Zoox, Inc.

Jacob Daniel Boydston

Chuang Wang

Joshua Kriser Cohen

James William Vaisey Philbin

Michael Carsten Bosse

Yuesong Xie

Giacomo Zavolta Taylor

Hang Ren

Yuan Zhang

Techniques for clustering sensor data are discussed herein. Sensors of a vehicle may detect data points in an environment. Clustering techniques can be used in a vehicle safety system to determine connection information between the data points. Further, the machine learned model can be trained to output motion data representing velocity data or a cluster association score. The connection information and/or motion data can be used by a vehicle computing device to detect objects in an environment and/or to control operation of a vehicle.

Perception system velocity determination

Techniques for determining a safety area for a vehicle are discussed herein. In some cases, a first safety area can be based on a vehicle travelling through an environment and a second safety area can be based on a steering control or a velocity of the vehicle. A width of the safety areas can be updated based on a position of a bounding box associated with the vehicle. The position can be based on the vehicle traversing along a trajectory. Sensor data can be filtered based on the sensor data falling within the safety area(s).

Lateral safety area

Inaccurate sensor data, such as by caused by reflections, may skew ground profile estimations that identify a profile or plane associated with a surface, such as a roadway. Correcting a ground profile that was based on inaccurate sensor data may include determining a location at which a slope of the ground profile meets or exceeds a maximum slope, determining a line associated with the location and having a slope of a magnitude equal to the maximum slope, and using the line to determine a subset of sensor data points that lies outside a region defined by the line. The subset of sensor data points that lies outside the region may be excluded from the set of sensor data points used to generate ground profile(s) and a new ground profile may be determined using the set of sensor data points, less the subset of sensor data points.

Ground profile estimation for sensor data

Determining safety area based on bounding box

A collision avoidance system may validate, reject, or replace a trajectory generated to control a vehicle. The collision avoidance system may comprise a secondary perception component that may receive sensor data, receive and/or determine a corridor associated with operation of a vehicle, classify a portion of the sensor data associated with the corridor as either ground or an object, determine a position and/or velocity of at least the nearest object, determine a threshold distance associated with the vehicle, and control the vehicle based at least in part on the position and/or velocity of the nearest object and the threshold distance.

Collision avoidance perception system

Classifying sensor data as being associated with ground (as opposed to an object) may comprise determining a number of channels of sensor data that have returns in them, setting a number of control points and a number of knots of a curve based at least in part on the number of channels that have returns, and fitting a curve having the number of control points and the number of knots to the sensor data. The curve may be used to distinguish sensor data associated with the ground from sensor data associated with an object. Determining the curve may additionally or alternatively include limiting an elevation value of a control point and/or knot based on elevation value(s) of the sensor data, weighting the sensor data based at least in part on elevation values associated with the sensor data, and/or adjusting knot spacing, et alia.

Growing community of inventors

Foster City, CA, United States of America

Jacob Daniel Boydston