Abstract: A computer-readable medium stores instructions executable by one or more processors to implement a self-driving control architecture for controlling an autonomous vehicle. A perception and prediction component receives sensor data, and generates (1) an observed occupancy grid indicating which cells are currently occupied in a two-dimensional representation of the environment, and (2) predicted occupancy grids indicating which cells are expected to be occupied later. A mapping component provides navigation data for guiding the vehicle toward a destination, and a cost map generation component is configured to generate, based on the observed occupancy grid, the predicted occupancy grid(s), and the navigation data, cost maps that each specify numerical values representing a cost, at a respective instance of time, of occupying certain cells in a two-dimensional representation of the environment.

Abstract: In one embodiment, a lidar system includes a light source configured to emit pulses of light. The lidar system also includes multiple optical links and multiple sensor heads. Each optical link couples the light source to a corresponding sensor head, and each optical link is configured to convey at least a portion of the emitted pulses of light from the light source to the corresponding sensor head. Each sensor head includes a scanner configured to scan pulses of light across a field of regard of the sensor head, where the scanned pulses of light include the portion of the emitted pulses of light conveyed from the light source to the sensor head by the corresponding optical link. Each sensor head also includes a receiver configured to detect at least a portion of the scanned pulses of light scattered or reflected by a target located downrange from the sensor head.

Abstract: A machine vision system comprises a camera configured to generate one or more images of a field of regard of the camera, a lidar system, and a processor. The lidar system includes a laser configured to emit light, where the emitted light is directed toward a region within the field of regard of the camera and a receiver configured to detect light returned from the emitted light. The processor is configured to receive an indication of a location based on the returned light and determine whether a solid object is present at the location based on the one or more images.

Abstract: A lidar system includes a light source configured to produce a beam of light, a scanner configured to scan a field of regard of the lidar system, and a receiver configured to detect light from the beam of light scattered by a remote target. The scanner includes a polygon mirror having a block with a first wall, a second wall, and several reflective surfaces angularly offset from one another along a periphery of the block, the polygon mirror configured to rotate about a scan-mirror rotation axis to scan the beam of light across the field of regard. The scanner further includes a bracket adjacent to the polygon mirror, where at least one of the polygon mirror or the bracket includes a noise-reducing feature configured to reduce acoustic noise produced when pressure waves generated by the polygon mirror during rotation are incident on one or more components of the scanner.

Abstract: In one embodiment, a method includes emitting, by a light source of a lidar system, multiple optical pulses using multiple alternating pulse repetition intervals (PRIs) that include a first PRI and a second PRI, where the first PRI and the second PRI are not equal. The method also includes detecting, by a receiver of the lidar system, multiple input optical pulses and generating, by a processor of the lidar system, multiple pixels. Each pixel of the multiple pixels corresponds to one of the input optical pulses, and each pixel includes a PRI associated with a most recently emitted optical pulse of the multiple optical pulses. The method also includes determining a group of neighboring pixels for a particular pixel of the multiple pixels and determining whether the particular pixel is range-wrapped based at least in part on the PRI associated with each pixel of the group of neighboring pixels.

Abstract: A lidar system comprises a light source configured to emit light, a scanner configured to direct the emitted light to scan a field of regard of the lidar system in accordance with a scan pattern, a receiver configured to detect the light scattered by one or more remote targets, and a controller configured to control motion of at least the second mirror to modify the scan pattern. The scanner includes a rotatable polygon mirror having a block having a first wall, a second wall, and reflective surfaces extending between the first and second walls, the reflective surfaces being angularly offset from one another along a periphery of the block. The scanner also includes a polygon mirror axle extending into the block through at least one of the first and second walls, about which the block rotates, and a second mirror pivotable along an axis orthogonal to the polygon mirror axle.

Abstract: In one embodiment, a lidar system includes a light source configured to emit pulses of light and a scanner configured to scan at least a portion of the emitted pulses of light across a field of regard. The lidar system also includes a receiver configured to detect at least a portion of the scanned pulses of light scattered by a target located a distance from the lidar system.

Abstract: A lidar system with improved signal-to-noise ratio in the presence of solar background noise. The lidar system can comprise a light source to emit light toward a target. The light source can have an operating wavelength which lies within a band that delineates a relative maximum in atmospheric absorption. The lidar system can also include a detector to detect scattered light from the target and a processor to determine a characteristic of the target based on a characteristic of the scattered light received at the detector.

Abstract: A method in a lidar system comprises emitting a pulse of light, detecting at least a portion of the emitted pulse of light scattered by a target located a distance from the lidar system, and determining the distance from the lidar system to the target based at least in part on a round-trip time of flight for the emitted pulse of light to travel from the lidar system to the target and back to the lidar system. The method further comprises emitting a series of pulses of light having particular pulse-frequency characteristics, detecting at least a portion of the series of emitted pulses of light scattered by the target, and comparing the pulse-frequency characteristics of the series of emitted pulses of light with corresponding pulse-frequency characteristics of the detected series of scattered pulses of light to determine a velocity of the target with respect to the lidar system.

Abstract: A computer-implemented method of determining a relative velocity between a vehicle and an object. The method includes receiving sensor data generated by one or more sensors of the vehicle configured to sense an environment following a scan pattern. The method also includes obtaining, based on the sensor data, a point cloud frame. The point cloud frame comprises a plurality of points of depth data and a time at which the depth data was captured. Additionally, the method includes selecting two or more points of the scan pattern that overlap the object. The selected points are located on or near a two-dimensional surface corresponding to the object, and the depth data for two or more of the selected points are captured at different times. The method includes calculating the relative velocity between the vehicle and the object based on the depth data and capture times associated with the selected points.

Abstract: A lidar system includes a transmitter that encodes successive transmit pulses with different pulse characteristics and a receiver that detects the pulse characteristics of each received (scattered or reflected) pulse and that distinguishes between the received pulses based on the detected pulse characteristics. The lidar system thus resolves range ambiguities by encoding pulses of scan positions in the same or different scan periods to have different pulse characteristics, such as different pulse widths or different pulse envelope shapes. The receiver includes a pulse decoder configured to detect the relevant pulse characteristics of the received pulse and a resolver that determines if the pulse characteristics of the received pulse matches the pulse characteristics of the current scan position or that of a previous scan position.

Abstract: A computer-implemented method of determining relative velocity between a vehicle and an object. The method includes receiving sensor data generated by one or more sensors of the vehicle. The one or more sensors are configured to sense an environment through which the vehicle is moving by following a scan pattern comprising component scan lines. The method also includes obtaining, based on the sensor data and by one or more processors, two or more point cloud frames representative of the environment and tracking, by the one or more processors, a point cloud object across the two or more point cloud frames. Additionally, the method includes determining, based on the tracking and by the one or more processors, a relative velocity of the point cloud object and correcting, by the one or more processors, the point cloud object based on the relative velocity of the point cloud object.

Abstract: Techniques for identifying and labeling distinct objects within 3-D images of environments in which vehicles operate, to thereby generate training data used to train models that autonomously control and/or operate vehicles, are disclosed. A 3-D image may be presented from various perspective views (in some cases, dynamically), and/or may be presented with a corresponding 2-D environment image in a side-by-side and/or a layered manner, thereby allowing a user to more accurately identify groups/clusters of data points within the 3-D image that represent distinct objects. Automatic identification/delineation of various types of objects depicted within 3-D images, automatic labeling of identified/delineated objects, and automatic tracking of objects across various frames of a 3-D video are disclosed. A user may modify and/or refine any automatically generated information. Further, at least some of the techniques described herein are equally applicable to 2-D images.

Abstract: In one embodiment, a laser system includes a seed laser configured to produce optical seed pulses. The laser system also includes a first fiber-optic amplifier configured to amplify the seed pulses by a first amplifier gain to produce a first-amplifier output that includes amplified seed pulses and amplified spontaneous emission (ASE). The laser system further includes a first optical filter configured to remove from the first-amplifier output an amount of the ASE. The laser system also includes a second fiber-optic amplifier configured to receive the amplified seed pulses from the first optical filter and amplify the received pulses by a second amplifier gain to produce output pulses. The output pulses have output-pulse characteristics that include: a pulse repetition frequency of less than or equal to 100 MHz; a pulse duration of less than or equal to 20 nanoseconds; and a duty cycle of less than or equal to 1%.

Abstract: A method for configuring a perception component of a vehicle having one or more sensors includes generating a first set of training data that includes first sensor data corresponding to a first setting of one or more sensor parameters, and an indicator of the first setting. The method also includes generating a second set of training data that includes second sensor data corresponding to a second setting of the sensor parameter(s), and an indicator of the second setting. The method further includes training the perception component, at least by training a machine learning based model using the first and second training data sets. The trained perception component is configured to generate signals descriptive of a current state of the vehicle environment by processing sensor data generated by the sensor(s), and one or more indicators indicating which setting of the sensor parameter(s) corresponds to which portions of the generated sensor data.

Abstract: A method for controlling a first sensor configured to sense an environment through which a vehicle is moving includes receiving sensor data generated by one or more sensors of the vehicle as the vehicle moves through the environment, identifying, by one or more processors and based on at least a portion of the received sensor data, one or more road portions along which the vehicle is expected to travel, and determining, by one or more processors, a configuration of the identified road portions, at least in part by determining a slope of at least one of the identified road portions. The method also includes determining, by one or more processors analyzing at least the determined configuration, an elevation of a field of regard of the first sensor that satisfies one or more visibility criteria, and causing the first sensor to be adjusted in accordance with the determined elevation.

Abstract: A vehicle includes a body with multiple interior and exterior surfaces and a lidar system including a set of one or more sensor units. Each sensor unit includes a light source configured to emit light, a scanner configured to direct the emitted light to scan a field of regard of the sensor unit according to a scan pattern, a receiver configured to detect the light scattered by one or more remote targets, and a housing enclosing the light source, the scanner, and the receiver. The housing of each sensor unit is embedded in one of the interior or exterior surfaces, so that a first portion of the housing projects out of the body and a second portion of the housing is inside the body.

Abstract: A computer-readable medium stores instructions executable by one or more processors to implement an aggregate self-driving control architecture (SDCA) for controlling an autonomous vehicle. The aggregate SDCA includes a plurality of SDCAs each including a different motion planner. Each motion planner is configured to receive signals descriptive of a current state of an environment through which the autonomous vehicle is moving, and each SDCA is configured to generate candidate decisions for controlling the autonomous vehicle by using the respective motion planner to process the received signals. The aggregate SDCA also includes a decision arbiter configured to receive the candidate decisions generated by the SDCAs, generate decisions for controlling the autonomous vehicle by processing the candidate decisions, and provide signals indicative of the generated decisions to one or more operational subsystems of the vehicle to effectuate maneuvering of the vehicle.

Abstract: A computer-implemented method of determining relative velocity between a vehicle and an object. The method includes receiving sensor data generated by one or more sensors of the vehicle. The one or more sensors are configured to sense an environment through which the vehicle is moving by following a scan pattern comprising component scan lines. The method also includes obtaining, based on the sensor data and by one or more processors, two or more point cloud frames representative of the environment and tracking, by the one or more processors, a point cloud object across the two or more point cloud frames. Additionally, the method includes determining, based on the tracking and by the one or more processors, a relative velocity of the point cloud object and correcting, by the one or more processors, the point cloud object based on the relative velocity of the point cloud object.

Abstract: Post-processing in a lidar system may be guided by camera information as described herein. In one embodiment, a camera system has a camera to capture images of the scene. An image processor is configured to classify an object in the images from the camera. A lidar system generates a point cloud of the scene and a modeling processor is configured to correlate the classified object to a plurality of points of the point cloud and to model the plurality of points as the classified object over time in a 3D model of the scene.