261-3298,共38页计算机、材料和连续体（英文）

基　　金：the Fundamental Research Grant Scheme-FRGS/1/2021/ICT09/MMU/02/1,Ministry of Higher Education,Malaysia.

摘　　要：This study comprehensively examines the current state of deep learning (DL) usage in indoor positioning.It emphasizes the significance and efficiency of convolutional neural networks (CNNs) and recurrent neuralnetworks (RNNs). Unlike prior studies focused on single sensor modalities like Wi-Fi or Bluetooth, this researchexplores the integration of multiple sensor modalities (e.g.,Wi-Fi, Bluetooth, Ultra-Wideband, ZigBee) to expandindoor localization methods, particularly in obstructed environments. It addresses the challenge of precise objectlocalization, introducing a novel hybrid DL approach using received signal information (RSI), Received SignalStrength (RSS), and Channel State Information (CSI) data to enhance accuracy and stability. Moreover, thestudy introduces a device-free indoor localization algorithm, offering a significant advancement with potentialobject or individual tracking applications. It recognizes the increasing importance of indoor positioning forlocation-based services. It anticipates future developments while acknowledging challenges such as multipathinterference, noise, data standardization, and scarcity of labeled data. This research contributes significantly toindoor localization technology, offering adaptability, device independence, and multifaceted DL-based solutionsfor real-world challenges and future advancements. Thus, the proposed work addresses challenges in objectlocalization precision and introduces a novel hybrid deep learning approach, contributing to advancing locationcentricservices.While deep learning-based indoor localization techniques have improved accuracy, challenges likedata noise, standardization, and availability of training data persist. However, ongoing developments are expectedto enhance indoor positioning systems to meet real-world demands.

关 键 词：Deep learning indoor localization wireless-based localization 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]