A Trustworthy AIoT-enabled Localization System via Federated Learning and Blockchain

TL;DR

This work tackles privacy and security concerns in RSS fingerprint-based indoor localization by proposing DFLoc, a decentralized federated learning framework anchored in a blockchain. It introduces a DFLoc Validator Mechanism and a PoS-based consensus to prevent single-point failures and malicious updates, while delivering two specialized networks for 3D localization: DFLoc-BFC for floor classification and DFLoc-LLR for longitude-latitude regression. Extensive experiments on the UJIIndoorLoc dataset show that DFLoc maintains high accuracy and robust resilience to attacks and faults, significantly outperforming centralized federated learning baselines in adverse conditions. The framework enables trustworthy, privacy-preserving AIoT localization in smart buildings and points toward future real-world deployments using on-chain data networks.

Abstract

There is a significant demand for indoor localization technology in smart buildings, and the most promising solution in this field is using RF sensors and fingerprinting-based methods that employ machine learning models trained on crowd-sourced user data gathered from IoT devices. However, this raises security and privacy issues in practice. Some researchers propose to use federated learning to partially overcome privacy problems, but there still remain security concerns, e.g., single-point failure and malicious attacks. In this paper, we propose a framework named DFLoc to achieve precise 3D localization tasks while considering the following two security concerns. Particularly, we design a specialized blockchain to decentralize the framework by distributing the tasks such as model distribution and aggregation which are handled by a central server to all clients in most previous works, to address the issue of the single-point failure for a reliable and accurate indoor localization system. Moreover, we introduce an updated model verification mechanism within the blockchain to alleviate the concern of malicious node attacks. Experimental results substantiate the framework's capacity to deliver accurate 3D location predictions and its superior resistance to the impacts of single-point failure and malicious attacks when compared to conventional centralized federated learning systems.

Figures (12)

• Figure 1: DFLoc system overview. Clients handle tasks including model training, verification, and block mining, constituting a blockchain that can aggregate and distribute the global model, replacing the central server.
• Figure 2: Operations of DFLoc. In each round of the DFLoc learning phase, after downloading the global model, each client is assigned a specific role and completes the corresponding task. Subsequently, all clients aggregate the global model and update it with the stake record onto the blockchain.
• Figure 3: Training process of CFL and DFLoc-LLR under malicious attacks.
• Figure 4: Effect of malicious attacks on CFL and DFLoc-LLR.
• Figure 5: Training process of CFL and DFLoc-BFC under malicious attacks.