Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep learning approaches to indoor wireless channel estimation for low-power communication

Samrah Arif, Muhammad Arif Khan, Sabih Ur Rehman

TL;DR

This work tackles indoor LP-IoT channel estimation by reframing RSSI as a key channel metric and applying two Fully Connected Neural Network models. Model A targets stationary indoor LoS/NLoS scenarios, while Model B adapts to changing receiver locations, both trained on real indoor RSSI data with time-series framing and Leaky ReLU activations. The results show substantial reductions in estimation error compared to traditional methods and other DL approaches, with Model A achieving up to ~99% relative MSE reduction and Model B ~90% relative reduction, while maintaining computational efficiency suitable for resource-constrained devices. The study demonstrates practical viability for RSSI-driven DL channel estimation in LP-IoT and outlines clear paths for deployment and extension to more complex indoor environments.

Abstract

In the rapidly growing development of the Internet of Things (IoT) infrastructure, achieving reliable wireless communication is a challenge. IoT devices operate in diverse environments with common signal interference and fluctuating channel conditions. Accurate channel estimation helps adapt the transmission strategies to current conditions, ensuring reliable communication. Traditional methods, such as Least Squares (LS) and Minimum Mean Squared Error (MMSE) estimation techniques, often struggle to adapt to the diverse and complex environments typical of IoT networks. This research article delves into the potential of Deep Learning (DL) to enhance channel estimation, focusing on the Received Signal Strength Indicator (RSSI) metric - a critical yet challenging aspect due to its susceptibility to noise and environmental factors. This paper presents two Fully Connected Neural Networks (FCNNs)-based Low Power (LP-IoT) channel estimation models, leveraging RSSI for accurate channel estimation in LP-IoT communication. Our Model A exhibits a remarkable 99.02% reduction in Mean Squared Error (MSE), and Model B demonstrates a notable 90.03% MSE reduction compared to the benchmarks set by current studies. Additionally, the comparative studies of our model A with other DL-based techniques show significant efficiency in our estimation models.

Deep learning approaches to indoor wireless channel estimation for low-power communication

TL;DR

This work tackles indoor LP-IoT channel estimation by reframing RSSI as a key channel metric and applying two Fully Connected Neural Network models. Model A targets stationary indoor LoS/NLoS scenarios, while Model B adapts to changing receiver locations, both trained on real indoor RSSI data with time-series framing and Leaky ReLU activations. The results show substantial reductions in estimation error compared to traditional methods and other DL approaches, with Model A achieving up to ~99% relative MSE reduction and Model B ~90% relative reduction, while maintaining computational efficiency suitable for resource-constrained devices. The study demonstrates practical viability for RSSI-driven DL channel estimation in LP-IoT and outlines clear paths for deployment and extension to more complex indoor environments.

Abstract

In the rapidly growing development of the Internet of Things (IoT) infrastructure, achieving reliable wireless communication is a challenge. IoT devices operate in diverse environments with common signal interference and fluctuating channel conditions. Accurate channel estimation helps adapt the transmission strategies to current conditions, ensuring reliable communication. Traditional methods, such as Least Squares (LS) and Minimum Mean Squared Error (MMSE) estimation techniques, often struggle to adapt to the diverse and complex environments typical of IoT networks. This research article delves into the potential of Deep Learning (DL) to enhance channel estimation, focusing on the Received Signal Strength Indicator (RSSI) metric - a critical yet challenging aspect due to its susceptibility to noise and environmental factors. This paper presents two Fully Connected Neural Networks (FCNNs)-based Low Power (LP-IoT) channel estimation models, leveraging RSSI for accurate channel estimation in LP-IoT communication. Our Model A exhibits a remarkable 99.02% reduction in Mean Squared Error (MSE), and Model B demonstrates a notable 90.03% MSE reduction compared to the benchmarks set by current studies. Additionally, the comparative studies of our model A with other DL-based techniques show significant efficiency in our estimation models.
Paper Structure (20 sections, 7 equations, 6 figures, 1 table)

This paper contains 20 sections, 7 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. System Model
  4. Problem Formulation
  5. DL Model Development
  6. Model A
  7. Model Training
  8. FCNN-based Model B
  9. Experimental Data Collection
  10. Scenario 1: Stationary Devices with Fixed Distances
  11. Case I
  12. Case II
  13. Scenario 2: Receiver's Location Variation Assessment
  14. Case I
  15. Case II
  16. ...and 5 more sections

Figures (6)

  • Figure 1: Visual representation of FCNN-based LP-IoT estimation model.
  • Figure 2: Visual representation of two FCNN-based estimation models using RSSI for LP-IoT networks.
  • Figure 3: Experimental setup for scenario 1-Fixed distance between transmitter and receiver.
  • Figure 4: Experimental setup for scenario 2-Varying locations of the receiver at a fixed distance from the transmitter.
  • Figure 5: FCNN-based Model A estimating RSSI at the next time stamp using historical RSSI from the LP-IoT stationary devices.
  • ...and 1 more figures