TL;DR
This paper presents a novel approach to capturing electrocardiogram (ECG) signals using multiple unipolar, capacitively coupled electrodes embedded in a regular chair. By processing these signals independently and then synthesizing them using a phase-space-reconstruction-based quality index, the system achieves robust ECG monitoring even with varying body types, clothing, and sitting postures, overcoming limitations of traditional differential measurement methods.
What Problem Does This Paper Address?
Traditional wearable ECG sensors, while useful, can be inconvenient for continuous, long-term monitoring as users may be reluctant to wear them all the time. To address this, researchers are exploring embedding sensors into everyday objects. Chairs, being common seating furniture, present a prime opportunity for unobtrusive health monitoring. However, existing chair-based ECG systems often rely on differential measurements, which are sensitive to variations in electrode-skin coupling impedance caused by factors like body shape, posture, and clothing thickness. This can lead to unstable signal quality and susceptibility to motion artifacts. This paper aims to develop a more robust chair-based ECG acquisition system that mitigates these issues.
Methods
The core innovation lies in the use of multiple unipolar, capacitively coupled electrodes integrated into the backrest of a chair. Unlike traditional differential setups that measure the difference between two electrodes, each unipolar electrode captures the biopotential signal independently. These signals are then amplified and sampled separately.
Key methodological aspects include:
- Capacitively Coupled Electrodes: Standard printed circuit boards (PCBs) are used, with the electrode layer insulated by a solder mask. These electrodes form a capacitor with the human skin, allowing for non-contact measurement through clothing.
- Unipolar Measurement: Each electrode's signal is processed independently, avoiding the impedance imbalance issues inherent in differential measurements.
- Phase Space Reconstruction for Quality Evaluation: A novel quality index is proposed based on phase space reconstruction. This method analyzes the distribution of gridded phase portraits of the ECG signals. A higher "Rate" (defined as the ratio of visited boxes to total boxes in the phase space) indicates lower signal quality, likely due to noise or artifacts.
- Weighted Signal Synthesis: The calculated quality index is used to assign weights to each unipolar channel. A weighted synthesis algorithm then combines these channels into a single, more reliable ECG signal. This approach is computationally efficient, outperforming Principal Component Analysis (PCA) in terms of processing time.
Datasets
The paper utilizes two main sources of data:
- MIT-BIH Database: Segments from this publicly available database were used for simulating and evaluating the proposed signal quality index.
- Experimental Data: ECG signals were collected from human subjects sitting on the developed chair system. This experimental data was used to validate the unipolar measurement approach, the quality evaluation index, and the signal synthesis method under various conditions (different backrest softness, clothing thickness, and subject postures).
Key Insights and Ideas
- Unipolar is Better for Chair-Based ECG: The unipolar approach significantly improves robustness against variations in electrode-skin coupling impedance, which are unavoidable in a chair-based system. This leads to more stable ECG signal acquisition compared to differential methods.
- Phase Space Reconstruction as a Superior Quality Metric: Traditional frequency-domain analysis proved insufficient for evaluating ECG signal quality in this context due to overlapping noise and signal frequencies. Phase space reconstruction offers a more effective way to distinguish clean signals from noisy ones.
- Soft Backrests Enhance Signal Quality: The softness of the chair's backrest plays a crucial role. A softer backrest allows the electrodes to conform better to the body's shape, leading to improved contact and clearer ECG signals, explaining why many existing "smart chairs" utilize soft backrests.
- Efficient Signal Synthesis: The proposed weighted synthesis method, leveraging the phase space quality index, provides a fast and effective way to combine information from multiple unipolar channels into a single, high-quality ECG signal, with significantly lower computational cost than PCA.
Key Results
- The unipolar measurement approach demonstrated robust performance across different sitting postures and body types, with ECG signals clearly observable even when differential measurements were heavily affected by motion artifacts.
- The proposed quality index based on phase space reconstruction effectively differentiated between clean and noisy ECG signals, outperforming frequency-domain methods.
- The weighted synthesis method successfully combined signals from four unipolar electrodes to produce a clear and reliable ECG waveform, suppressing noise and missing features present in individual channels.
- The synthetic method achieved an average execution time of 0.14 ms for 8 seconds of ECG data, significantly faster than PCA (1.58 ms).
- In a performance test with six subjects, the system achieved an overall heartbeat detection sensitivity (Se) of 99.13% and predictability (P) of 99.28%, demonstrating high accuracy in identifying heartbeats.
- The system showed good performance even with clothing thicknesses up to 1.48 mm, highlighting the advantage of capacitively coupled electrodes.
Limitations and Future Work
- P-wave Visibility: The synthetic ECG signals, while good for QRS complex detection, showed less obvious P-wave visibility compared to standard lead-I ECG signals or some individual channels.
- Motion Artifacts in Extreme Cases: While robust, the synthetic results were not entirely stable when motion artifacts were large and occurred across most channels simultaneously (as seen around the 40s mark in Figure 12a). Adjusting the scaling factor 'σ' in the synthesis equation can improve quality in such cases.
- Long-Term Evaluation: The authors suggest that future work should focus on long-term evaluation under real-life conditions and explore new measurement configurations. This would provide a more comprehensive understanding of the system's performance in everyday scenarios.
Summary updated