Geospatial clustering of areas using neural networks for autonomous systems and applications

Abstract

The present disclosure relates to geospatial clustering of areas for autonomous systems and applications using neural networks. In various examples, a cell model is used to determine clusters of cells that share similarities, the cell model dividing a geographic area into one or more cells. Sensor data is provided to one or more machine learning models trained to classify the sensor data into one or more cells of the cell model. Based on the classification of sensor data into cells of the cell model, similarities between pairs of cells of the cell model can be determined and used to form clusters of the cells that are sufficiently similar to help manage training data for training machine learning models to help autonomous or semi-autonomous machines in surrounding environments.

Description

Background Technology

[0001] Autonomous and semi-autonomous driving systems, as well as advanced driver assistance systems (ADAS), can use sensors, such as cameras, to form an understanding of the vehicle's surroundings. This understanding can include information about the location of objects, obstacles, road signs, road surfaces, and / or other markers. Autonomous driving systems rely on machine learning models and / or neural networks to assist in the information gathering and decision-making process, and effective machine learning models or neural networks require training using a set of real and / or synthetic training data.

[0002] The training data may include image data and / or other sensor data (e.g., lidar, radar, ultrasound, etc.) describing the environment of the autonomous machine. For example, sensor data representing sensor data representations (e.g., images, point clouds, projected images, etc.) can be used to train machine learning models intended for use with autonomous vehicles. These sensor data represent a street-level or road environment, including environmental and landscape elements that provide the “look and feel” of the environment, as well as objects such as other vehicles, road signs, obstacles, structures, or any other objects of interest to the autonomous vehicle system. Training machine learning models to produce accurate estimates, such as when detecting specific signs (e.g., road signs, road markings, etc.) or other road elements, may involve managing large training datasets where specific objects of interest (e.g., road signs) are depicted in various locations, orientations, visibility, landscapes, and / or arrangements. However, due to issues such as physical accessibility, geographical challenges, and political and / or practical reasons, certain objects of interest may appear differently in various real-world geographic regions, which can pose difficulties in managing large training datasets. Therefore, generating a sufficiently large training dataset representing a specific geographic region is either costly (e.g., challenging in terms of data processing time and labor), infeasible (e.g., in terms of accessibility), or both.

[0003] Typically, systems used to manage (curating) large datasets containing sensor data representing objects of interest for training autonomous machines to operate in a specific geographic area utilize large, unlabeled or unanalyzed datasets. These datasets may be captured by vehicles equipped with cameras and / or other sensors as part of a data collection session (e.g., images captured from the angle of a vehicle's path on a road), or sometimes by private customer vehicles operating in that geographic area. Collecting large amounts of training data can be impossible and / or impractical due to limited physical and / or practical accessibility in certain areas (e.g., cost, regulations, etc.) or a limited established market for data collection tools. For example, legal and / or financial regulations in a geographic area can incur significant costs in managing a sufficient amount of data to train a machine learning model relevant to that specific area and performing well in deployment. Summary of the Invention

[0004] Embodiments of this disclosure relate to geospatial clusters of regions using neural networks for autonomous systems and applications. Systems and methods for training and deploying machine learning models to determine geospatial region clusters are disclosed. For example, the machine learning model can be trained with training data labeled using unit models and used to classify captured sensor data into one or more units of a unit model corresponding to a geospatial region.

[0005] Compared to traditional methods (as described above), this disclosure provides a method for clustering or grouping geographic regions into semantic regions using machine learning models—such as deep neural networks (DNNs)—along with map and geospatial data. These semantic regions include geographic areas similar based on perceptual features (e.g., visual appearance of terrain, buildings, road signs, objects, etc.), without limiting semantic regions to political boundaries (e.g., national, state, provincial, municipal boundaries, etc.). The training data used to manage training data includes desired objects of interest associated with specific regions or groups of regions. Using the disclosed method, collected sensor data can be used to identify similar geographic regions (e.g., visual similarity, etc.) that can be clustered into semantic regions. Attached Figure Description

[0006] The following describes in detail, with reference to the accompanying figures, the system and method for using neural networks for regional geospatial clustering in autonomous systems and applications, wherein:

