Argo Ai Gmbh

Ford Global Technolgoies, LLC

Malwarebytes Corporation

Volkswagen Group of America Investments, LLC

Other

Andrew Thomas Hartnett

Constantin Savtchenko

G Peter K Carr

Greydon Foil

Jagjeet Singh

George Peter Kenneth Carr

Lorenzo Nardi

Brett Browning

Neal Andrew Seegmiller

Christopher Cunningham

Douglas Stuart Swanson

Slawomir W Bak

Michael Pacilio

Nikolai Popov

Mark Ollis

Gregory Boyd Nichols

John Russell Lepird

Deva K Ramanan

Matthew Lee Gilson

Friederike Schneemann

Alexander Metz

William Tyler Krampe

Siddhesh Shyam Khandelwal

Peter W Rander

Shawn Cook

William Junbo Qi

Bangjie Liu

Andreas Eberlein

Elena Corina Grigore

Systems and methods for planning a trajectory for a vehicle are disclosed. The methods include generating first predicted states of an object in an environment of the vehicle at one or more first time steps during a first prediction horizon using a machine learning model and sensor data about the environment. The first predicted states are used as an input to a physical model to generate second predicted states of the object at one or more second time steps during a second prediction horizon where the one or more second time steps are after the one or more first time steps. A trajectory for the vehicle is generated using the first predicted states or the second predicted states of the object.

Methods and systems for long-term trajectory prediction by extending a prediction horizon

Disclosed herein are system, method, and computer program product aspects for testing autonomous vehicle intent using protolanes. For example, the method includes selecting a first lane segment of a roadway within a geonet, wherein the geonet comprises a plurality of lane segments. It is determined that the first lane segment corresponds to a protolane of a set of protolanes, wherein the protolane is associated with one or more additional lane segments of the plurality of lane segments. The protolane is associated with the first lane segment based on the determination that the first lane segment corresponds to the protolane.

Protolanes for testing autonomous vehicle intent

Disclosed herein are system, method, and computer program product embodiments for clustering lane segments of a roadway in order to improve and simplify autonomous vehicle behavior testing. The approaches disclosed herein provide a hybrid methodology of dividing lane segments into hard features and soft features, and using a metric learning model trained in a supervised process on the entirety of lane segment features to cluster the lane segments based on the soft features. These clustered lane segments can then be assigned to what is termed as protolanes, where a single set of tests applied to a given protolane is considered valid across all of the lane segments assigned to the protolane.

Lane segment clustering using hybrid distance metrics

Systems and methods for assigning a lane to an object in an environment of an autonomous vehicle are disclosed. The methods include assigning an instantaneous probability to each of a plurality of lanes in the environment based on a current state of the object, generating a transition matrix for each of the plurality of lanes, and identifying the lane in which the object is moving at the current time t based on the instantaneous probability and the transition matrix. The instantaneous probability is a measure of likelihood that the object is in that lane at a current time. The transition matrix encodes one or more probabilities that the object transitioned either into that lane or out of that lane at the current time.

Methods and systems for tracking a mover's lane over time

Systems and methods for operating a robot. The methods comprise: performing, by a processor, operations to detect an object that is moving; identifying, by the processor, detected behavior of the object that constitutes an unrecognized behavior; predicting, by the processor, future movement of the object based on a circle having a radius that is function of a velocity of the object; and controlling operations of the robot based on the predicting.

Methods and systems for autonomous vehicle inference of routes for actors exhibiting unrecognized behavior

Disclosed herein are system, method, and computer program product aspects for validating protolanes. For example, the method includes determining a protolane level flowchart from one or more flowcharts corresponding to one or more gates of a protolane selected from a set of protolanes of a geonet, wherein the one or more flowcharts correspond to conditions that can be encountered by an actor approaching the corresponding one or more gates, and wherein the geonet comprises a plurality of lane segments associated with the set of protolanes. A graph through the protolane level flowchart is generated that corresponds to actions of an actor. Reasoning of the actor is validated by testing the actions of the actor against the graph through the protolane level flowchart.

Validating protolanes with actor behavior flowcharts

Systems and methods for predicting actions of an actor before entering a conflicted area are disclosed. The methods may include detecting the presence of an actor in an environment of an autonomous vehicle while the autonomous vehicle and the actor are approaching the conflicted area, determining whether the autonomous vehicle has precedence over the actor for traversing the conflicted area, assigning a kinematic target to the moving object that requires the moving object to come to a stop at an yield point before entering the conflicted area if the autonomous vehicle has precedence over the actor, determining whether a plurality of forecasted trajectories for the actor need to be generated, and controlling movement of the autonomous vehicle to traverse the conflicted area accordingly.

