Waymo

The technology relates to cargo vehicles. National, regional and/or local regulations set requirements for operating cargo vehicles, including how to distribute and secure cargo, and how often the cargo should be inspected during a trip. However, such regulations have been focused on traditional human-driven vehicles. Aspects of the technology address various issues involved with securement and inspection of cargo before a trip, as well as monitoring during the trip so that corrective action may be taken as warranted. For instance, imagery and other sensor information may be used to enable proper securement of cargo before starting a trip. Onboard sensors along the vehicle monitor the cargo and securement devices/systems during the trip to identify issues as they arise. Such information is used by the onboard autonomous driving system (or a human driver) to take corrective action depending on the nature of the issue.

The disclosed systems and techniques are directed to tracking and predicting live availability of parking resources by a fleet of vehicles. The disclosed techniques include receiving communications vehicles of a fleet with identification of object(s) located at an edge of a driving environment and edge visibility data for the edge from sensing systems of the vehicles, updating a map of live parking space occupancy for the driving environment with the received identification and the edge visibility data, determining, based on the updated map of live parking space occupancy and a historical parking space availability, a likelihood value associated with a parking space in the driving environment remaining unoccupied within a target time, and directing a vehicle of the fleet to the parking space based at least on the likelihood value.

A light detection and ranging (lidar) device may be coupled to a vehicle and configured to scan a surrounding environment to determine ranges to one or more objects in the surrounding environment of the vehicle. The lidar device may generate data that can be used to form a range image, which includes or is based on range data determined for the one or more objects. The lidar device may also generate data that can be used to form a corresponding background image, which includes background light intensity data that the lidar device measures during the scan. The background image or background image data may be used to add range data to the range image, correct range data in the range image, and/or evaluate the quality of the range data in the range image. In this way, the background image or background image data can be used to generate an enhanced range image that includes range data that is more comprehensive and/or more reliable than the range data included in the original range image.

Example embodiments relate to camera arrangements for vehicular object detection and avoidance. An example system includes a vehicle and at least one camera of a first camera type attached to the vehicle. The system also includes a plurality of cameras of a second camera type attached to the vehicle. Further, the system includes a plurality of cameras of a third camera type attached to the vehicle. Moreover, the system includes a computing device communicatively coupled to the at least one camera of the first camera type, plurality of cameras of the second camera type, and plurality of cameras of the third camera type. The computing device is configured to identify objects located within a first range of distances from the vehicle, objects located within a second range of distances from the vehicle, and objects located within a third range of distances from the vehicle.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining respective sensor data captured by one or more sensors of an autonomous vehicle at each of a sequence of time steps, the sequence of time steps comprising one or more context time steps followed by one or more prediction time steps; generating respective ground truth birds-eye-view (BEV) representations of the respective sensor data for each of the prediction time steps; for each prediction time step, processing the respective sensor data at one or more preceding time steps in the sequence using a future prediction neural network to generate a predicted BEV representation for the prediction time step; and training the future prediction neural network based on, for each prediction time step, an error between the ground truth BEV representation for the prediction time step and the predicted BEV representation for the prediction time step.

The disclosed systems and techniques are directed to identifying and responding to presence of vulnerable road users (VRUs) in driving environments that are at risk of loss of control of their driving trajectories. The techniques include collecting, using a sensing system of an autonomous vehicle, sensing data for an environment of the autonomous vehicle and processing the sensing data by one or more machine learning models to identify a plurality of reference points associated with a VRU in the environment. The techniques further include identifying one or more height differentials for the plurality of reference points, determining that the VRU is at risk of loss of control, based at least on a change of the one or more height differentials, and causing a control system of the autonomous vehicle to perform an avoidance action.

The described aspects and implementations enable reduced use in computational resources by an autonomous vehicle (AV) by predicting third-party collisions for an AV. A method includes obtaining object indications for objects in a driving environment of an AV. An object indication may include a shape definition and one or more predicted future locations of a corresponding object. The method includes projecting each shape definition onto each predicted future location of the one or more predicted future locations of the corresponding object. The method includes determining, based on projected shape definitions of the objects, that the projected shape definitions of at least two objects overlap. The method includes, responsive to the overlap between at least one of the projected shape definitions meeting a collision criterion, modifying the operation of the AV to avoid an area of the overlapped projected shape definitions of the least two objects.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing a prediction task using sensor data. The method includes obtaining scene data characterizing a scene in an environment at a current time point, wherein the scene comprises an autonomous vehicle and a plurality of agents, wherein the scene data comprises sensor data captured by one or more sensors of the autonomous vehicle and scene context data; generating, from the sensor data using a multimodal language model (MLM) neural network, one or more text outputs that each describe one or more aspects of the scene; generating, from at least the one or more text outputs describing the one or more aspects of the scene and the scene context data, a prediction input to a prediction neural network; and processing the prediction input using the prediction neural network to generate a prediction output for the prediction task.

The disclosed systems and techniques are directed to identifying and responding to presence of target objects in occluded areas of driving environments. The techniques include training, using perception data associated with a first driving scene, a first machine learning model (MLM) to determine a location, within the first driving scene, of a target object masked with a masking transformation. The techniques further include training, using an output of the first MLM for a training driving scene, a second MLM to generate a map of probabilities of one or more target objects to be in an occluded region of the training driving scene, the training driving scene comprising at least one of the first driving scene or a second driving scene, and causing the second MLM to be deployed on an autonomous vehicle.

Aspects of the disclosure relate to a method of managing a fleet of autonomous vehicles providing trip services. The method includes receiving information identifying an intermediate destination and a final destination for a trip 900. In this example, the intermediate destination 920 is a destination where an autonomous vehicle 100 will drop off and wait for a passenger in order to continue the trip, and the final destination 930 is a destination where the trip ends. The method also includes determining an amount of waiting time the vehicle is likely to be waiting for the passenger at the intermediate destination, determining how a vehicle of the fleet of autonomous vehicles should spend the amount of waiting time, and sending an instruction to the vehicle, based on the determination of how the vehicle should spend the amount of waiting time.