BTS: Bifold Teacher-Student in Semi-Supervised Learning for Indoor Two-Room Presence Detection Under Time-Varying CSI

TL;DR

This paper introduces BTS, a semi-supervised teacher-student model for indoor two-room presence detection using CSI data, which maintains high accuracy despite environmental changes and reduces labeling effort.

Contribution

The paper proposes a novel bifold teacher-student SSL approach with an enhanced loss function for robust indoor presence detection under dynamic conditions.

Findings

Achieves around 98% accuracy after retraining with unlabeled data.

Maintains 93% accuracy under environmental layout changes.

Outperforms existing SSL models in detection accuracy.

Abstract

In recent years, indoor human presence detection based on supervised learning (SL) and channel state information (CSI) has attracted much attention. However, existing studies that rely on spatial information of CSI are susceptible to environmental changes which degrade prediction accuracy. Moreover, SL-based methods require time-consuming data labeling for retraining models. Therefore, it is imperative to design a continuously monitored model using a semi-supervised learning (SSL) based scheme. In this paper, we conceive a bifold teacher-student (BTS) learning approach for indoor human presence detection in an adjoining two-room scenario. The proposed SSL-based primal-dual teacher-student network intelligently learns spatial and temporal features from labeled and unlabeled CSI datasets. Additionally, the enhanced penalized loss function leverages entropy and distance measures to distinguish drifted data, i.e., features of new datasets affected by time-varying effects and altered from the original distribution. Experimental results demonstrate that the proposed BTS system accomplishes an averaged accuracy of around 98% after retraining the model with unlabeled data. BTS can sustain an accuracy of 93% under the changed layout and environments. Furthermore, BTS outperforms existing SSL-based models in terms of the highest detection accuracy of around 98% while achieving the asymptotic performance of SL-based methods.