TRIS-HAR: Transmissive Reconfigurable Intelligent Surfaces-assisted Cognitive Wireless Human Activity Recognition Using State Space Models

Junshuo Liu,Yunlong Huang,Xin Shi,Rujing Xiong,Jianan Zhang,Tiebin Mi,Robert C. Qiu

2024-10-03

Abstract:Human activity recognition (HAR) using radio frequency (RF) signals has garnered considerable attention for its applications in smart environments. However, traditional systems often struggle with limited independent channels between transmitters and receivers, multipath fading, and environmental noise, which particularly degrades performance in through-the-wall scenarios. In this paper, we present a transmissive reconfigurable intelligent surface (TRIS)-assisted through-the-wall human activity recognition (TRIS-HAR) system. The system employs TRIS technology to actively reshape wireless signal propagation, creating multiple independent paths to enhance signal clarity and improve recognition accuracy in complex indoor settings. Additionally, we propose the Human intelligence Mamba (HiMamba), an advanced state space model that captures temporal and frequency-based information for precise activity recognition. HiMamba achieves state-of-the-art performance on two public datasets, demonstrating superior accuracy. Extensive experiments indicate that the TRIS-HAR system improves recognition performance from 85.00% to 98.06% in laboratory conditions and maintains high performance across various environments. This approach offers a robust solution for enhancing RF-based HAR, with promising applications in smart home and elderly care systems.

Signal Processing

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the performance degradation problem faced by existing radio frequency (RF) - based human activity recognition (HAR) systems in complex indoor environments, especially in through - wall scenarios. Specifically, traditional systems have limitations in the following aspects:

1. **Limited number of independent channels**: In traditional systems, the number of independent channels between the transmitter and the receiver is limited, resulting in a single signal propagation path.
2. **Multipath fading**: The multipath effect existing in the wireless environment can significantly reduce the signal quality, especially in complex indoor environments.
3. **Environmental noise**: The noise in the environment will further affect the clarity and recognition accuracy of the signal.
4. **Performance degradation in through - wall scenarios**: In cases where it is necessary to penetrate obstacles such as walls, signal attenuation and distortion are intensified, making activity recognition extremely difficult.

To solve these problems, the paper proposes a through - wall human activity recognition system (TRIS - HAR) based on transmissive reconfigurable intelligent surface (TRIS). This system improves the existing technology in the following ways:

- **Utilizing TRIS technology**: TRIS can actively reshape the wireless signal propagation path and create multiple independent paths, thereby enhancing signal clarity and improving recognition accuracy.
- **Introducing the HiMamba model**: This is an advanced state - space model that combines time - domain and frequency - domain information and can more accurately capture and recognize human activities.

### System advantages

1. **Signal enhancement**: By adjusting the phase and amplitude, TRIS can effectively improve the signal propagation quality, especially in multi - room or multi - wall environments.
2. **Rich spatio - temporal information**: The multiple independent and configurable paths generated by TRIS provide rich spatial information, which is helpful for detecting and recognizing subtle human activities.
3. **High accuracy**: The experimental results show that the recognition accuracy of the TRIS - HAR system has been increased from 85.00% to 98.06% and maintains high performance in different environments.

### Summary

The main contributions of the paper are:
- Proposing a brand - new through - wall human activity recognition system and applying TRIS to the HAR field for the first time.
- Designing the HiMamba model, which processes time - domain and frequency - domain information through a two - stream framework to achieve effective modeling and recognition of CSI data.
- Verifying the superior performance of the system through extensive experiments and demonstrating its adaptability and effectiveness in complex indoor environments.

These improvements make the TRIS - HAR system have broad application prospects in fields such as smart homes and elderly care.