Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

AdaFed-LDR: Adaptive Federated Learning with Layerwise Dynamics Regularization for Robust Wi-Fi Localization.

Kaito Harada1, Hirofumi Natori2, Makoto Koike2

  • 1Graduate School of Integrated Science and Technology, Shizuoka University, Hamamatsu 432-8011, Shizuoka, Japan.

Sensors (Basel, Switzerland)
|May 27, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Thymic hyperplasia after chemotherapy for ovarian malignancy: two cases complicated by hyperthyroidism.

Journal of surgical case reports·2026
Same author

Case report of endometrioid carcinoma arising from the paraovarian cyst.

Journal of surgical case reports·2025
Same author

Analysis of oceanic suspended particulate matter in the western North Pacific using the complex amplitude sensor.

Scientific reports·2024
Same author

Source contribution of black carbon aerosol during 2020-2022 at an urban site in Indo-Gangetic Plain.

The Science of the total environment·2024
Same author

Black carbon scavenging by low-level Arctic clouds.

Nature communications·2023
Same author

Physical and chemical properties of PM<sub>1</sub> in Delhi: A comparison between clean and polluted days.

The Science of the total environment·2023

AdaFed-LDR enhances Wi-Fi Channel State Information (CSI) localization by balancing stability and plasticity in federated learning. This method improves accuracy in known and unseen environments, addressing privacy and data heterogeneity challenges.

Area of Science:

  • * Computer Science
  • * Electrical Engineering
  • * Signal Processing

Background:

  • * Wi-Fi Channel State Information (CSI) enables high-precision indoor localization.
  • * Federated learning for CSI localization faces challenges with data privacy and statistical heterogeneity (non-IID) due to environment dependency.
  • * This heterogeneity leads to a stability-plasticity trade-off, impacting performance in known versus unseen environments.

Purpose of the Study:

  • * To address the stability-plasticity trade-off in federated learning for CSI-based indoor localization.
  • * To propose a novel method, AdaFed-LDR, that enhances localization accuracy across diverse environments while preserving data privacy.
  • * To improve the adaptability of federated learning models to unseen environments without compromising performance in known ones.
Keywords:
Channel State Information (CSI)domain adaptationfederated learningindoor positioning systemlayerwise dynamicsmultipath interferencenon-IIDprivacy preservation

Related Experiment Videos

Main Methods:

  • * Development of AdaFed-LDR, combining server-side Confidence-Weighted Adaptive Aggregation and client-side Layerwise Dynamics Regularization (LDR).
  • * Confidence-Weighted Adaptive Aggregation recalibrates client contributions using feature covariance changes.
  • * Layerwise Dynamics Regularization (LDR) applies depth-dependent constraints to preserve general features and allow environment-specific adaptation.

Main Results:

  • * AdaFed-LDR achieved a mean localization error (MLE) of 0.41 cm in known environments, an 88.2% improvement over FedAvg.
  • * In domain generalization to unseen environments, AdaFed-LDR achieved an MLE of 218.2±2.8 cm, outperforming FedPos.
  • * With minimal adaptation data (one sample per reference point), MLE improved to 21 cm, demonstrating significant adaptability.

Conclusions:

  • * AdaFed-LDR effectively addresses the stability-plasticity trade-off in federated CSI localization.
  • * The combined approach of adaptive aggregation and LDR yields superior performance compared to individual components.
  • * AdaFed-LDR offers a reproducible and effective solution for robust indoor localization across multiple environments.