patent.fyi
Applications  /  US 2026/0354708
United States Patent Application Publication

US 2026/0354708 A1

Published
Nov 12, 2026
Filed
Jun 24, 2026
Earliest date on the face
Dec 25, 2023
Classification
A61B 5/1126
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Chin motion sensing

What ifWhat if the sensor sat at the center of the chin?
Claim language“mounted on a lateral side of a lower jaw”

A three-axis accelerometer on an adhesive patch measures movement at the chin. A band-pass filter isolates repeated jaw motion, and a classifier uses the timing and amplitude of successive strokes to identify chewing.

1.2Local peak timing

What ifWhat if local amplitude peaks marked the chewing strokes?
Claim language“detection of a rising zero-cross”

A low-pass filter smooths the jaw sensor signal, and a five-sample window locates peaks by comparing its center sample with the samples on either side. The detector checks peak prominence and the intervals between successive peaks to identify chewing bouts.

1.2Single boundary after smoothing

What ifWhat if a filtered signal used the same zero boundary on both sides of a crossing?
Claim language“with hysteresis”

A five-sample moving average suppresses short fluctuations before the detector tests for a transition from a negative sample to a nonnegative sample. A crossing starts an amplitude measurement window, and the detector uses the resulting stroke amplitudes and timing to identify chewing.

1.2Waveform based detector reset

What ifWhat if the next candidate became eligible only after the waveform completed a return excursion?
Claim language“based on a dead time setting”

After a rising crossing, a state machine waits for the signal to fall below a negative reset level before accepting another crossing. The measured waveform determines when the detector rearms, and the detector uses the amplitude and spacing of the resulting strokes to identify chewing.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Chew counting in food texture panels

A lateral lower-jaw motion sensor connects to a trial logger at each food texture assessment station. The operator starts a recording window as the participant begins chewing a coded sample and closes it when the participant completes the trial. The detector uses hysteresis and dead time to identify rising zero crossings, then compares the ensuing signal excursion with a mastication threshold. The logger associates the chew count and event times with the sample code.

Component substitution

Angular velocity sensing at the jaw

A microelectromechanical gyroscope serves as the sensor in a housing secured to the lateral side of the lower jaw. One sensing axis aligns with the approximate axis of jaw opening and closing, and the detector subtracts a stationary bias measured at setup. The detector uses hysteresis and dead time to identify rising zero crossings in the signed angular velocity. After each accepted crossing, it compares the positive angular velocity peak with a mastication threshold.

Analog circuitry for chew event detection

A lateral lower-jaw motion sensor drives a comparator with positive feedback and switching levels on either side of the sensor's resting voltage. Each rising comparator edge accepted by a nonretriggerable one-shot starts a dead-time interval and resets a gated peak-hold circuit. A shorter timing pulse opens the peak-hold acquisition window, and a second comparator checks the held voltage against an adjustable mastication threshold at the end of that window. An output latch emits one count pulse when the held voltage reaches the threshold.

Parameter or scale variation

Timed detection for rapid chewing

A lateral lower-jaw accelerometer supplies a signed acceleration channel at 200 samples per second. The detector arms at or below −0.03 g and confirms a rising zero crossing when the signal subsequently reaches +0.03 g. It starts a 220 ms dead time at confirmation and compares the highest sample in the following 100 ms with a 0.10 g mastication threshold. The detector counts the event when the peak reaches the threshold and requires a fresh negative excursion after dead time before arming again.

Combination with existing systems

Chew markers for swallowing video

A lateral lower-jaw sensor and detector connect to a clinical swallowing video recorder through a digital event input. The detector buffers the rising zero-crossing time when hysteresis and dead time permit a candidate, then tests the subsequent signal peak against the mastication threshold. For each confirmed chew, the recorder stores the buffered crossing time against the corresponding frame using a shared clock. The review interface displays the markers on the video timeline.

Automated or data-driven variant

Automatic calibration from guided chewing

A lateral lower-jaw sensor streams a signed waveform to software that records five seconds of rest and ten prompted chews. The software centers the signal on its resting median and sets symmetric hysteresis levels from the 95th percentile of absolute resting deviations, with a floor of two sensor quantization steps. It sets dead time to 40 percent of the median interval between calibration crossings and sets the mastication threshold to 60 percent of the median positive peak following those crossings. During operation, the software applies the calibrated hysteresis and dead time before comparing each candidate peak with the calibrated threshold.

