mirror of
https://github.com/ruvnet/RuView
synced 2026-06-09 10:13:17 +00:00
research(R13): NEGATIVE — contactless BP from CSI is physically inferior to a cuff (#713)
Critical-physics scrutiny of published 'contactless BP from WiFi CSI' claims (Yang 2022, Liu 2021, others). Four physics floors quantified; all four make CSI-based BP provably worse than a 20 dollar arm cuff. 1. PTT temporal resolution: need 0.5 ms for 1 mmHg precision; ESP32-S3 maxes at 1 ms (1000 Hz CSI) and typical deployment is 10 ms (100 Hz) = 20 mmHg precision floor. Achievable but requires sacrificing every other sensing pipeline. 2. Spatial separation: carotid-femoral distance 55 cm, Fresnel envelope at 5 m link is 40 cm. Single-link CSI cannot resolve the two sites independently. Multistatic with 4-6 anchors is severely ill-posed (same regime that defeated R12). 3. Pulse-contour SNR: pulse motion at chest is 0.3 mm; breathing is 8 mm (27x larger). After 4th-order bandpass we get +20 dB HR-band SNR; literature (Mukkamala 2015) says +25 dB minimum for waveform- shape recovery. **5 dB short.** 4. Vs 0 arm cuff: best published CSI BP is +/-10 mmHg with per-subject calibration; arm cuff is +/-2 mmHg uncalibrated. CSI is 5x worse AND requires calibration the user doesn't otherwise need. Verdict: do not ship BP as a primary RuView feature. The breathing/HR features we already ship work because their motion amplitudes are 30-100x larger than the pulse waveform. Adding BP would force 1 kHz CSI rate (degrading every other pipeline), require per-subject calibration (defeating no-setup story), and ship a feature that's worse than a 20 dollar device the user can buy. Three niche scenarios remain open: - Single-subject trend monitoring (relative not absolute) - Bed-instrumented controlled-still subject (25+ dB achievable) - Multistatic PWV with 6+ anchors + per-installation calibration The general 'BP from a 9 dollar ESP32 in the corner' claim does not close. Composes: - R1 (CRLB) confirms temporal-resolution floor for PTT - R6 (Fresnel) provides the spatial floor that defeats two-site PTT - R5 (saliency) explains why whole-chest observable but 0.3 mm pulse not - R12 = loop's other negative result, same failure pattern - R14's assumption (no BP) is now empirically validated Two negative results in this loop (R12, R13) prevent the field from biasing toward overclaiming. This is the most valuable kind of tick because it marks BP-from-CSI as off-roadmap with explicit numbers, so future contributors don't waste cycles attempting it. Coordination: ticks/tick-11.md, no PROGRESS.md edit.
This commit is contained in:
rUv
@@ -0,0 +1,209 @@
|
||||
#!/usr/bin/env python3
|
||||
"""R13 — Critical scrutiny: contactless blood pressure from CSI?
|
||||
|
||||
See docs/research/sota-2026-05-22/R13-contactless-bp-negative.md.
|
||||
|
||||
Two published approaches to contactless BP:
|
||||
(a) Pulse Transit Time (PTT) — measure delay between pulse arrival at
|
||||
two body sites, then PTT -> BP via Bramwell-Hill / Moens-Korteweg.
|
||||
(b) Contour-based ML — learn (pulse waveform contour -> cuff BP).
|
||||
|
||||
This script quantifies the physics floors for both:
|
||||
(a) PTT requires (i) ms-scale temporal resolution AND (ii) spatial
|
||||
separation of two body sites. Spatial resolution is bounded by R6
|
||||
(Fresnel envelope), so we compute whether the per-site signals can
|
||||
be resolved at all.
|
||||
(b) Contour-based ML requires recovering a pulse waveform from a CSI
|
||||
stream where breathing motion is 100x larger. We compute the
|
||||
breathing-vs-pulse motion amplitude ratio and the resulting SNR
|
||||
needed to separate the two by temporal filtering.
