Abstract: The described aspects and implementations enable efficient detection and classification of objects with machine learning models that deploy a bird's-eye view representation and are trained using depth ground truth data. In one implementation, disclosed are system and techniques that include obtaining images, generating, using a first neural network (NN), feature vectors (FVs) and depth distributions pixels of images, wherein the first NN is trained using training images and a depth ground truth data for the training images. The techniques further include obtaining a feature tensor (FT) in view of the FVs and the depth distributions, and processing the obtained FTs, using a second NN, to identify one or more objects depicted in the images.

Abstract: The described aspects and implementations enable efficient object detection and tracking. In one implementation, disclosed is a method and a system to perform the method, the system including the sensing system configured to obtain sensing data characterizing an environment of the vehicle. The system further includes a data processing system operatively coupled to the sensing system and configured to process the sensing data using a first (second) set of neural network (NN) layers to obtain a first (second) set of features for a first (second) region of the environment, the first (second) set of features is associated with a first (second) spatial resolution. The data processing system is further to process the two sets of features using a second set of NN layers to detect a location of obj ect(s) in the environment of the vehicle and a state of motion of the object(s).

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network using contrastive learning. One of the methods includes obtaining a network input representing an environment; processing the network input using a first subnetwork of the neural network to generate a respective embedding for each location in the environment; processing the embeddings for each location in the environment using a second subnetwork of the neural network to generate a respective object prediction for each location; determining, for each of a plurality of pairs of the plurality of locations in the environment, whether the respective object predictions of the pair of locations characterize the same possible object or different possible objects; computing a respective contrastive loss value for each of the plurality of pairs of locations; and updating values for a plurality of parameters of the first subnetwork using the computed contrastive loss values.

Abstract: A method includes obtaining, by a processing device, input data derived from a set of sensors associated with an autonomous vehicle (AV), extracting, by the processing device from the input data, a plurality of sets of features, generating, by the processing device using the plurality of sets of features, a fused bird's-eye view (BEV) grid. The fused BEV grid is generated based on a first BEV grid having a first scale and a second BEV grid having a second scale different from the first scale. The method further includes providing, by the processing device, the fused BEV grid for object detection.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network using contrastive learning. One of the methods includes obtaining a network input representing an environment; processing the network input using a first subnetwork of the neural network to generate a respective embedding for each location in the environment; processing the embeddings for each location in the environment using a second subnetwork of the neural network to generate a respective object prediction for each location; determining, for each of a plurality of pairs of the plurality of locations in the environment, whether the respective object predictions of the pair of locations characterize the same possible object or different possible objects; computing a respective contrastive loss value for each of the plurality of pairs of locations; and updating values for a plurality of parameters of the first subnetwork using the computed contrastive loss values.

Abstract: Systems and techniques for pedestrian path predictions are disclosed herein. For example, an environment, features of the environment, and pedestrians within the environment may be identified. Models for the pedestrians may be generated based on features of the environment. A model may be indicative of goals of a corresponding pedestrian and predicted paths for the corresponding pedestrian. Pedestrian path predictions for the pedestrians may be determined based on corresponding predicted paths. A pedestrian path prediction may be indicative of a probability that the corresponding pedestrian will travel a corresponding predicted path. Pedestrian path predictions may be rendered for the predicted paths, such as using different colors or different display aspects, thereby enabling a driver of a vehicle to be presented with information indicative of where a pedestrian is likely to travel.

Abstract: Systems and techniques for pedestrian path predictions are disclosed herein. For example, an environment, features of the environment, and pedestrians within the environment may be identified. Models for the pedestrians may be generated based on features of the environment. A model may be indicative of goals of a corresponding pedestrian and predicted paths for the corresponding pedestrian. Pedestrian path predictions for the pedestrians may be determined based on corresponding predicted paths. A pedestrian path prediction may be indicative of a probability that the corresponding pedestrian will travel a corresponding predicted path. Pedestrian path predictions may be rendered for the predicted paths, such as using different colors or different display aspects, thereby enabling a driver of a vehicle to be presented with information indicative of where a pedestrian is likely to travel.