Specific version of a part

Contactless jaw tracking

A camera on a tabletop stand records the chin and nose at 60 frames per second. A tracker subtracts nose motion from chin motion to reduce the contribution of head movement, then measures the duration and spacing of successive jaw strokes.

Waveform template matching

Each rising crossing opens a 600-millisecond sampling window, and the processor normalizes the sampled waveform to unit amplitude. The processor computes dynamic time warping distances to stored chewing, speech, and jaw-opening templates and assigns the label of the closest template. A series of consecutive chewing labels identifies a chewing bout.

Adhesive jaw accelerometer

A three-axis accelerometer sits in a silicone housing held against the skin over the lateral mandibular body by a medical adhesive pad. One axis runs approximately along the lower edge of the jaw, and the sensor samples all three axes at 100 samples per second.

Jaw mounted strain film

A flexible piezoelectric film sits in an adhesive strip across the skin over the lateral mandibular angle. Jaw movement bends the film, and a charge amplifier converts the resulting electrical charge into a voltage signal.

Crossing detector with blanking

Two comparators monitor a baseline-centered signal with switching levels at −20 and +20 millivolts. A latch registers a rising zero-crossing after the signal passes both levels in order, and a 150-millisecond timer blocks further events. A peak detector measures the next 120 milliseconds of the sensor signal and compares the peak with a calibrated amplitude threshold to identify a chewing stroke.

Digital crossing and amplitude gate

A processor samples the sensor at 200 samples per second, subtracts a running baseline, and uses switching levels of −0.02 g and +0.02 g to register rising zero-crossings with hysteresis. After each registered crossing, it blocks new crossings for 40 samples and compares the maximum signal amplitude over the first 30 samples with a calibrated chewing threshold. Three successive strokes above the threshold with intervals of 0.3 to 1.5 seconds indicate a chewing bout.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 2012/0238908 A1
Published
Sep 20, 2012
Filed
Mar 9, 2012
Class
A61B 5/11
Document BUS 2022/0341961 A1
Published
Oct 27, 2022
Filed
Jan 24, 2022
Class
G01P 3/36
Claim 1
Document A
Document B
1.1a sensor to be mounted on a lateral side of a lower jaw; and
For example, a mastication detection device is proposed where a mastication action is detected by attaching a sensor which detects the movement above the temporomandibular joint.
A · US 2012/0238908 A1 · ¶ 8
1.2a mastication detection unit that performs detection of a rising zero-cross on a detection signal from the sensor with hysteresis and based on a dead time setting, and when the rising zero-cross is detected, performs comparison of the detection signal from the sensor with a threshold for detecting a mastication motion and detects, based on a result of the comparison, whether a series of motions made by a person wearing the sensor is a mastication motion.
Then, a portion in the form Pas where a power which is larger than the power threshold Pth′ is held for a period between the lower limit time threshold Tthl and the upper limit time threshold Tthh is determined as the mastication portion, and for example, the detection pulse is output.
A · US 2012/0238908 A1 · ¶ 61
The comparison circuit has multiple hysteresis comparators each being configured to compare a couple of phase shifted signals among the multiple phase shifted signals using an operation hysteresis level.
B · US 2022/0341961 A1 · ¶ 12

An empty cell means only that no passage was reproduced there.

Statements in these documents
Accordingly, it is necessary that mastication is performed with awareness in order to obtain a sufficient number of mastications, but realization of this is difficult.
A · US 2012/0238908 A1 · ¶ 8
However, in a case where a unique sensor such as this is used, there is a problem in that costs increase.
A · US 2012/0238908 A1 · ¶ 8
It is seen from FIG. 4 that if emission light intensity of the light source 10 is adjusted corresponding to different longitudinal distances (referring to FIG. 3B), the pulse width of the index output can have a significant variation.
B · US 2022/0341961 A1 · ¶ 7
Accordingly, the present disclosure further provides an optical encoder that adjusts emission light intensity of a light source corresponding to different longitudinal distances in conjunction with keeping a pulse width of the index output to be substantially identical by regulating a gain of index output.
B · US 2022/0341961 A1 · ¶ 8
Cite aspatent.fyi, US 2026/0354708 A1, claim 1, first published 2026-11-19T02:47:19-05:00, https://patent.fyi/US20260354708/, SHA-256 bc5d55baa8e94d8e3767af731aff37210a82a71e295977b3c70058c90a4f7a10
First published
Permanent addresshttps://patent.fyi/US20260354708/
Content fingerprint (SHA-256)bc5d55baa8e94d8e3767af731aff37210a82a71e295977b3c70058c90a4f7a10
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi