Abstract: Provided are methods for deep learning-based camera calibration, which can include receiving first and second images captured by a camera, processing the first image using a first neural network to determine a depth of the first image, processing the first image and the second image using a second neural network to determine a transformation between a pose of the camera for the first image and a pose of the camera for the second image, generating a projection image based on the depth of the first image, the transformation of the pose of the camera, and intrinsic parameters of the camera, comparing the second image and the projection image to determine a reprojection error, and adjusting at least one of the intrinsic parameters of the camera based on the reprojection error. Systems and computer program products are also provided.

Abstract: Among other things, planning a motion of a machine having moving capabilities is based on strategic guidelines derived from various basic principles, such as laws, ethics, preferences, driving experiences, and road environments.

Abstract: Among other things, stored data is maintained indicative of potential stopping places that are currently feasible stopping places for a vehicle within a region. The potential stopping places are identified as part of static map data for the region. Current signals are received from sensors or one or more other sources current signals representing perceptions of actual conditions at one or more of the potential stopping places. The stored data is updated based on changes in the perceptions of actual conditions. The updated stored data is exposed to a process that selects a stopping place for the vehicle from among the currently feasible stopping places.

Abstract: Among other things, techniques are described for receiving, from at least one sensor of a vehicle, sensor data associated with a surface along a path to be traveled by a vehicle; using a surface classifier to determine a classification of the surface based on the sensor data; determining, based on the classification of the surface, drivability properties of the surface; planning, based on the drivability properties of the surface, a behavior of the vehicle when driving near the surface or on the surface; and controlling the vehicle based on the planned behavior.

Abstract: Provided are systems and methods for a deep learning based beam control. Sensor data associated with the environment and the corresponding detected objects from a perception system are obtained. Object features and image features are extracted. The extracted object features and image features are fused into fused features. A beam control status is predicted according to the fused features, wherein the beam control status indicates a high beam illumination intensity or a low beam illumination intensity of a light emitting device.

Abstract: Disclosed are systems and techniques for using blockchain technology to maintain and validate a vehicle ledger. The technique includes receiving, at a master node in the system, a request to update a vehicle ledger associated with a first vehicle node comprising the system. If first criteria are met, the system updates the vehicle ledger, including: generating an updated version of the vehicle ledger using vehicle data stored in a master ledger associated with the master node, and transmitting the updated version of the vehicle ledger to the first vehicle node. If the first criteria are not met, the system forgoes updating the vehicle ledger. The vehicle data corresponds to a first vehicle associated with the first vehicle node. The master ledger is implemented using a blockchain that contains vehicle records for vehicles associated with the system. The blockchain includes a first block including vehicle data corresponding to the first vehicle.

Abstract: Among other things, techniques are described for steering angle calibration. An autonomous vehicle receives a steering angle measurement and a yaw rate measurement, and estimates a steering angle offset using the steering angle measurement, the yaw rate measurement, and a wheel base of the autonomous vehicle. An estimated yaw rate is determined based on a yaw rate model, the steering angle measurement and the estimated steering angle offset. The yaw rate measurement and the estimated yaw rate are compared and an action is initiated on the autonomous vehicle in response to the comparing.

Abstract: Among other things, a world model is maintained of an environment of a vehicle. A hypothetical object in the environment that cannot be perceived by sensors of the vehicle is included in the world model.

Abstract: Among other things, techniques are described for self-cleaning door handles. A door handle for a vehicle includes a housing including a hand grip; a hygienic film that is formed around a portion of a circumference of the hand grip; a plurality of cylindrical rollers rotatably connected to the housing and are configured to rotationally guide the hygienic film along the portion of the circumference of the hand grip; and a light that is configured to emit electromagnetic radiation in an ultraviolent spectrum and configured to irradiate at least one portion of the hygienic film to reduce a quantity of bacteria on the at least one portion of the hygienic film.

Abstract: Enclosed are embodiments for scenario-based behavior specification and validation.

Abstract: This disclosure relates in general to systems and methods for tracking objects proximate an autonomous vehicle. In particular, an object tracking system capable of re-identifying objects it has temporarily lost line of sight to is described. Re-identification of the objects allows earlier object detections to be used more effectively to predict motion likely to be taken by the objects.

