PulseSensor Research · Knowledge Base v1
Papers, patent disclosures, products, implementations, reproducible experiments, open questions, and roadmap projects belong in one connected view—not in separate silos.
Pilot topic · PPG-derived respiratory rate
This pilot is a navigable example, not a claim of commercial lineage, scientific equivalence, patent coverage, or freedom to operate.
Stevens et al. tested four PulseSensors and open-source software with 15 healthy adults in four positions. The best reported respiratory-rate combination reached Spearman ρ = 0.82 and mean difference 1.90 breaths/minute.
Limit: small, single-center cohort; movement and population generalization remain open.
Read the CC BY paper →Active screen: Owlet’s US 12,159,717 discloses a CNN → FFT → dense-layer respiratory-rate predictor.
Expired screen: US 7,881,790 discloses cycle-synchronized sparse sampling.
Abandoned screen: US 2013/0172759 discloses PPG monitoring in a narcotic-reversal system.
Compare disclosures and status →The current PulseSensor is a single-channel green-light reflectance PPG sensor with raw analog output. Stevens et al. selected it for access, wavelength, size, and cost.
Boundary: it does not measure SpO2 or blood pressure and is not medical equipment.
See the current kit →Record synchronized raw PPG and a respiratory reference. Predefine windows, rejection rules, and metrics. Compare baseline wander, pulse-amplitude variation, and beat-interval variation.
Safety: no breath-holding or forced hyperventilation. Without a reference, report modulation—not respiratory-rate accuracy.
Open the experiment →Publish de-identified raw PPG, synchronized reference signals, participant-level train/test splits, signal-quality labels, code, and failure cases.
Decision gate: compare classical modulation fusion with learned models only after the reference and dataset protocol are fixed.
See the build plan →The three lanes · preserved and connected
The patent layer crosses all three lanes; it does not replace or isolate any of them.
Forty-seven papers on PPG foundations, respiratory modulation, algorithms, motion, signal quality, and validation.
Graph role: scientific evidence and methods.
Read the bibliography → 2Verified published uses, the PulseSensor signal chain, open-source tools, limitations, and precedent.
Graph role: products, implementations, and observed use.
See the research record → 3Reproducible starter protocols, advanced questions, and practical next experiments for classrooms, homes, and labs.
Graph role: replication, open questions, and future projects.
Choose a project →How to read the graph
Each connection should expose its evidence type and uncertainty instead of letting proximity imply proof.
The paper reports four PulseSensors, Arduino Nano 33 IoT, 15 participants, four positions, and 15 respiratory-rate algorithm combinations.
Its open hardware/software choices and stated methods make independent replication feasible. That is a planning inference, not proof that every setup will reproduce its results.
The pilot’s patents are related by topic. No claim chart, infringement analysis, validity opinion, or freedom-to-operate conclusion has been made.
Pilot patent reading notes
These are concise technical readings of public patent documents, checked against the local source pack.
Inventors Sean Kerman, Tanner Christensen, Chris Hettinger, and Jeffrey Humpherys disclose a learned pipeline that uses convolutional layers to identify/clean the PPG signal, an FFT layer to represent respiratory frequencies, and a dense layer to predict respiratory rate. The claims also describe derivative, clipping, normalization, and prior-rate template options.
Evidence limit: the document describes and claims a system; it does not replace independent validation across devices, populations, and motion conditions.
Inventor Robert G. Turcott discloses reducing acquisition and processing by taking one hemodynamic sample per respiratory cycle at a consistent phase, optionally triggered by inspiration or expiration. The concept is historically useful for thinking about low-power, cycle-synchronized measurement.
Legal limit: the US record was screened as expired fee-related on 2026-08-01. Before building around a specific claim, check restoration, related family members, other jurisdictions, and any separate rights.
Inventors Richard J. Melker and Donn M. Dennis disclose central-site PPG monitoring in a system that can trigger administration of a narcotic reversal agent when respiratory or other parameters cross preset bounds.
Evidence and safety limit: this is a high-stakes medical disclosure, not a PulseSensor project recommendation, a clinical validation result, or evidence that an adjacent use is legally clear. The application was screened abandoned; related rights require separate checking.
US 10,206,612 B2: separates venous and arterial components from red/infrared PPG using second-order blind identification, then derives respiratory information and SpO2 from different components. US 11,304,654 B2: derives a respiratory signal, peak-to-valley values, and peak counts to detect abnormal respiration. US 9,066,680 B1: compares respiratory-rate estimates from multiple physiological inputs to express confidence.
These are distinct technical disclosures with distinct sensor assumptions. None independently validates a single-channel green PulseSensor implementation.
Reproducible experiment layer
Use a synchronized respiratory reference if the question is accuracy. Without one, the safe question is whether breathing-related modulation is visible and repeatable.
Roadmap project · research classification
This is a future research project, not a current product capability.
Research-grade sourcing
Mutable status is dated. Links point to public records, published research, or existing PulseSensor pages.