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.
1.1Movement assessment from observer ratings
What ifWhat if the system uses observer ratings of movement quality and has no motion sensor?
Claim language“capture motion data of a user”
After each trial, an observer enters ratings from zero to four for steadiness, fluidity and coordination on a keypad. A lookup table maps the rating combinations to a summary category, and the display shows the category with the lowest-rated dimension.
1.2Feature extraction from raw samples
What ifWhat if the processor extracts features directly from raw motion samples without filtering them?
Claim language“filter the motion data”
A motion logger supplies raw digital acceleration and angular velocity samples at 100 samples per second. For each two-second window, the processor computes the median absolute deviation of acceleration magnitude and the median angular speed directly from the samples. The classifier uses the joint ranges of these statistics to assign a stability category.
1.2Continuous samplewise movement scoring
What ifWhat if the processor updates the movement score with each sample without dividing the stream into windows?
Claim language“segment the motion data into time-based windows”
The processor maintains running means, variances and cross-covariance for acceleration magnitude and angular speed using exponential weights. A classifier updates the coordination score from the resulting correlation and variability estimates after every sample.
1.3Smoothness from one thresholded feature
What ifWhat if the interpretation module assigns a smoothness category from one scalar feature without comparing it with other features?
Claim language“based on relationships among the extracted motion features”
The processor calculates root-mean-square jerk magnitude from successive acceleration samples in each two-second interval. Two fixed thresholds divide that value into low, intermediate and high abruptness categories.
1.4Scores for internal recordkeeping
What ifWhat if the system records movement quality classifications internally and provides no feedback?
Claim language“generate responsive feedback”
The processor writes each classification and its timestamp to nonvolatile memory. An internal routine computes weekly counts by category and stores them as session summaries.
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
Movement feedback for ceramic shaping
Two overhead cameras track markers on a ceramic worker’s fingertips and wrists while the worker shapes clay on a wheel. A processor reconstructs the three-dimensional marker positions, filters the trajectories, and extracts fingertip path curvature and wrist displacement in time windows. It uses the ratio of fingertip curvature to wrist displacement and the distribution of that ratio across fingers to classify fine movement control across shaping motions. A display updates the control indicator beside the wheel.
Component substitution
Joint motion from optical fiber strain
Fiber Bragg gratings replace the inertial sensors in a flexible sleeve, with opposing pairs of gratings arranged across the elbow to measure bending in two planes. An optical interrogator reads the reflected wavelengths, and the processor subtracts each pair’s wavelength shifts and filters the resulting bend estimates. In each time window, it compares bend excursion with the path length traced by the two bend signals to classify steadiness during arbitrary joint movements. A display presents the steadiness result for each window.
Parameter or scale variation
Short windows for fine hand movement
A six-axis motion sensor on the back of the hand samples at 400 Hz, and a processor applies a 40 Hz low-pass filter before dividing the signals into 250 millisecond windows that advance every 50 milliseconds. The processor extracts angular excursion and integrated absolute angular acceleration from each 100-sample window. It scales the excursion by the integrated acceleration and window duration to classify smoothness across fine hand movements. An on-screen indicator refreshes every 50 milliseconds.
Combination with existing systems
Motion feedback on a rehabilitation treadmill
Shoe-mounted motion sensors feed a movement processor connected to a pressure-sensing rehabilitation treadmill, and the two systems timestamp measurements against a shared clock. The processor filters the shoe trajectories and the center-of-pressure trace, then extracts lateral excursion from both sources in the same time windows. It compares pressure excursion with total shoe excursion and compares the two shoes’ shares of that excursion to classify lateral control across treadmill movements. The treadmill display shows the classification beside belt speed.
Automated or data-driven variant
Learning feature relationships from paired ratings
A depth camera records joint trajectories, and a processor filters and segments them into time windows. The processor calculates ratios of distal to proximal path length and net displacement to total path length for each limb. A regularized ranking model learns weights for these ratios from human ratings of which of two clips shows steadier movement, with training pairs drawn across different movement patterns. Software applies the learned weights to incoming windows and updates a steadiness indicator.
Feedback weighted by tracking uncertainty
Three synchronized cameras feed a processor that reconstructs and filters limb trajectories and estimates position uncertainty from the disagreement among camera views. In each time window, the processor propagates that uncertainty into two feature ratios, one comparing net displacement with path length and one comparing peak lateral excursion with total excursion. It combines the ratios using inverse-variance weights to classify path control across movement patterns. The display shows a tracking status indicator when the estimated uncertainty in the combined result exceeds a configured threshold.
Specific version of a part
Shank-mounted inertial sensor
An elastic cuff holds a three-axis accelerometer and a three-axis gyroscope against the lateral shank. A microcontroller samples both sensors at 100 hertz and timestamps the six channels with a shared clock.
Overlapping windows with frequency features
The processor applies a fourth-order low-pass filter with a 10-hertz cutoff to each axis sampled at 100 hertz. It divides the stream into four-second windows with 50 percent overlap and computes the dominant frequency and the fraction of spectral energy near that frequency for each window.
Coordination from intersegment timing
The interpretation module cross-correlates shank and thigh angular velocity traces within each window and retains the peak correlation and its lag. A two-dimensional lookup table maps those values to coordination categories.
Vibration after sustained instability
A microcontroller drives a wrist-mounted vibration motor for 150 milliseconds when three consecutive windows receive a low-stability classification. A five-second lockout sets the minimum time between pulses.
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.
1.1a sensing system configured to capture motion data of a user;
(Although the set-top box is illustrated as being distinct from the image capture device 108, it may have the image capture device 108 integrated therein.) As the user 301 performs the exercise, or other athletic movement, the image capture device 108 captures images (e.g., motion capture) of the anatomical structure of the user 301 during the biomechanical movement.
A · US 2021/0154529 A1 · ¶ 51
Inertial measurement units (IMUs) can be attached to the legs of a patient which can report back kinematic data describing the angular velocity of each limb.
B · US 2022/0241591 A1 · ¶ 35
1.2a processing system configured to receive the motion data, filter the motion data, segment the motion data into time-based windows and extract one or more motion features from the motion data;
1.3an interpretation module configured to classify movement quality of the user based on relationships among the extracted motion features; and
For example, a particular position of an anatomical structure (e.g., limb) outside of a predefined range of motion may itself not be enough to violate a biomechanical rule, but a movement of that anatomical structure outside of that predefined range at an acceleration that exceeds a predetermined permissible acceleration may be violate the one or more biomechanical rules.
A · US 2021/0154529 A1 · ¶ 36
1.4an output system configured to generate responsive feedback based on the movement quality classification;
The processor generates a virtual sports feedback cue to be rendered, in at least substantially real-time, via a virtual sports performance expert during performance of the one or more biomechanical movements of the user.
A · US 2021/0154529 A1 · abstract
1.5wherein the movement quality classification is performed independent of predefined motion templates, predefined exercise classifications or activity-specific models and is applicable across arbitrary movement patterns.
An empty cell means only that no passage was reproduced there.
Statements in these documents
In essence, the sports performance expert performs a subjective assessment, which may not take into account invisible metrics (e.g., fatigue) or metrics that are difficult for a human to capture (e.g., joint articulation angles) with any degree of precision and/or consistency.
A · US 2021/0154529 A1 · ¶ 2
As an alternative, a computer program may have a computer readable storage device with a computer readable program stored thereon that implements the functionality of the aforementioned interactive visualization systems.
A · US 2021/0154529 A1 · ¶ 7
The present invention generally relates to using kinematic information about a patient to modify deep brain stimulation parameters to maintain therapeutic benefit.
B · US 2022/0241591 A1 · ¶ 2
FoG is sometimes non-responsive to medication which can otherwise alleviate other PD symptoms.
B · US 2022/0241591 A1 · ¶ 4
Cite aspatent.fyi, US 2026/0354707 A1, claim 1, first published 2026-11-19T02:47:19-05:00, https://patent.fyi/US20260354707/, SHA-256 c49f97f1f244d479546004234d721162130b16fde31d0fb04413ff4cafdd1e43