Abstract: Among other things, an apparatus includes a processor, and storage for instructions executable by the processor to, in connection with a trip of a person in an autonomous vehicle, select a specific location at which the person will be picked up for the trip or a specific location at a destination of the trip, and present through a user interface of a device visible information that depicts the specific location.

Abstract: Provided are methods for agent prioritization, which can include determining a primary agent set and generating, based on the primary agent set, a trajectory for the autonomous vehicle. Some methods described also include determining an interaction parameter of agents in the environment. Systems and computer program products are also provided.

Abstract: Provided are methods for object tracking, which can include receiving sensor data characterizing respective detected objects. The methods can also include generating a data structure based on the data characterizing the respective detected objects. The data structure can include a graph of nodes representing states of the objects and edges representing hypothetical transitions in states of the objects. The methods can also include applying a predictive model to the data structure. The predictive model can be trained to receive the state as inputs and produce an identification of a set of nodes and edges corresponding to the one of the respective detected objects. The methods can further include providing data based on the identification of the set of nodes and edges to a planning system of the vehicle and causing the vehicle to operate based on providing the data. Systems and computer program products are also provided.

Abstract: Among other things, we describe techniques for operation of a vehicle based on measured load characteristics and/or passenger comfort. One or more sensors of the vehicle can measure passenger data and/or load data of the vehicle. The passenger data and/or load data of the vehicle can be used by the vehicle to determine how to navigate within the surrounding environment.

Abstract: Provided are methods for generating future maneuver parameters in a planner, which include receiving at least one first set of parameters associated with one or more previous maneuvers of a vehicle and at least one second set of parameters associated with a maneuver goal of the vehicle, generating, using the first and second sets of parameters, a future maneuver parameter corresponding to a future maneuver of the vehicle, training at least one data model by comparing the generated future maneuver parameter to one or more reference maneuver parameters, generating, based on training, a corrected future maneuver parameter. The corrected future maneuver parameter includes a future maneuver of the vehicle and a correction to the future maneuver of the vehicle. The first set of parameters includes the generated corrected future maneuver parameter, which is used to correct at least one first parameter. The training is executed using the corrected first parameter.

Abstract: Provided are methods for augmenting data related to generation of vehicle trajectories, which include predicting, using a machine learning model, a first trajectory of a vehicle at a first time in an environment surrounding the vehicle and including at least one object, detecting a deviation of the predicted first trajectory at the first time from a first ground truth trajectory of the vehicle and determining that, at the first time, the deviation satisfies a threshold, predicting, using the machine learning model, a second trajectory of the vehicle based on the predicted first trajectory of the vehicle and a second ground truth trajectory of at least one object at a second time being subsequent to the first time, and generating a training dataset for training the machine learning model using the predicted first and second trajectories of the vehicle. Systems and computer program products are also provided.

Abstract: Provided are methods for counterexample-guided update of a motion planner, which may include training a machine learning model to generate a trajectory for a vehicle. The training may generate a first trained machine learning model having a first plurality of feature weights. The machine learning model may be trained based on a counterexample for which the first trained machine learning model fails to generate a correct trajectory for the vehicle. The training may generate a second trained machine learning model having a second plurality of feature weights. A third plurality of feature weights may be determined by updating the first plurality of feature weights based on the second plurality of feature weights. The machine learning model may be updated to generate the trajectory of the vehicle by applying the third plurality of feature weights. Systems and computer program products are also provided.

Abstract: Provided are systems and methods for camera alignment using pre-distorted targets. Some methods described include selecting a configuration of shapes, and determining targets by pre-distorting the shapes according to the inverse of the distortion function of the lens system to be aligned. Images of pre-distorted targets are then compared to the original configuration of shapes, to perform camera alignment. Alignment is thus accomplished in simpler and more accurate manner. Systems and computer program products are also provided.

Abstract: Techniques are provided for improving a perception processing pipeline for object detection that fuses image segmentation data (e.g., segmentation scores) with LiDAR points. The disclosed techniques are implemented using an architecture that accepts point clouds and images as input and estimates oriented 3D bounding boxes for all relevant object classes. In an embodiment, a method comprises: matching temporally, using one or more processors of a vehicle, points in a three-dimensional (3D) point cloud with an image; generating, using an image-based neural network, semantic data for the image; decorating, using the one or more processors, the points in the 3D point cloud with the semantic data; and estimating, using a 3D object detector with the decorated points as input, oriented 3D bounding boxes for the one or more objects.