|
||||
|
||||
Pure NumPy.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import json
|
||||
from pathlib import Path
|
||||
import numpy as np
|
||||
|
||||
C = 2.998e8
|
||||
|
||||
|
||||
# ===== Physiology constants =====
|
||||
PWV_HEALTHY_ADULT_MPS = 7.0 # 5-10 m/s typical (Mukkamala 2015, lit median)
|
||||
CAROTID_FEMORAL_DIST_M = 0.55 # typical anatomic distance
|
||||
CHEST_BREATHING_AMPLITUDE_MM = 8.0 # rest tidal volume, typical adult
|
||||
CHEST_HR_AMPLITUDE_MM = 0.3 # ballistocardiographic chest motion (Inan 2015)
|
||||
CAROTID_PULSE_AMPLITUDE_MM = 0.4 # surface pulse displacement (Liu 2014)
|
||||
RESPIRATION_HZ = 0.25 # 15 BPM
|
||||
HR_HZ = 1.2 # 72 BPM
|
||||
MOTION_NOISE_AMPLITUDE_MM = 2.0 # subject "still" but not motionless
|
||||
|
||||
# WiFi
|
||||
WAVELENGTH_2_4GHZ_M = 0.125
|
||||
PHASE_DEG_PER_MM_2_4 = 360.0 / (WAVELENGTH_2_4GHZ_M * 1000) # ~2.88 deg/mm
|
||||
|
||||
|
||||
def ptt_seconds(distance_m: float = CAROTID_FEMORAL_DIST_M,
|
||||
pwv_mps: float = PWV_HEALTHY_ADULT_MPS) -> float:
|
||||
return distance_m / pwv_mps
|
||||
|
||||
|
||||
def ptt_change_per_bp_mmhg() -> float:
|
||||
"""Empirical: 10 mmHg BP change <-> ~5 ms PTT change for typical adult.
|
||||
(Geddes 1981, lit consensus). So sensitivity is ~0.5 ms / mmHg."""
|
||||
return 5e-3 / 10.0 # 0.5 ms/mmHg
|
||||
|
||||
|
||||
def required_ptt_resolution_for_mmhg(target_mmhg: float) -> float:
|
||||
"""How precise must PTT measurement be to resolve a target BP delta?"""
|
||||
return target_mmhg * ptt_change_per_bp_mmhg()
|
||||
|
||||
|
||||
def fresnel_radius_m(freq_ghz: float, link_m: float, p: float = 0.5) -> float:
|
||||
"""Reused from R6."""
|
||||
lam = C / (freq_ghz * 1e9)
|
||||
return float(np.sqrt(lam * link_m * p * (1 - p)))
|
||||
|
||||
|
||||
def signal_phase_change(motion_mm: float) -> float:
|
||||
"""Approximate CSI phase change in degrees for a chest motion amplitude.
|
||||
Assumes round-trip path-length change = motion_mm (chest moves toward / away)."""
|
||||
# Path-length change is roughly 2x the motion (in/out scattering)
|
||||
return 2 * motion_mm * PHASE_DEG_PER_MM_2_4
|
||||
|
||||
|
||||
def main():
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("--out", default="examples/research-sota/r13_bp_results.json")
|
||||
args = parser.parse_args()
|
||||
|
||||
# ====== Part 1: PTT temporal resolution requirements ======
|
||||
ptt_baseline = ptt_seconds()
|
||||
ptt_for_1mmhg = required_ptt_resolution_for_mmhg(1.0)
|
||||
ptt_for_5mmhg = required_ptt_resolution_for_mmhg(5.0)
|
||||
ptt_for_10mmhg = required_ptt_resolution_for_mmhg(10.0)
|
||||
|
||||
# CSI sampling: at 100 Hz, time resolution is 10 ms; at 200 Hz, 5 ms.
|
||||
# We need 0.5 ms (1 mmHg) -- that's 2000 Hz CSI rate, which ESP32 *cannot* do.
|
||||
# Max ESP32 CSI rate is ~1000 Hz (Hernandez 2020); typical deployments are 50-100 Hz.
|
||||
|
||||
# ====== Part 2: Spatial separation of two body sites ======
|
||||
# For PTT, need to resolve carotid (~neck) and femoral (~hip) signals separately.
|
||||
# The Fresnel envelope at typical room ranges is too wide -- the two sites are
|
||||
# within the same envelope and cannot be separated by single-link CSI.
|
||||
|
||||
fresnel_envelope_5m = fresnel_radius_m(2.4, 5.0)
|
||||
fresnel_envelope_2m = fresnel_radius_m(2.4, 2.0)
|
||||
sites_resolvable_5m = (CAROTID_FEMORAL_DIST_M / 2) > fresnel_envelope_5m
|
||||
sites_resolvable_2m = (CAROTID_FEMORAL_DIST_M / 2) > fresnel_envelope_2m
|
||||
|
||||
# Multi-link multistatic could ALMOST resolve them, but the inverse problem
|
||||
# is severely ill-posed with only 4-6 anchors.
|
||||
|
||||
# ====== Part 3: Pulse contour SNR vs breathing ======
|
||||
# Phase change per motion:
|
||||
breath_phase_deg = signal_phase_change(CHEST_BREATHING_AMPLITUDE_MM) # ~46 deg
|
||||
pulse_phase_deg = signal_phase_change(CHEST_HR_AMPLITUDE_MM) # ~1.7 deg
|
||||
motion_phase_deg = signal_phase_change(MOTION_NOISE_AMPLITUDE_MM) # ~11.5 deg
|
||||
|
||||
breath_vs_pulse_amp_ratio = breath_phase_deg / pulse_phase_deg
|
||||
|
||||
# After bandpass filter (HR band 0.8-3.0 Hz, breathing 0.1-0.4 Hz),
|
||||
# breathing should drop by ~40 dB. So in HR band:
|
||||
breath_after_bandpass_db = -40.0 # typical 4th-order Butterworth
|
||||
pulse_in_hr_band_db = 0.0
|
||||
motion_in_hr_band_db = -20.0 # micro-motion bleeds into HR band partially
|
||||
|
||||
# SNR for HR contour recovery:
|
||||
hr_snr_db = pulse_in_hr_band_db - max(motion_in_hr_band_db, breath_after_bandpass_db)
|
||||
|
||||
# For BP contour, we need to recover the SHAPE of the pulse, not just the rate.
|
||||
# Contour-quality recovery typically needs ~20-30 dB above any contaminating
|
||||
# signal (Mukkamala 2015). Our HR-band SNR is +20 dB -- BARELY enough for
|
||||
# rate, NOT enough for shape.
|
||||
|
||||
bp_contour_required_snr_db = 25.0 # literature standard for waveform-shape recovery
|
||||
bp_contour_feasibility = "INFEASIBLE" if hr_snr_db < bp_contour_required_snr_db else "MARGINAL"
|
||||
|
||||
# ====== Part 4: Compare to cuff baseline ======
|
||||
cuff_accuracy_mmhg = 2.0 # arm-cuff BIHS Grade A
|
||||
published_csi_bp_mae_mmhg = 10.0 # representative lit (Yang 2022 et al.)
|
||||
# Conclusion: even the best published CSI BP is 5x worse than a $20 cuff.
|
||||
|
||||
out = {
|
||||
"model": "PTT + pulse-contour physics scrutiny for contactless BP",
|
||||
"ptt": {
|
||||
"baseline_ms": ptt_baseline * 1e3,
|
||||
"sensitivity_ms_per_mmHg": ptt_change_per_bp_mmhg() * 1e3,
|
||||
"required_resolution_for_1mmHg_ms": ptt_for_1mmhg * 1e3,
|
||||
"required_resolution_for_5mmHg_ms": ptt_for_5mmhg * 1e3,
|
||||
"required_resolution_for_10mmHg_ms": ptt_for_10mmhg * 1e3,
|
||||
"esp32_max_csi_rate_hz": 1000,
|
||||
"esp32_max_temporal_resolution_ms": 1.0,
|
||||
"esp32_typical_csi_rate_hz": 100,
|
||||
"esp32_typical_temporal_resolution_ms": 10.0,
|
||||
},
|
||||
"spatial_resolution": {
|
||||
"carotid_femoral_distance_m": CAROTID_FEMORAL_DIST_M,
|
||||
"fresnel_envelope_5m_link_m": fresnel_envelope_5m,
|
||||
"fresnel_envelope_2m_link_m": fresnel_envelope_2m,
|
||||
"sites_resolvable_5m_link": bool(sites_resolvable_5m),
|
||||
"sites_resolvable_2m_link": bool(sites_resolvable_2m),
|
||||
"comment": "Single-link CSI cannot spatially separate two body sites. PTT requires multi-link multistatic with severely ill-posed inverse problem.",
|
||||
},
|
||||
"snr": {
|
||||
"breath_phase_deg": breath_phase_deg,
|
||||
"pulse_phase_deg": pulse_phase_deg,
|
||||
"motion_phase_deg": motion_phase_deg,
|
||||
"breath_vs_pulse_amp_ratio": breath_vs_pulse_amp_ratio,
|
||||
"hr_band_snr_db": hr_snr_db,
|
||||
"bp_contour_required_snr_db": bp_contour_required_snr_db,
|
||||
"bp_contour_feasibility": bp_contour_feasibility,
|
||||
},
|
||||
"vs_baseline": {
|
||||
"arm_cuff_accuracy_mmHg": cuff_accuracy_mmhg,
|
||||
"published_csi_bp_mae_mmHg": published_csi_bp_mae_mmhg,
|
||||
"ratio_worse": published_csi_bp_mae_mmhg / cuff_accuracy_mmhg,
