Wi-Fi CSI Action Recognition with ESP32
2024 - Present • Hardware-based Action Recognition
Project Overview
Developing action recognition systems using Wi-Fi Channel State Information (CSI) with ESP32 microcontrollers to predict human actions using only WiFi signals. This research addresses privacy concerns associated with camera-based systems by exploring contactless human activity recognition through wireless signals, combining hardware design with machine learning.
Research Motivation
Traditional action recognition systems rely on cameras, which raise significant privacy concerns. This project explores an alternative approach using WiFi signals to predict human actions without visual surveillance. By leveraging Wi-Fi Channel State Information (CSI), we can detect and classify human activities while maintaining privacy.
Key Features
- ESP32-based Wi-Fi CSI data collection
- Privacy-preserving action recognition (no cameras required)
- Contactless human activity detection
- Hardware-software integration
- Machine learning for signal processing
Research Challenges
WiFi-based action recognition faces several technical challenges:
- Noise Interference: WiFi signals are susceptible to environmental noise, interference from other devices, and multipath propagation effects
- Signal Variability: CSI data can vary significantly due to environmental factors, device positioning, and signal attenuation
- Feature Extraction: Identifying meaningful features from raw CSI data that correlate with human actions
- Real-time Processing: Achieving low-latency action recognition suitable for practical applications
- Robustness: Developing models that generalize across different environments and device configurations
Research Focus
Exploring novel approaches to human activity recognition without cameras or wearables, using wireless signal analysis and embedded systems. The goal is to develop a privacy-preserving solution that can accurately predict human actions using only WiFi signals, while addressing the inherent challenges of noise interference and signal variability.
Technologies: ESP32 • Wi-Fi CSI • Python • Machine Learning • Embedded Systems
Links: