patent.fyi
Applications  /  US 2026/0356373
United States Patent Application Publication

US 2026/0356373 A1

Published
Nov 12, 2026
Filed
Apr 8, 2026
Earliest date on the face
Apr 11, 2025
Classification
B60W 50/00
First published here
Nov 20, 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.

1.1Nonlinear feedback through a table

What ifWhat if a nonlinear lookup table maps the state vector to the input vector?
Claim language“a feedback matrix (K)”

The controller locates the measured lateral offset and heading error on two table axes and interpolates among the four surrounding steering commands. The table uses shallow slopes near the centerline and caps the steering command at large offsets.

1.2Sampled dynamics from test data

What ifWhat if the evaluator fits a discrete transition model directly to sampled vehicle data?
Claim language“time discretizing the time-continuous dynamic model”

A test rig records steering commands, command arrival times, lateral offset, and yaw rate every 10 milliseconds. A regression fits the next sampled state from the current state, the last five issued commands, and the measured age of the command at the actuator.

1.3Direct simulation of sampled dynamics

What ifWhat if the evaluator runs the full discrete state update for each sampled delay history?
Claim language“a surrogate model”

The evaluator generates 2,000 delay histories from a fitted stochastic delay model and advances the full vehicle state for each history. A timestamped command queue supplies the steering input at each update, and the resulting trajectories provide state means and variances at each sample time.

1.4State percentiles at each sample

What ifWhat if the evaluator computes lower and upper state percentiles at each sample time?
Claim language“a mean (E(xk))”

The evaluator propagates delay realizations through the surrogate and sorts the resulting lateral offsets separately at each sample time. It compares the 1st and 99th percentiles with the left and right lateral offset limits.

1.5Average exceedance across the maneuver

What ifWhat if the evaluator compares a maneuver-wide exceedance score with a single threshold?
Claim language“searching for one or more discrete time points”

At each sample, the evaluator computes the positive excess of the absolute mean lateral offset plus two standard deviations over a 0.30-meter limit. It averages the excess across the maneuver and flags the timing condition when the average exceeds 0.02 meters.

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

Timing assessment for chilled water plants

A chilled water plant uses supply temperature and pump differential pressure as states and pump speed and valve position as feedback inputs. A continuous thermal and hydraulic model represents command delay as a stochastic process, and its discrete representation stores recent commands to supply delayed actuator inputs. A reduced state-space model propagates the mean and variance of supply temperature at each time point. The software classifies timing as critical if the expected squared temperature tracking error, calculated as squared mean error plus variance, exceeds a specified limit at any time point.

Component substitution

Finite state model for command delay

A finite state Markov jump linear model serves as the surrogate for a linearized and discretized vehicle model with delayed commands. Each mode selects a command from a different position in a history buffer, and a transition matrix describes changes between delay modes. The solver propagates mode probabilities and mode-weighted first and second state moments, then sums those moments to obtain the overall mean and covariance at each step. The software classifies timing as critical when the mean squared steering tracking error exceeds its limit at any step.

Parameter or scale variation

Two rate sampling for suspension assessment

An active suspension model uses body displacement and suspension travel as states and damper commands as feedback inputs. The discretizer runs at a 2 millisecond step over a 200 millisecond horizon, and the controller updates its command every 10 milliseconds. The actuator receives a command after a stochastic delay of zero to five model steps and holds it until a newer command arrives. The surrogate computes suspension travel mean and variance at all 100 steps and classifies timing as critical if the sum of squared mean travel and travel variance exceeds the square of 30 millimeters at any step.

Combination with existing systems

Timing checks on a controller test bench

A hardware-in-the-loop bench connects a production vehicle controller to a real-time plant simulator through a message interface that timestamps command transmission and receipt. The assessor estimates delay transition probabilities from the timestamp records and feeds them into a discrete surrogate of the simulated plant with the controller feedback gains. The surrogate propagates the mean and variance of yaw rate tracking error for each scripted maneuver. The bench records a critical result with the maneuver identifier and first failing time step when the expected squared tracking error crosses its threshold.

Automated or data-driven variant

Delay estimation from vehicle message logs

A logger records command sequence numbers, transmission and receipt timestamps from synchronized clocks, and message queue depth. A fitting routine pairs timestamps by sequence number, bins the measured delays by model time step, and estimates a separate delay transition matrix for each queue depth band. An assessment program selects the matrix for the recorded band at each step and propagates state means and covariances through the discrete surrogate. The program calculates expected squared lateral tracking error from the resulting moments and reports the time indices where it exceeds the specified limit.

