Abstract: The present disclosure generally relates to an object detection system. For example, aspects of the present disclosure relate to systems and techniques for performing object detection using sensor information, such as elevation and/or velocity information from one or more light-based sensors. One example apparatus generally includes one or more processors operably configured to: obtain sensor information indicating at least two objects in an environment; determine at least one of a velocity or an elevation associated with each object of the at least two objects; consolidate the at least two objects into a common object based on the at least one of the velocity or the elevation; and output an indication of the common object.

Abstract: An example device for processing image data includes a memory configured to store image data; and one or more processors implemented in circuitry and configured to: determine a set of keypoints representing objects in an image of the image data captured by a camera of a vehicle; determine depth values for the objects in the image; determine positions of the objects relative to the vehicle using the set of keypoints and the depth values; and at least partially control operation of the vehicle according to the positions of the objects. For example, the depth values may represent descriptors for the keypoints or be used to determine the descriptors for the keypoints.

Abstract: Systems and methodologies for determining validity of a location of a vehicle include obtaining a first location of the vehicle corresponding to a first time. A first RSU signal including an indication of a location of the first RSU is received at the vehicle from a first Roadside Unit (RSU). One or more location measurements are obtained of the vehicle relative thee first RSU based on one or more sensors of the vehicle. The first location of the vehicle is compared with the first RSU-based location of the vehicle.

Abstract: Disclosed are systems, apparatuses, processes, and computer-readable media for processing image data. For example, an apparatus can compute initial embeddings from a plurality of images. The apparatus can construct a graph comprising nodes representing the initial embeddings. The apparatus can further perform, based on the graph, a plurality of message passing steps successively to generate final embeddings. The apparatus can classify, using a classification engine, one or more objects in each of the plurality of images based on the final embeddings. The apparatus can further compute a classification loss based on the classifying of the one or more objects.

Abstract: Systems and techniques are provided for processing sensor data. For example, a process can include obtaining input data and processing, using a non-machine learning based feature detector feature detector, the input data to determine one or more feature points in the input data. The process can further include determining, using a machine learning system, a respective feature descriptor for each respective feature point of the one or more feature points.

Abstract: A method includes determining a current state of an environment of an autonomous agent, such as a vehicle. The method also includes determining, via a first neural network, a set of actions based on the current state. The method further includes determining whether further analysis of the set of actions is desired. The method selects an action from the set of actions using a model-based solution based on a reward and a risk of the action when further analysis is desired. The method also includes selecting the action from the set of actions according to a metric when further analysis is not desired. The method controls the autonomous agent to perform the selected action.

Abstract: Apparatus, methods, and computer-readable media for facilitating a cloud-based vehicle XR user experience are disclosed herein. An example method for wireless communication at a user equipment (UE) includes transmitting a request for a vehicle extended reality (XR) session. The vehicle XR session may be based on a first user XR stream including a vehicle XR component associated with a vehicle and a first user XR component associated with a first user. The first user may have an association with the vehicle. The example method also includes transmitting uplink information associated with the first user XR stream. The example method also includes receiving rendering information associated with the first user XR stream. The rendering information may be based on the uplink information.

Abstract: Techniques disclosed provide for enhanced V2X communications by defining information Elements (IE) for V2X messaging between V2X entities. For a transmitting vehicle that sends a V2X message to a receiving vehicle, these IEs are indicative of a detected vehicle model type detected by the transmitting vehicle of a detected vehicle; a pitch rate of the transmitting vehicle, a detected vehicle, or a detected object; a roll rate of the transmitting vehicle, a detected vehicle, or a detected object; a yaw rate of a detected vehicle, or a detected object; a pitch rate confidence; a roll rate confidence; an indication of whether a rear brake light of a detected vehicle is on; or an indication of whether a turning signal of a detected vehicle is on; or any combination thereof. With this information, the receiving vehicle is able to make more intelligent maneuvers than otherwise available through traditional V2X messaging.

Abstract: Methods, system, non-transitory media, and devices for supporting safety compliant computing in a heterogeneous computing system, such as a vehicle heterogeneous computing system are disclosed. Various aspects include methods enabling a vehicle, such as an autonomous vehicle, a semi-autonomous vehicle, etc., to achieve algorithm safety for various algorithms on a heterogeneous compute platform with various safety levels.

Abstract: Certain aspects of the present disclosure provide techniques for techniques for sensor data sharing. Certain aspects provide a method for wireless communication by a first wireless device. The method generally includes receiving, from one or more sensors, corresponding one or more raw sensor data sets and transmitting, to a second wireless device, at least one message comprising information regarding an object detected based on processing the one or more raw sensor data sets, wherein the information includes one or more characteristics of the object derived based on processing the one or more raw sensor data sets, wherein: when one or more conditions are met, the information further includes at least one raw sensor data set of the one or more raw sensor data sets; and when the one or more conditions are not met, the information does not include any raw sensor data.

Abstract: Techniques for autonomous uplink (AUL) transmissions are provided that allow for efficient use of shared radio frequency spectrum band resources. A user equipment (UE) may determine a duration of an AUL transmission and modify an uplink waveform or provide an indication to a base station of one or more channel resources that may be available for base station transmissions, in order to more fully utilize shared radio frequency spectrum band resources within a maximum channel occupancy time (MCOT). A base station may activate or deactivate AUL transmissions through downlink control information (DCI) transmitted to the UE. A UE and base station may exchange various other control information to provide relatively efficient autonomous uplink transmissions and use of the shared radio frequency spectrum band resources.

