Zoox, Inc.

Jifei Qian

Subhasis Das

Chuang Wang

Anton Mario Bongio Karrman

Michael Carsten Bosse

Francesco Papi

Patrick Blaes

Joshua Kriser Cohen

Cooper Stokes Sloan

Shiwei Sheng

Yassen Ivanchev Dobrev

Arvind Srivastav

Liujiang Yan

Yifan Zuo

Techniques for detecting radar data inaccuracies using a multistage filtering technique are discussed herein. A vehicle may capture radar data including a set of radar points. The vehicle can input the radar data into a machine-learning model configured to detect objects and/or generate bounding boxes associated with such detected objects. In some instances, the vehicle may cluster the radar points such that a cluster includes the radar points located proximate to and/or within a bounding box. To identify double bounce object detections, the vehicle may determine whether a detected object is a false-positive double bounce object detection based on evaluating the azimuth, range, and relative velocity double bounce criteria. Based on the detected object satisfying the double bounce criteria for azimuth, range, and relative velocity, the vehicle may modify the radar data.

Radar double bounce detection

Techniques for updating data operations in a perception system are discussed herein. A vehicle may use a perception system to capture data about an environment proximate to the vehicle. The perception system may output the data about the environment to a system configured to determine positions of objects relative to the perception system over time. The positions of the objects may be used to estimate an object velocity and may be compared against machine learning model outputs in a self-supervised manner to train the machine learning model to output object velocities based on inputs from the perception system. The output of the machine learning model may include a two-dimensional velocity for objects in the environment. The two-dimensional velocity may be used for a vehicle system such that the vehicle can make environmentally aware operational decisions, which may improve reaction time(s) and/or safety outcomes of the vehicle.

Self-supervised global velocity determination for perception system

Techniques for determining an output from a plurality of sensor modalities are discussed herein. Features from a radar sensor, a lidar sensor, and an image sensor may be input into respective models to determine respective intermediate outputs associated with a tracks associated with an object and associated confidence levels. The Intermediate outputs from a radar model, a lidar model, and an vision model may be input into a fused model to determine a fused confidence level and fused output associated with the track. The fused confidence level and the individual confidence levels are compared to a threshold to generate the track to transmit to a planning system or prediction system of an autonomous vehicle. Additionally, a vehicle controller can control the autonomous vehicle based on the track and/or on the confidence level(s).

Track confidence model

Some radar sensors may provide a Doppler measurement indicating a relative velocity of an object to a velocity of the radar sensor. Techniques for determining a two-or-more-dimensional velocity from one or more radar measurements associated with an object may comprise determining a data structure that comprises a yaw assumption and a set of weights to tune the influence of the yaw assumption. Determining the two-or-more-dimensional velocity may further comprise using the data structure as part of regression algorithm to determine a velocity and/or yaw rate associated with the object.

Radar-tracked object velocity and/or yaw

Techniques for updating data operations in a perception system are discussed herein. A vehicle may use a perception system to capture data about an environment proximate to the vehicle. The perception system may receive state data stored in cyclic buffer of globally registered detection and occasionally converted to gridded point cloud in a local reference frame. The two-dimensional gridded point cloud may be processed using one or more neural networks to generate semantic data associated with a scene or physical environment surrounding the vehicle such that the vehicle can make environment aware operational decisions, which may improve reaction time(s) and/or safety outcomes of the autonomous vehicle.

Perception system

Growing community of inventors

Campbell, CA, United States of America

Jifei Qian