Methods and systems for predicting actions of an object by an autonomous vehicle to determine feasible paths through a conflicted area

Methods by which an autonomous vehicle may predict actions of other actors are disclosed. A vehicle will assign either a high priority rating or a low priority rating to each actor that it detects. The vehicle will then generate a forecast for each of the detected actors. Some of not all high priority actors will receive a high resolution forecast. Low priority actors, and optionally also some of the high priority actors, will receive a low resolution forecast. The system will the forecasts to predict actions for the actors. The autonomous vehicle will then use the predicted actions to determine its trajectory.

Method and system for predicting behavior of actors in an environment of an autonomous vehicle

A system includes an on-board electronic device of an autonomous vehicle, and a computer-readable medium having one or more programming instructions. The system receives one or more forecast messages pertaining to a track, where each of the forecast messages includes a unique identifier associated with the track, and receives one or more inference messages pertaining to the track, where each of the inference messages includes the unique identifier. The system aggregates the one or more forecast messages and the one or more inference messages to generate a message set, and applies a set of processing operations to the message set to generate a feedback message. The system identifies one or more events from the feedback message, automatically generates an annotation for one or more of the events that is identified, and embeds the generated annotations in an event log for the autonomous vehicle.

On-board feedback system for autonomous vehicles

Systems/methods for operating an autonomous vehicle. The methods comprise, by a computing device: using sensor data to track an object that was detected in proximity to the autonomous vehicle; classifying the object into at least one dynamic state class of a plurality of dynamic state classes; transforming the at least one dynamic state class into at least one goal class of a plurality of goal classes; transforming the at least one goal class into at least one proposed future intention class of a plurality of proposed future intention classes; determining at least one predicted future intention of the object based on the proposed future intention class; and/or causing the autonomous vehicle to perform at least one autonomous driving operation based on the at least one predicted future intention determined for the object.

Systems and methods for determining future intentions of objects

A system receives a road network map that corresponds to a road network that is in an environment of an autonomous vehicle. For each of the one or more lane segments, the system identifies one or more conflicting lane segments from the plurality of lane segments, each of which conflicts with the lane segment, and adds conflict data pertaining to a conflict between the lane segment and the one or more conflicting lane segments to a set of conflict data. The system analyzes the conflict data to identify a conflict cluster that is representative of an intersection. The system groups predecessor lane segments and the successor lane segments as inlets or outlets of the intersection, generates an outer geometric boundary of the intersection, generates an inner geometric boundary of the intersection, creates a data representation of the intersection and adds the data representation to the road network map.

Methods and system for constructing data representation for use in assisting autonomous vehicles navigate intersections

Systems and methods of determining trajectories of an actor in an environment in which a vehicle is operating are provided. The method includes, by an object detection system of a vehicle in an environment, detecting an actor that may move within a scene in the environment. The method further includes using context of the scene to determine a reference polyline for the actor and determining a kinematic history of the actor. The method additionally includes using the kinematic history to predict a waypoint, which is a predicted position of the actor at a conclusion of a waypoint time period, and identifying a segment of the reference polyline, the segment extending from a current location to a point along the reference polyline that is closest to the waypoint and determining a trajectory for the actor conditioned by the segment of the reference polyline.

Waypoint prediction for vehicle motion planning

Methods of forecasting intentions of actors that an autonomous vehicle (AV) encounters in are disclosed. The AV uses the intentions to improve its ability to predict trajectories for the actors, and accordingly making decisions about its own trajectories to avoid conflict with the actors. To do this, for any given actor the AV determines a class of the actor and detects an action that the actor is taking. The system uses the class and action to identify candidate intentions of the actor and evaluating a likelihood of each candidate intention. The system repeats this process over multiple cycles to determine overall probabilities for each of the candidate intentions. The AV's motion planning system can use the probabilities to determine likely trajectories of the actor, and accordingly influence the trajectory that the AV will itself follow in the environment.

Assessing present intentions of an actor perceived by an autonomous vehicle

Systems and methods for forecasting trajectories of objects. The method includes obtaining a prediction model trained to predict future trajectories of objects. The prediction model is trained over a first prediction horizon selected to encode inertial constraints in a predicted trajectory and over a second prediction horizon selected to encode behavioral constraints in the predicted trajectory. The method also include generating a planned trajectory of an autonomous vehicle by receiving state data corresponding to the autonomous vehicle, receiving perception data corresponding to an object, predicting a future trajectory of the object based on the perception data and the prediction model, and generating the planned trajectory of the autonomous vehicle based on the future trajectory of the object and the state data.

Methods and systems for trajectory forecasting with recurrent neural networks using inertial behavioral rollout

A method of determining a trajectory for an autonomous vehicle is disclosed. An ego-vehicle may detect a moving actor in an environment. To choose between candidate trajectories for the ego-vehicle, the system will consider the cost of each candidate trajectory to the moving actor. The system will use the candidate trajectory costs for the candidate trajectories to select one of the candidate trajectories via which to move the ego-vehicle. An autonomous vehicle system of the ego-vehicle may then move the ego-vehicle in the environment along the selected trajectory.

Growing community of inventors

Pittsburgh, PA, United States of America

Andrew Thomas Hartnett