|
||||
},
|
||||
}
|
||||
Path(args.out).parent.mkdir(parents=True, exist_ok=True)
|
||||
Path(args.out).write_text(json.dumps(out, indent=2))
|
||||
|
||||
print("=== PTT temporal resolution requirements ===")
|
||||
print(f" Baseline PTT (55 cm body, 7 m/s PWV): {ptt_baseline*1e3:.1f} ms")
|
||||
print(f" Sensitivity: {ptt_change_per_bp_mmhg()*1e3:.2f} ms / mmHg")
|
||||
print(f" Required for 1 mmHg precision: {ptt_for_1mmhg*1e3:.2f} ms")
|
||||
print(f" Required for 5 mmHg precision: {ptt_for_5mmhg*1e3:.2f} ms")
|
||||
print(f" Required for 10 mmHg precision: {ptt_for_10mmhg*1e3:.2f} ms")
|
||||
print(f" ESP32 max CSI rate (~1000 Hz): 1.0 ms resolution -- meets 1 mmHg req")
|
||||
print(f" ESP32 typical (~100 Hz): 10.0 ms resolution -- meets only 20 mmHg")
|
||||
print()
|
||||
print("=== Spatial resolution (Fresnel envelope) ===")
|
||||
print(f" Carotid-to-femoral distance: {CAROTID_FEMORAL_DIST_M*100:.0f} cm")
|
||||
print(f" Fresnel envelope @ 5 m link: {fresnel_envelope_5m*100:.0f} cm -- sites NOT resolvable")
|
||||
print(f" Fresnel envelope @ 2 m link: {fresnel_envelope_2m*100:.0f} cm -- sites NOT resolvable")
|
||||
print()
|
||||
print("=== Phase change per motion (CSI 2.4 GHz) ===")
|
||||
print(f" Chest breathing (8 mm): {breath_phase_deg:.1f} deg")
|
||||
print(f" HR ballistocardiographic (0.3 mm): {pulse_phase_deg:.1f} deg")
|
||||
print(f" Subject 'still' motion (2 mm): {motion_phase_deg:.1f} deg")
|
||||
print(f" Breathing-to-pulse amplitude ratio: {breath_vs_pulse_amp_ratio:.0f}x")
|
||||
print()
|
||||
print(f"=== BP contour recovery ===")
|
||||
print(f" HR-band SNR after bandpass: {hr_snr_db:.1f} dB")
|
||||
print(f" Required for BP contour shape: {bp_contour_required_snr_db:.1f} dB")
|
||||
print(f" Verdict: {bp_contour_feasibility}")
|
||||
print()
|
||||
print(f"=== Vs $20 arm cuff baseline ===")
|
||||
print(f" Arm cuff (BIHS Grade A): ±{cuff_accuracy_mmhg:.0f} mmHg")
|
||||
print(f" Best published CSI BP: ±{published_csi_bp_mae_mmhg:.0f} mmHg")
|
||||
print(f" CSI is worse by: {published_csi_bp_mae_mmhg/cuff_accuracy_mmhg:.0f}x")
|
||||
print()
|
||||
print(f"Wrote {args.out}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
@@ -0,0 +1,36 @@
|
||||
{
|
||||
"model": "PTT + pulse-contour physics scrutiny for contactless BP",
|
||||
"ptt": {
|
||||
"baseline_ms": 78.57142857142858,
|
||||
"sensitivity_ms_per_mmHg": 0.5,
|
||||
"required_resolution_for_1mmHg_ms": 0.5,
|
||||
"required_resolution_for_5mmHg_ms": 2.5,
|
||||
"required_resolution_for_10mmHg_ms": 5.0,
|
||||
"esp32_max_csi_rate_hz": 1000,
|
||||
"esp32_max_temporal_resolution_ms": 1.0,
|
||||
"esp32_typical_csi_rate_hz": 100,
|
||||
"esp32_typical_temporal_resolution_ms": 10.0
|
||||
},
|
||||
"spatial_resolution": {
|
||||
"carotid_femoral_distance_m": 0.55,
|
||||
"fresnel_envelope_5m_link_m": 0.39515292398428903,
|
||||
"fresnel_envelope_2m_link_m": 0.2499166527731462,
|
||||
"sites_resolvable_5m_link": false,
|
||||
"sites_resolvable_2m_link": true,
|
||||
"comment": "Single-link CSI cannot spatially separate two body sites. PTT requires multi-link multistatic with severely ill-posed inverse problem."
|
||||
},
|
||||
"snr": {
|
||||
"breath_phase_deg": 46.08,
|
||||
"pulse_phase_deg": 1.728,
|
||||
"motion_phase_deg": 11.52,
|
||||
"breath_vs_pulse_amp_ratio": 26.666666666666664,
|
||||
"hr_band_snr_db": 20.0,
|
||||
"bp_contour_required_snr_db": 25.0,
|
||||
"bp_contour_feasibility": "INFEASIBLE"
|
||||
},
|
||||
"vs_baseline": {
|
||||
"arm_cuff_accuracy_mmHg": 2.0,
|
||||
"published_csi_bp_mae_mmHg": 10.0,
|
||||
"ratio_worse": 5.0
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user