1.1Road geometry from map
A satellite navigation receiver and inertial sensors locate the vehicle along a lane centerline stored as surveyed coordinates. The planner extracts the next 120 meters and fits a spline to those coordinates.
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.
A satellite navigation receiver and inertial sensors locate the vehicle along a lane centerline stored as surveyed coordinates. The planner extracts the next 120 meters and fits a spline to those coordinates.
Wheel speed pulses feed a distance accumulator that triggers the camera at two-meter increments. The road estimator processes each new frame and uses its capture pose to align the detected lane edges with the road model.
The lane estimator supplies the lateral offset and heading error at the vehicle position. A feedback controller converts those errors into a steering angle, while a separate speed controller uses upcoming road curvature.
The controller aligns five curvature profiles by distance along the road using vehicle odometry. At each road point, it sorts the five signed curvature values and passes the third value to the speed planner.
A forward range sensor measures how far the planned path remains visible and clear. The planner caps speed so the distance traveled during actuator delay plus the braking distance at a specified deceleration fits within the clear distance.
Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.
A downward tilted camera on a motorized warehouse cart captures floor markings at regular intervals. The controller projects the markings onto the floor plane, fits a guide path from the latest frame, and uses wheel odometry to register recent path estimates in a common floor frame. It averages curvature values at corresponding positions along the aisle and sets speed limits from a lateral acceleration limit for the payload. The steering and drive controllers follow the latest guide path and apply the resulting speed profile.
A Kalman filter replaces the arithmetic mean at each point on a distance indexed road grid. The controller transforms successive camera based road estimates into a common local frame and assigns each curvature measurement to its corresponding grid point. Each filter uses a random walk curvature model and updates its estimate when a measurement arrives, with measurement variance set from the road fit residual. The controller derives the speed profile from the filtered curvature and refreshes the target trajectory from the latest road shape.
A vehicle operating at up to 20 kilometers per hour captures road images at 20 frames per second and estimates road shape over the next 15 meters. The controller registers the latest eight estimates using vehicle odometry and averages curvature at corresponding points spaced 0.5 meters apart. It caps speed at each point using a lateral acceleration limit of 1.2 meters per second squared and propagates the caps backward using a deceleration limit of 1.5 meters per second squared. The vehicle follows the trajectory from the latest frame at the resulting target speeds.
A forward camera supplies road geometry, and a forward radar measures the range and relative speed of a lead vehicle. The controller refreshes the target trajectory from the latest road estimate and derives a curve speed cap from curvature averaged across recent frames. An adaptive cruise controller computes a following speed from the radar measurements and a set time gap. The longitudinal controller uses the lower of the curve speed cap and the following speed, while the steering controller tracks the target trajectory.
A chassis controller estimates tire road friction from wheel speeds, applied drive or brake torque, and measured vehicle acceleration when tire slip provides usable data. The trajectory controller registers and averages recent camera based curvature estimates at corresponding road positions. It computes the lateral acceleration budget as the estimated friction coefficient multiplied by gravitational acceleration and a fixed margin factor below one, then converts that budget and the averaged curvature into a pointwise speed cap. The controller propagates the speed caps backward using a braking limit and tracks the latest target trajectory at the resulting speeds.
A regression model assigns an error variance to each point of a road shape estimated from a camera frame. The model uses local image contrast, blur, and road fit residuals as inputs and uses curvature errors measured against surveyed road geometry as training targets. The controller registers recent estimates by road position and averages their curvature values at each point using inverse variance weights. It sets the target speed from the weighted curvature and follows a trajectory generated from the latest road shape.
A forward-facing camera captures an image every 50 milliseconds. The processor projects detected lane markings onto an estimated road surface and fits cubic splines to the resulting lane boundaries.
After each road shape estimate, the planner samples the left and right lane boundaries at one-meter intervals along the road and computes their midpoints. It fits a spline from the vehicle position and heading toward those midpoints to form the target trajectory.
The processor aligns curvature estimates from the latest 250 milliseconds on a common distance grid and computes their arithmetic mean at each grid point. The speed planner divides a lateral acceleration limit by the magnitude of the mean curvature and takes the square root, with a maximum speed cap that also applies at zero curvature. It propagates lower speeds backward along the grid using a specified braking deceleration.
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.
The vehicle 2 may include a camera for taking a picture of a region around the vehicle 2. The camera may generate an image representing the region around the vehicle 2 at predetermined intervals, and output the generated image to the trajectory recorder via the in-vehicle network.A · US 2024/0253662 A1 · ¶ 24
The techniques can send an indication of the reference trajectory to the vehicle for use in planning a trajectory for the vehicle. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on the trajectory.B · US 12,130,621 B2 · abstract
An empty cell means only that no passage was reproduced there.
The use of such an unnatural trajectory for generating a trajectory serving as a reference may produce an inappropriate reference trajectory.A · US 2024/0253662 A1 · ¶ 4
It is an object of the present invention to provide a reference trajectory generating device that can generate a trajectory serving as a reference in a predetermined section of a road appropriately.A · US 2024/0253662 A1 · ¶ 5
However, in certain situations, the vehicle may be unable to navigate past a portion of the environment that impedes progress of the vehicle.B · US 12,130,621 B2 · ¶ 1
FIG. 6 is a flowchart depicting an example process for sending a reference trajectory to a vehicle to guide the vehicle.B · US 12,130,621 B2 · ¶ 8
patent.fyi, US 2026/0356350 A1, claim 1, first published 2026-11-20T00:27:46-05:00, https://patent.fyi/US20260356350/, SHA-256 89cd8e1dc26a11270cf1546690466677acb74d2cdb99af8766d9639e772e454d