Abstract: Systems, methods, and computer-readable media are disclosed for identifying light flares in images. An example method may involve receiving an image from an imaging device, the image including data indicative of a flare artifact originating from a region of the image. The example method may also involve determining, based on the image data, a first array of pixels extending radially outwards from the region and a second array of pixels extending radially outwards from the region. The example method may also involve creating, based on the image data, a flare array, the flare array including the first array of pixels and the second array of pixels. The example method may also involve determining, based on the flare array, a peak flare artifact value indicative of a size of the flare artifact. The example method may also involve determining, based on the peak flare artifact value, a flare artifact score for the imaging device.

Abstract: Systems, methods, and computer-readable media are disclosed for a systems and methods for intra-shot dynamic LIDAR detector gain. One example method my include emitting, by an optical ranging system at a first time, a first light pulse. The example method may also include increasing, after the first time, a sensitivity of a photodetector of the optical ranging system from a first sensitivity at the first time to a second sensitivity at a second time. The example method may also include decreasing the sensitivity of the photodetector of the optical ranging system from the second sensitivity at third time to the first sensitivity at a fourth time, wherein the fourth time is after the photodetector receives return light based on the first light pulse. The example method may also include emitting, by the optical ranging system at the fourth time, a second light pulse.

Abstract: Devices, systems, and methods are provided for classifying detected objects as static or dynamic. A device may determine first light detection and ranging (LIDAR) data associated with a convex hull of an object at a first time, and determine second LIDAR data associated with the convex hull at a second time after the first time. The device may generate, based on the first LIDAR data and the second LIDAR data, a vector including values of features associated with the first convex hull and the second convex hull. The device may determine, based on the vector, a probability that the object is static. The device may operate a machine based on the probability that the object is static.

Abstract: Devices, systems, and methods are provided for enhanced multispectral sensor calibration. A device may include a first layer having copper, a second layer having solder material, the second layer above the first layer, and a third layer having a white silkscreen material, the third layer above the second layer. Regarding the device, the first layer may be used for calibration of a thermal sensor, the second layer may be used for calibration of an image sensor and calibration of a light detection and ranging (LIDAR) sensor, and the third layer may be used for the calibration of the image sensor and the calibration of the LIDAR sensor.

Abstract: Devices, systems, and methods are provided for testing and validation of a camera. A device may determine a content object placed in a line of sight of a camera device, wherein the content object provides informational and visual content. The device may capture one or more images of the content object. The device may cause an electromagnetic event using an electromagnetic interference device, causing an electromagnetic event to affect an image quality of at least one of the one or more images captured by the camera device. The device may assign a structural similarity index (SSIM) score to the at least one image of the one or more images, wherein the SSIM score indicates a camera validation status of the camera device. The device may assign an SSIM score between on one image taken without the presence of interference and one image taken in the presences of the interference.

Abstract: Devices, systems, and methods are provided for enhanced sensor cleaning validation. A device may determine a baseline performance measurement associated with a clean performance baseline of a sensor. The device may actuate a cleaning mechanism to remove at least a portion of an obstruction deposited on the sensor. The device may determine a first post-clean performance measurement associated with the sensor. The device may determine a degradation measurement between the baseline performance measurement and the first post-clean performance measurement, wherein the degradation measurement indicates an effectiveness of the cleaning mechanism.

Abstract: Systems, methods, and computer-readable media are disclosed for a vibrated polarizing beam splitter for improved return light. An example method may involve emitting, by an emitter, a first light pulse. The example method may also involve reflecting, by a polarizing beam splitter in a first position, the first light pulse, wherein the polarizing beam splitter is at a first angle of incidence in the first position. The example method may also involve adjusting, subsequent to the polarizing beam splitter reflecting the first light pulse, a position of the polarizing beam splitter from the first position to a second position, wherein the polarizing beam splitter is at a second angle of incidence in the second position. The example method may also involve transmitting, by the polarizing beam splitter, a return light pulse through the polarizing beam splitter, the return light pulse based on the first light pulse. The example method may also involve detecting, by a detector, the return light pulse.

Abstract: System and methods are disclosed herein for preventing debris accumulation and/or for removing debris about a lens of a sensor of a vehicle. The sensor may include one or more LIDAR units disposed about the vehicle. An airflow device may be disposed about the lens of the sensor. The airflow device may be configured to provide an airflow about the lens of the sensor. The airflow provided to the lens may be a constant flow of air about the lens at a low pressure and/or bursts of air about the lens at a higher pressure.

Abstract: Systems, methods, and computer-readable media are disclosed for systems and methods for improved LIDAR return light capture efficiency. One example method may include emitting, by an emitter, a first light pulse in a first path. The example method may also include transmitting, by a polarizing beam splitter in the first path and aligned with an aperture of a reflective element, a portion of the first light pulse, wherein the reflective element is disposed of in the first path. The example method may also include reflecting, by a reflective surface of the reflective element, a second light pulse in a second path, the second light pulse including a return pulse based on the first light pulse being reflected from an object. The example method may also include detecting, by a detector, the detector disposed in the second path of the second light pulse.

Abstract: Devices, systems, and methods are provided for enhanced sensor alignment. A device may determine a first array of displacement sensors proximate to a first test structure. The device may determine a second array of displacement sensors proximate to a second test structure. The device may apply a test condition to the first array, the second array, the first test structure, and the second test structure. The device may collect a first output from applying the test condition to the first test structure. The device may collect a second output from applying the test condition to the second test structure. The device may generate a first deviation vector associated with the first output. The device may generate a second deviation vector associated with the second output. The device may determine a first design status of the first structure based on the first deviation vector. The device may determine a second design status of the second structure based on the second deviation vector.

Abstract: A GmAPD FPA having increased tolerance optical overstress includes a limit resistor that is monolithically integrated into each pixel in the FPA, and which limits the magnitude of the current entering the read out integrated circuit.

Abstract: A node is provided for capturing information about moving objects at an intersection. The node includes a plurality of first cameras that are positioned to capture first digital images of an intersection from different fields of view and a second camera positioned to capture second digital images in a field of view that is wider than that of each first camera. The node includes a processor that detects in the first and second digital images a set of objects of interest of the intersection, determines motion of each detected object of interest in the set from consecutive images of the first digital images or the second digital images. The node generates, for each object of interest of the set, augmented perception data that includes location data in the global coordinate system and the determined motion of each object of interest in the set.

Abstract: To determine a cause of a fault in a robotic system, a diagnostic service of the robotic system will receive primary signals from various processes running on the robotic system. The service will access a graph representation of functions as stored in memory, and the service will perform the functions on one or more of the primary signals to yield one or more derived signals. A sink of the robotic system will subscribing to a causal trace that includes a value for a specified one of the derived signals and an identification of each signal from which the specified derived signal was derived. During runtime, the sink will receive updates to the causal trace as the value for the specified derived signal changes. The diagnostic service will use the causal trace to identify a process that caused the fault.

Abstract: Systems and methods are provided for identifying and representing a traffic signal device. The method includes determining a location and orientation of the vehicle and receiving a real world image. The method further includes analyzing information about the vehicle's location and environment and using this information and the vehicle's orientation to generate a raster image illustrating an approximation of a view of the real world image, including one or more traffic signal devices. Additionally, the method includes providing the real world image and the raster image as inputs to a neural network to classify a traffic signal device in the real world image as the primary traffic signal device and determine a set of coordinates indicating a location of the primary traffic signal device, generating a classified real world image which includes a bounding box indicating the set of coordinates, and receiving the classified real world image.

Abstract: A system detects multiple instances of an object in a digital image by receiving a two-dimensional (2D) image that includes a plurality of instances of an object in an environment. For example, the system may receive the 2D image from a camera or other sensing modality of an autonomous vehicle (AV). The system uses a first object detection network to generate a plurality of predicted object instances in the image. The system then receives a data set that comprises depth information corresponding to the plurality of instances of the object in the environment. The data set may be received, for example, from a stereo camera of an AV, and the depth information may be in the form of a disparity map. The system may use the depth information to identify an individual instance from the plurality of predicted object instances in the image.

Abstract: Systems/methods for controlling an Autonomous Vehicle (“AV”). The methods comprise: generating LiDAR datasets measuring a distance from AV to an object at different times; plotting each LiDAR dataset on a 3D graph; defining an amodal cuboid on each 3D graph; selecting different positions for amodal cuboid in 3D graphs; determining coordinates for an amodal cuboid center point at different positions within 3D graphs; using Center Point Coordinates (“CPCs”) to compute velocity and acceleration values for the object at each time; using CPCs to compute Single Frame Score (“SFS”) values; using velocity values, acceleration values and SFS values to compute scores for different cuboid positions in 3D graphs; using the scores as inputs to a Viterbi algorithm for determining a most likely sequence of states for the object given the LiDAR datasets; and using the most likely sequence of states for the object by AV to facilitate autonomous driving operation(s).

Abstract: A method for calibration of vision sensors includes, by a processor: selecting a calibration sequence that has a base calibration pattern and calibration angles, generating a calibrating target that may include the calibration pattern at a selected calibration angle of the angles, and causing a display screen to display a digital image representative of the target at the selected angle relative to an originating border of the selected angle. Each of the calibration angles is associated with a different originating border of the screen. The method includes by a vision sensor capturing at least one image of the calibrating target displayed on the screen. The calibration is repeated for each calibration angle of the sequence. The method includes performing calibration of the vision sensor in response to image signal processing of extracted calibration features in the at least one image of the calibrating targets.

Abstract: Systems and methods for operating a vehicle. The methods comprise: generating, by a computing device, a vehicle trajectory for the vehicle that is in motion; detecting an object within a given distance from the vehicle; generating, by the computing device, at least one possible object trajectory for the object which was detected; using, by the computing device, the vehicle trajectory and the at least one possible object trajectory to determine whether there is an undesirable level of risk that a collision will occur between the vehicle and the object; modifying, by the computing device, the vehicle trajectory when a determination is made that there is an undesirable level of risk that the collision will occur; and automatically causing the vehicle to move according to the modified vehicle trajectory.

Abstract: A method and a system for generating a mesh representation of a surface. The method includes receiving a three-dimensional (3D) point cloud representing the surface, identifying and discarding one or more outliers in the 3D point cloud to generate a filtered point cloud using a Gaussian process, adding one or more additional points to the filtered point cloud to generate a reconstruction dataset, and using Poisson surface reconstruction to generate an implicit surface corresponding to the surface from the reconstruction dataset.

Abstract: A robotic system simultaneously monitors multiple processes running on the robotic system in an efficient manner that can help reduce communication and processing resource requirements. A diagnostic service of the robotic system receives primary signals from multiple tasks operating in the robotic system. For each of the primary signals, the service performing a first instance of a function on the primary signal to create a first derived signal for the primary signal. For each of the primary signals that is a keyed signal and associated with a multivalent key, the service will create an additional instance of the function to create an additional derived signal for each additional valence. The service will then using the each instance of the function to create an aggregated signal, and it will use the aggregated signal to simultaneously monitor each of the processes running on the robotic system.