Abstract: Various embodiments include methods and systems for autonomous driving systems for using map data in performing an autonomous driving function. Various embodiments may include accessing map data regarding an object or feature in the vicinity of the vehicle, accessing confidence information associated with the map data, and using the confidence information in performing an autonomous or semi-autonomous driving action by the processor. The confidence information may be stored in the map database or in a separate data structure accessible by the processor. Methods of generating map safety and confidence information may include receiving information regarding a map object or feature including a measure of confidence in the information, using the received measure of confidence to generate safety and confidence information regarding the object or feature, and storing the safety and confidence information for access by vehicle autonomous and semi-autonomous driving systems.

Abstract: In some aspects, a device of a vehicle may obtain information relating to an environment in which the vehicle is located. The device may determine using a machine learning model, a driving context of the vehicle based at least in part on the information relating to the environment, and a set of hyperparameters for a model, that is used to determine a driving behavior for the vehicle, based at least in part on the driving context. The device may determine, using the model configured with the set of hyperparameters, the driving behavior for the vehicle. The device may cause autonomous operation of the vehicle in accordance with the driving behavior. Numerous other aspects are described.

Abstract: A method includes determining a current state of an environment of an autonomous agent, such as a vehicle. The method also includes determining, via a first neural network, a set of actions based on the current state. The method further includes determining whether further analysis of the set of actions is desired. The method selects an action from the set of actions using a model-based solution based on a reward and a risk of the action when further analysis is desired. The method also includes selecting the action from the set of actions according to a metric when further analysis is not desired. The method controls the autonomous agent to perform the selected action.

Abstract: The present disclosure generally relates to an object detection system. For example, aspects of the present disclosure relate to systems and techniques for performing object detection using sensor information, such as elevation and/or velocity information from one or more light-based sensors. One example apparatus generally includes one or more processors operably configured to: obtain sensor information indicating at least two objects in an environment; determine at least one of a velocity or an elevation associated with each object of the at least two objects; consolidate the at least two objects into a common object based on the at least one of the velocity or the elevation; and output an indication of the common object.

Abstract: Embodiments include methods performed by a processor of a vehicle control unit for managing a driving condition anomaly. In some embodiments, the vehicle may receive a first driving condition based on data from a first vehicle sensor, receive a second driving condition based on data from another data source, determine a driving condition anomaly based on the first driving condition and the second driving condition, send a request for information to a driving condition database, receive the requested information from the driving condition database, and resolve the driving condition anomaly based on the requested information from the driving condition database.

Abstract: Systems, methods, and devices for radio frequency (RF) ranging-aided localization and crowdsourced mapping are provided. In one aspect, a method performed by a user equipment (UE) includes obtaining sensor data comprising first radio frequency (RF) ranging data and imaging data. The method further includes tagging the first RF ranging data with location information and semantic information, wherein the semantic information is based on the imaging data, and wherein the semantic information indicates a first portion of the RF ranging data is associated with a static object type and a second portion of the RF ranging data is associated with a temporary-static object type different from the static object type. The method further includes transmitting, to a RF ranging assistance server, the first RF ranging data tagged with the location information and the semantic information.

Abstract: Techniques for autonomous uplink (AUL) transmissions are provided that allow for efficient use of shared radio frequency spectrum band resources. A user equipment (UE) may determine a duration of an AUL transmission and modify an uplink waveform or provide an indication to a base station of one or more channel resources that may be available for base station transmissions, in order to more fully utilize shared radio frequency spectrum band resources within a maximum channel occupancy time (MCOT). A base station may activate or deactivate AUL transmissions through downlink control information (DCI) transmitted to the UE. A UE and base station may exchange various other control information to provide relatively efficient autonomous uplink transmissions and use of the shared radio frequency spectrum band resources.

Abstract: System and method for processing a camera frame in a mobile device by partitioning the camera frame into different sections based on the distance from a vehicle to each of the sections and the required resolution of each of the sections. A mobile device comprises: a memory; a processor communicatively coupled to the memory, the processor configured to: receive a camera frame from a camera mounted on a vehicle traveling on a road; determine a drivable path of the vehicle; project the drivable path onto the camera frame; partition a part of the camera frame containing the drivable path into at least one section based on a distance from the vehicle to each of the at least one section; and determine a required resolution of each of the at least one section based on the distance from the vehicle to each of the at least one section.

Abstract: Various embodiments include methods and devices for training and implementing a tree-based behavior prediction model for use in autonomous vehicle control systems. Some embodiments may include labeling real-world autonomous vehicle run data to indicate an insight of the data, selecting an insight decision tree of the tree-based behavior prediction model for training using the labeled data, training the insight decision tree using the labeled data to classify a probability of an insight associated with the insight decision tree, and updating the tree-based behavior prediction model based on training the insight decision tree. Some embodiments may include selecting an insight decision tree of a tree-based behavior prediction model configured for classifying a probability of an insight associated with the insight decision tree, executing the insight decision tree, and outputting a probability of an insight determined from executing the insight decision tree using the data.