Specific version of a part

Correlated steering response delay

The continuous vehicle model uses a feedback matrix to compute the steering command from lateral offset and heading error. A bounded stochastic delay parameter varies over time and sets which earlier command the steering actuator receives.

Fixed step integration with interpolation

The evaluator forms a discrete transition with a four-stage integration rule and a 5-millisecond step. At each integration stage, it interpolates the actuator input from the two stored commands that bracket the stage time minus the modeled delay.

Polynomial surrogate for delay histories

The evaluator represents the sampled delay history with six independent, unit-variance random coefficients using a covariance decomposition. At each sample time, it fits a quadratic polynomial for each state component using trajectories from the discrete vehicle model.

Two sided lateral offset check

At each sample time, the evaluator forms an interval from the mean lateral offset minus two standard deviations to the mean plus two standard deviations. It records the sample time when the lower endpoint falls below minus 0.30 meters or the upper endpoint rises above 0.30 meters.

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 2020/0400778 A1
Published
Dec 24, 2020
Filed
Sep 4, 2020
Class
G01S 7/006
Document BUS 2024/0135185 A1
Published
Apr 25, 2024
Filed
Feb 10, 2023
Class
G06N 3/088
Claim 1
Document A
Document B
1.1receiving a time-continuous dynamic model for describing the system to be controlled, wherein the time-continuous dynamic model is (ƒ(x, u, p(ζ))) represented by a specified function containing time-dependent arguments comprising: a state vector (x) comprising one or more state variables; an input vector (u) comprising one or more input variables, which input vector is connected to the state vector (x) by a feedback matrix (K) describing a state controller of the system; and an uncertainty vector (p(ζ)) with at least one component comprising a time-dependent uncertainty parameter representing a stochastic process associated with an uncertain time effect;
1.2forming a time-discrete representation of the time-continuous dynamic model by time discretizing the time-continuous dynamic model;
1.3determining a surrogate model of the time-discrete representation of a state-space model;
In the surrogate modeling stage, we consider Polynomial Chaos Expansions (PCEs) to learn the mapping from the latent distribution to the output space.
B · US 2024/0135185 A1 · ¶ 51
1.4solving the surrogate model to determine a mean (E(xk)) of the state vector (xk) and a deviation, in particular a variance (V(xk)), of the state vector (zk) at a plurality of discrete time points (tk);
MMD is a moment matching technique, and therefore, enables us to match high order moments of the output distribution to the model response.
B · US 2024/0135185 A1 · ¶ 51
1.5searching for one or more discrete time points (t*k) from the plurality of discrete time points using the determined mean of the state vector (xk) and the determined deviation, in particular the variance, of the state vector (xk), for which a specified rough requirement criterion for the state vector is not satisfied; and
The device may determine the first channel state variance of CSI entries for different packets during the time period. When the first channel state variance exceeds a threshold value (e.g., threshold value 302 of FIG. 3A), such may indicate motion in between the device and the second device.
A · US 2020/0400778 A1 · ¶ 94
1.6classifying the uncertain time effects as critical for controlling the system if the one or more discrete time points (t*k) have been identified, for which the rough requirement criterion for the state vector is not satisfied, and otherwise classifying the uncertain time effects as non-critical for controlling the system.

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

Statements in these documents
Wireless devices are becoming widely prevalent and are increasingly requesting access to wireless channels.
A · US 2020/0400778 A1 · ¶ 3
However, uncertainty quantification (UQ) techniques are computationally intensive and become prohibitively expensive for high-dimensional data.
B · US 2024/0135185 A1 · ¶ 6
The method additionally includes learning, by the computing device, a polynomial chaos expansion to map the new data samples in the latent space to the corresponding data output to learn the set of distributions and their relation to perform estimation with high-dimensional dataset under uncertainty such as missing values by estimating the values using the set of distributions.
B · US 2024/0135185 A1 · ¶ 7
Cite aspatent.fyi, US 2026/0356373 A1, claim 1, first published 2026-11-20T00:27:46-05:00, https://patent.fyi/US20260356373/, SHA-256 e47fece74b713ebb8c96df13b6a5ff8f23cf48db8ab4fbb67f9395986c3970db
First published
Permanent addresshttps://patent.fyi/US20260356373/
Content fingerprint (SHA-256)e47fece74b713ebb8c96df13b6a5ff8f23cf48db8ab4fbb67f9395986c3970db
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi