Waymo LLC

Google Inc.

Junhua Mao

Congcong Li

Yang Song

Xinwei Shi

Dragomir Dimitrov Anguelov

Ruizhongtai Qi

Zhishuai Zhang

Qian Yu

Zhifeng Chen

Chen Wu

Khaled Refaat

Jonathon Shlens

Vijay Vasudevan

Chen Sun

Edward Stephen Walker, Jr

Jiquan Ngiam

Ouais Alsharif

Yin Zhou

Jiyang Gao

Wei Han

Benjamin James Caine

Zhengdong Zhang

Brandon Chauloon Yang

Christoph Sprunk

Andre Liang Cornman

Alexander Gorban

Jingxiao Zheng

Ting Liu

Jeonhyung Kang

Alper Ayvaci

Ruichi Yu

Kevin Patrick Murphy

Lo Po Tsui

Zijian Guo

Tian Lan

Nichola Abdo

Jonathan Chandler Stroud

Jiachen Li

Xiaohan Jin

Huayi Zeng

Yukai Liu

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for agent behavior prediction using keypoint data. One of the methods includes obtaining data characterizing a scene in an environment, the data comprising: (i) context data comprising data characterizing historical trajectories of a plurality of agents up to the current time point; and (ii) keypoint data for a target agent; processing the context data using a context data encoder neural network to generate a context embedding for the target agent; processing the keypoint data using a keypoint encoder neural network to generate a keypoint embedding for the target agent; generating a combined embedding for the target agent from the context embedding and the keypoint embedding; and processing the combined embedding using a decoder neural network to generate a behavior prediction output for the target agent that characterizes predicted behavior of the target agent after the current time point.

Pedestrian behavior prediction with 3D human keypoints

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for estimating a 3-D pose of an object of interest from image and point cloud data. In one aspect, a method includes obtaining an image of an environment; obtaining a point cloud of a three-dimensional region of the environment; generating a fused representation of the image and the point cloud; and processing the fused representation using a pose estimation neural network and in accordance with current values of a plurality of pose estimation network parameters to generate a pose estimation network output that specifies, for each of multiple keypoints, a respective estimated position in the three-dimensional region of the environment.

Multi-modal 3-D pose estimation

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium that determine yield behavior for an autonomous vehicle, and can include identifying an agent that is in a vicinity of an autonomous vehicle navigating through a scene at a current time point. Scene features can be obtained and can include features of (i) the agent and (ii) the autonomous vehicle. An input that can include the scene features can be processed using a first machine learning model that is configured to generate (i) a crossing intent prediction that includes a crossing intent score that represents a likelihood that the agent intends to cross a roadway in a future time window after the current time, and (ii) a crossing action prediction that includes a crossing action score that represents a likelihood that the agent will cross the roadway in the future time window after the current time.

Pedestrian crossing intent yielding

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for searching an autonomous vehicle sensor data repository. One of the methods includes maintaining a collection of sensor samples and, for each sensor sample, an embedding of the sensor sample; receiving a request specifying a query sensor sample, wherein the query sensor sample characterizes a query environment region; and identifying, from the collection of sensor samples, a plurality of relevant sensor samples that characterize similar environment regions to the query environment region, comprising: processing the query sensor sample through the embedding neural network to generate a query embedding; and identifying, from sensor samples in a subset of the sensor samples in the collection, a plurality of sensor samples that have embeddings that are closest to the query embedding.

Searching an autonomous vehicle sensor data repository

Aspects of the subject matter disclosed herein include methods, systems, and other techniques for training, in a first phase, an object classifier neural network with a first set of training data, the first set of training data including a first plurality of training examples, each training example in the first set of training data being labeled with a coarse-object classification; and training, in a second phase after completion of the first phase, the object classifier neural network with a second set of training data, the second set of training data including a second plurality of training examples, each training example in the second set of training data being labeled with a fine-object classification.

Neural networks for coarse- and fine-object classifications

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating rare pose data. One of the methods includes obtaining a three-dimensional model of a dynamic object, wherein the dynamic object has multiple movable elements that define a plurality of poses of the dynamic object. A plurality of template poses of the dynamic object are used to generate additional poses for the dynamic object including varying angles of one or more key joints of the dynamic object according to the three-dimensional model. Point cloud data is generated for the additional poses generated for the dynamic object.

Rare pose data generation

Some aspects of the subject matter disclosed herein include a system implemented on one or more data processing apparatuses. The system can include an interface configured to obtain, from one or more sensor subsystems, sensor data describing an environment of a vehicle, and to generate, using the sensor data, (i) one or more first neural network inputs representing sensor measurements for a particular object in the environment and (ii) a second neural network input representing sensor measurements for at least a portion of the environment that encompasses the particular object and additional portions of the environment that are not represented by the one or more first neural network inputs; and a convolutional neural network configured to process the second neural network input to generate an output, the output including a plurality of feature vectors that each correspond to a different one a plurality of regions of the environment.

Object classification using extra-regional context

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing data generated by a sensing system that rotationally senses an environment. In one aspect, a method comprises partitioning a predetermined period of time into a plurality of sub-periods, wherein the predetermined period of time is a period of time for which data generated by the sensing system constitutes a complete rotational sensing of the environment; for each sub-period: receiving current data generated by the sensing system during the sub-period and characterizing a respective partial scene of the environment; processing the current data using an object detection neural network to generate a current object detection output that is specific to the respective partial scene of the environment.

Streaming object detection within sensor data

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing point cloud data using spatio-temporal-interactive networks.

Spatio-temporal-interactive networks

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a classifier to detect open vehicle doors. One of the methods includes obtaining a plurality of initial training examples, each initial training example comprising (i) a sensor sample from a collection of sensor samples and (ii) data classifying the sensor sample as characterizing a vehicle that has an open door; generating a plurality of additional training examples, comprising, for each initial training example: identifying, from the collection of sensor samples, one or more additional sensor samples that were captured less than a threshold amount of time before the sensor sample in the initial training example was captured; and training the machine learning classifier on first training data that includes the initial training examples and the additional training examples to generate updated weights for the machine learning classifier.

Training a classifier to detect open vehicle doors

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating vehicle intent predictions using a neural network. One of the methods includes obtaining an input characterizing one or more vehicles in an environment; generating, from the input, features of each of the vehicles; and for each of the vehicles: processing the features of the vehicle using each of a plurality of intent-specific neural networks, wherein each of the intent-specific neural networks corresponds to a respective intent from a set of intents, and wherein each intent-specific neural network is configured to process the features of the vehicle to generate an output for the corresponding intent.

Vehicle intent prediction neural network

Aspects of the disclosure relate to training and using a model for identifying actions of objects. For instance, LIDAR sensor data frames including an object bounding box corresponding to an object as well as an action label for the bounding box may be received. Each sensor frame is associated with a timestamp and is sequenced with respect to other sensor frames. Each given sensor data frame may be projected into a camera image of the object based on the timestamp associated with the given sensor data frame in order to provide fused data. The model may be trained using the fused data such that the model is configured to, in response to receiving fused data, the model outputs an action label for each object bounding box of the fused data. This output may then be used to control a vehicle in an autonomous driving mode.

Object action classification for autonomous vehicles

Growing community of inventors

Palo Alto, CA, United States of America

Junhua Mao