1.1Vehicle telemetry as input
The device reads wheel speed and forward radar range at 20 Hz and calculates time headway from the range and vehicle speed. It uses the headway and range rate to select a following-distance advisory.
Generated automatically. Not reviewed by a person. Not legal advice or an opinion.
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.
The device reads wheel speed and forward radar range at 20 Hz and calculates time headway from the range and vehicle speed. It uses the headway and range rate to select a following-distance advisory.
An onboard card reader loads two model files and a manifest that lists their input dimensions and checksums. The device verifies both files and copies them to internal flash memory for inference.
The device calculates the fraction of valid camera frames with closed eyelids over a rolling 30-second window. A threshold table maps that fraction to an advisory code, and the device passes the code to the correction stage.
The device checks how long the signal behind the advisory has remained above a configured threshold. It changes the advisory to a pending state until the signal has stayed above the threshold for three continuous seconds.
The device sends a timestamp and a corrected advisory code over a local data link. When the controller receives a confirmed fatigue advisory, it queues a low-workload scenario for the next training segment.
Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.
A crane simulator workstation acquires gaze position, eyelid closure, and pulse intervals from an eye tracker and a wrist sensor. The workstation downloads a trained advisory model and a trained correction model from a server that stores both models in a database. The advisory model uses the sensor features to select a visual reminder, and the correction model uses that reminder and the same features to set its display duration and placement. The workstation displays the corrected reminder on the simulator screen.
A fuzzy inference engine replaces the trained correction model on the user device. The server supplies the first-stage model and a database-backed configuration that defines triangular membership functions for eyelid closure fraction and median blink duration. The device selects a rule table using the first-stage advisory type, evaluates its rules against the current features, and calculates corrected urgency as a weighted average of the rule outputs. The display maps the corrected urgency to one of three intensity levels.
A user device samples eyelid position at 60 frames per second and pulse waveforms at 100 samples per second. Every 250 milliseconds, it calculates eyelid closure fraction over the preceding 2 seconds and median pulse interval over the preceding 30 seconds. A first-stage model downloaded from the server calculates an advisory score from both features, and a second downloaded model uses the score and both features to select the displayed advisory level. The device starts inference after it collects the first 30 seconds of data.
A driver monitoring device connects to an existing lane departure warning controller over an in-vehicle data bus. The device runs a server-supplied advisory model on gaze and eyelid features, then passes its output and those features to a server-supplied correction model that selects an advisory display priority. A display scheduler reads the controller's warning-active flag and presents the corrected advisory after the lane departure warning clears. The scheduler discards queued advisories older than 3 seconds.
A simulator logs contextual features, first-stage advisory scores, and driver responses in a server database. The simulator marks an advisory as actionable when the driver completes the associated control input within 5 seconds of the prompt, using only scenarios in which that input is called for. A scheduled server job fits a logistic regression correction model using the logged features and scores as inputs and the response marks as labels. The user device downloads both models and displays a prompt when the corrected probability exceeds a configured threshold.
A user device aligns eye tracking and pulse measurements by timestamp and calculates a valid-sample fraction for each channel over a rolling 5-second window. It fills missing feature values with training-set medians supplied by the server and applies the downloaded first-stage model to estimate the need for an advisory. A downloaded correction model receives that estimate, the filled features, and the valid-sample fractions and selects an advisory message, a sensor adjustment prompt, or no message. The device presents the selected output on its display.
A driver-facing near-infrared camera captures eye images at 30 frames per second. The user device flags blinks and frames with low tracking confidence, then calculates the fraction of usable non-blink frames with at least 80 percent eyelid closure over a 60-second window.
The user device downloads two trained models over an encrypted wireless link. A shared manifest specifies the feature order, input dimensions, and version of each model, and the device activates the pair after checking both files.
The device feeds the initial drowsiness score, eye-tracking confidence, and recent eyelid closure fraction into a trained logistic regression model. The model produces a corrected score that the device converts to a prompt level using two configured thresholds.
The user device maps the corrected advisory code to a short message on the instrument display. A confirmed fatigue advisory displays a prompt to take a break and remains visible until the user acknowledges it.
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.
Multifunction Biotelemetry Support System for Psychophysiology Monitoring of psychophysiological state in real time, measuring physiological values which are further evaluated uses at least one sensory unit equipped with at least 1 translation unit for 2 WiFi access points and one 3 control unit.B · US 2016/0183797 A1 · ¶ 25
The first module determines a risk factor based on a recommendation associated with a vehicle and contextual data associated with the vehicle.A · US 2015/0325121 A1 · ¶ 5
An empty cell means only that no passage was reproduced there.
Accordingly, it is desirable to provide methods and systems for providing decision support with the recommendations and presenting the recommendations and the decision support to a user of the vehicle.A · US 2015/0325121 A1 · ¶ 3
In one embodiment, a method includes: receiving a recommendation associated with a vehicle; receiving contextual data associated with the vehicle; determining a risk factor based on the recommendation and the contextual data; and generating notification data based on the risk factor to notify a user of the vehicle of the risk factor associated with the recommendation.A · US 2015/0325121 A1 · ¶ 4
Current solutions in the commonly available setup similarly do not allow for real-time evaluation or classification of psychophysiological states online as the control software, which is primarily designed for diagnostics lacks appropriate design for such substantially different tasks as real-time evaluation.B · US 2016/0183797 A1 · ¶ 4
The above disadvantages are overcome to a great extent by the Multifunction Biotelemetry Support System for Psychophysiology Monitoring measuring and further evaluating physiology variables in real time.B · US 2016/0183797 A1 · ¶ 8
patent.fyi, US 2026/0356384 A1, claim 1, first published 2026-11-20T00:27:46-05:00, https://patent.fyi/US20260356384/, SHA-256 08c775024110109e9e0d91617bb54ce5a23ad555446635f015a4282c93e10b40