Ford Global Technolgoies, LLC

Argo Ai Gmbh

Randall Schur

Constantin Savtchenko

Pietro Blandizzi

Kevin Lee Wyffels

Alexander Metz

This document discloses system, method, and computer program product embodiments for operating a vehicle, comprising: using kinematic models to generate forecasted trajectories of an actor (the kinematic models being respectively associated with different actor types that are assigned to an actor detected in an environment of the vehicle); selecting a first kinematic model based on the forecasted trajectories and a kinematic state of the actor; using the first kinematic model to predict a first path for the actor; selecting a second kinematic model responsive to movement of the actor no longer being consistent with typical movement of an object of one of the different actor types that is associated with the first kinematic model; using the second kinematic model to predict a second path for the actor; and controlling operations of the vehicle based on the first and second paths.

Method and system for determining a mover model for motion forecasting in autonomous vehicle control

This document discloses system, method, and computer program product embodiments for operating a robot. For example, the method includes: detecting an object in proximity to the robot; generating a path of travel for the object that is defined by a plurality of first data points representing different object locations in an area; generating cost curves respectively associated with the plurality of first data points; generating a polyline comprising a plurality of second data points associated with displacements of the cost curves from a particular location in the area; defining a predicted path of travel for the object based on the polyline; and using the predicted path of travel for the object to facilitate at least one autonomous operation of the robot.

Systems and methods for vehicle motion planning

Disclosed herein are system, method, and computer program product embodiments for state identification for road actors with uncertain measurements based on compliant priors. The perception system of an autonomous vehicle (AV) may detect an object with an uncertain kinematic state based on sensor information received from a sensing device associated with the AV. A first, second, and third distribution of velocity values may be generated based on the sensor information, map information, and a kinematic state for each additional object of a plurality of additional objects in proximity to the detected object with the uncertain kinematic state, respectively. A velocity value for the detected object with the uncertain kinematic state may be generated based on a comparison of the first, second, and third distributions of velocity values. The AV may perform a driving maneuver based on the velocity value for the detected object.

State identification for road actors with uncertain measurements based on compliant priors

Systems/methods for operating an autonomous vehicle. The methods comprise: detecting an object in proximity to the autonomous vehicle; determining a path of travel for the object that comprises a number of data points that is equal to a given number of vehicle locations selected based on a geometry of a lane in which the object; generating cost curves respectively associated with the data points, each cost curve representing a displacement cost to be at a particular location along a given cross line that (i) passes through a respective data point of said data points and (ii) extends perpendicular to and between boundary lines of the lane; determining a polyline representing displacements of the cost curves from a center of the lane; defining a predicted path of travel for the object based on the polyline; and using the predicted path of travel for the object to facilitate autonomous driving operation(s).

Methods of determining which kinematic model an autonomous vehicle (AV) should use to predict motion of a detected moving actor are disclosed. One or more sensors of the AV sensors will detect a moving actor. The AV will assign one or more probable classes to the actor, and it will process the information to determine a kinematic state of the actor. The system will query a library of kinematic models to return one or more kinematic models that are associated with each of the probable classes. The system will apply each of the returned kinematic models to predict trajectories of the actor. The system will then evaluate each of the forecasted trajectories of the actor against the kinematic state of the actor to select one of the returned kinematic models to predict a path for the actor. The system will then use the predicted path to plan motion of the AV.

Growing community of inventors

Pittsburgh, PA, United States of America

Randall Schur