[0007] Figure 1 This is an example data flow diagram illustrating a process for identifying geospatial area clusters according to some embodiments of the present disclosure;

[0008] Figure 2These are examples of unit model descriptions for segmenting geographic regions into clusters based on road similarity, according to some embodiments of this disclosure;

[0009] Figure 3 These are examples of unit model descriptions for segmenting geographic regions into clusters based on sign similarity according to some embodiments of the present invention;

[0010] Figure 4 These are examples of annotated images with annotations applied to sensor data to identify objects, according to some embodiments of this disclosure;

[0011] Figure 5 These are examples of similarity matrices used to represent pairwise similarity of unit models according to some embodiments of this disclosure;

[0012] Figure 6-8 This is a flowchart illustrating an example method for geospatial clustering of regions using a neural network according to some embodiments of the present disclosure;

[0013] Figure 9A These are illustrations of example autonomous vehicles according to some embodiments of the present disclosure;

[0014] Figure 9B According to some embodiments of this disclosure Figure 9A Examples of camera positions and field of view for autonomous vehicles;

[0015] Figure 9C According to some embodiments of this disclosure Figure 9A A block diagram of an example system architecture for an example autonomous vehicle;

[0016] Figure 9D Cloud-based servers and according to some embodiments of this disclosure Figure 9A A system diagram illustrating communication between autonomous vehicles;

[0017] Figure 10 This is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure; and

[0018] Figure 11 This is a block diagram of an example data center suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0019] Systems and methods related to geospatial clustering of regions using neural networks for autonomous machine systems and applications are disclosed. Although this disclosure may be described with reference to an example autonomous vehicle 900 (which may be alternatively referred to herein as "vehicle 900" or "ego-machine 900"), examples are referenced. Figures 9A-9D(Description provided), but this is not intended to be limiting. For example, the systems and methods described herein can be, but are not limited to, non-autonomous vehicles, semi-autonomous vehicles (e.g., in one or more adaptive driver assistance systems (ADAS)), manned and unmanned robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, spacecraft, ships, shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, engineering vehicles, underwater vehicles, drones, and / or other vehicle types. Furthermore, while this disclosure may be described with respect to object detection and recognition for autonomous machines, this is not intended to be limiting, and the systems and methods described herein can be used in augmented reality, virtual reality, mixed reality, robotics, security and surveillance, autonomous or semi-autonomous machine applications, and / or any other technological space where object recognition or detection can be used.

[0020] Compared to traditional methods (as described above), this disclosure provides a method for clustering or grouping geographic regions into semantic regions using machine learning models—such as deep neural networks (DNNs)—along with map and geospatial data. These semantic regions include geographic areas similar based on perceptual features (e.g., visual appearance of terrain, buildings, road signs, objects, etc.), without limiting semantic regions to political boundaries (e.g., national, state, provincial, municipal boundaries, etc.). This is used to manage training data that includes desired objects of interest associated with specific regions or groups of regions. Using the disclosed method, collected sensor data can be used to identify similar geographic regions (e.g., visual similarity, etc.) that can be clustered into semantic regions.

[0021] In some embodiments, semantic regions may be associated with a cell model. A cell model may represent the division of a geographic region (e.g., the Earth's surface, continents, states, astronomical objects, etc.) into one or more cells. Cell models may be defined based on parameters such as the minimum and / or maximum number of cells per country, the total number of cells, road network density, and / or land area. For a non-limiting example, a cell model may be defined to divide the Earth into several distinct cells (e.g., 400, 500, 650, 719, etc.). In some embodiments, country, state, city, region, or other boundaries may be preserved in the cell model, and regions corresponding to a specific country may be associated with one or more constituent cells. For a non-limiting example, the United States region may be divided into multiple cells in the cell model, such as 20 or 74 cells. A cell model may be defined such that it allows a cell to be divided into one or more sub-cells. For example, a cell may be divided into multiple cells based on road network density and cell size using a binary spatial partitioning algorithm. For example, a unit can be divided into multiple units of different sizes based on the density of the road network (e.g., highways, primary roads, secondary roads, tertiary roads, local roads, etc.) within a region of multiple sub-units.

[0022] In some embodiments, sensor data (e.g., recorded by a vehicle's sensors during a data collection session) may be associated with at least one cell of a cell model. For example, the recording vehicle may drive a route while recording timestamps, video data, image data, other sensor data, and / or GNSS data, and the data may be mapped to a specific cell of the cell model. For example, sensor data may be associated with a specific cell of the cell model using GNSS data, orientation data (e.g., determined using one or more inertial measurement unit (IMU) sensors), sensor pose or mounting location data, and / or other data from drive data. In some embodiments, sensor data may undergo additional processing, such as extracting portions of an image (or other sensor data representation, such as point clouds, projected images, etc.) represented by the sensor data. For example, the sensor data representation may be cropped based on detected objects (e.g., road signs, markers, vehicles, etc.) depicted in the sensor data. In some embodiments, the size of the sensor data may be additionally or alternatively adjusted. For example, randomly selected portions of an image may be cropped to conform to a specified dimension (e.g., a spatial dimension that a machine learning model, algorithm, etc., is configured to receive as input). In some embodiments, sensor data can be modified to adjust the brightness, color, and / or contrast of the depicted sensor data representation. In some embodiments, sensor data can be tagged with metadata. For example, sensor data can be tagged with metadata tags that assign a cell identifier, indicating a specific cell of a cell model to which the sensor data may correspond. For example, using the GNSS location associated with an image, the corresponding cell of a cell model can be determined, and the image can be tagged with the cell identifier of the corresponding cell. For example, an image with a corresponding GNSS location in the United States can be tagged with the cell identifier "US12," indicating that the GNSS location belongs to cell 12 within the United States.

[0023] In at least one embodiment, the system can perform a similarity matching operation to determine the classification of sensor data (e.g., cropped image data) as cells of a cell model. For example, a deep neural network (DNN) can be used to predict which cell of the cell model a sensor data frame should belong to. The DNN can be trained using sensor data already labeled with cell identifiers, which are used as ground reality cell classifications.

[0024] In the same or additional examples, the similarity level between two units of a unit model can be determined based on the pairwise similarity of the extracted units (e.g., using a DNN) and the interpretation of uncertainty in unit classification as similarity between the two units. For example, when a DNN has uncertainty in predicting the classification of two units between sensor data frames, the two units can be interpreted as having perceptual similarity. In some embodiments, the similarity between two or more units of a unit model can be represented as a probability distribution—e.g., using a Softmax algorithm. For example, the probability distribution can indicate the probability that an image classified as a first unit can also be classified as a second unit (within a threshold range). In some embodiments, the probability distribution can be represented as a matrix indicating the pairwise similarity between units of a unit model. For example, an NxN matrix can be used to represent the pairwise similarity of a set of units of size N. In some embodiments, a similarity score can be determined for each pair of units in the unit model.

[0025] In at least one embodiment, based on pairwise similarity between units in a unit model, units can be clustered into semantic regions using distance-based clustering operations and / or algorithms (such as k-medoids or k-means clustering). A semantic region can represent a set of units in a unit model that is identified as similar units independent of geographic boundaries and / or geographic proximity. For example, units can be clustered together based on similarities extracted between units associated with Argentina, units associated with Guyana, and units associated with Mexico. In some embodiments, one or more unit clusters in the unit model can be presented. As a result, location-associated training data can be used instead of or additionally associated with data from a second location associated with a first unit, the second location associated with a second unit, and the second unit associated with the first unit. Therefore, when training a machine learning model for label detection in Guyana, training data from Argentina and Mexico can be used to increase the robustness of the training dataset while maintaining the accuracy and precision of the machine learning model due to the similarity between training data from different regions. Thus, data management resources and / or requirements can be planned, estimated, and executed more effectively across various geographic regions.

[0026] refer to Figure 1 , Figure 1This is an example data flow diagram illustrating a process 100 for identifying geospatial region clusters according to some embodiments of the present disclosure. It should be understood that such and other arrangements described herein are illustrated by way of example only. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted together. Furthermore, many of the elements described herein are functional entities that can be implemented as discrete or distributed components, or in combination with other components and implemented in any suitable combination and location. The various functions described herein as being performed by entities can be performed by hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in memory. In some embodiments, the systems, methods, and processes described herein can use... Figures 9A-9D Example of autonomous vehicle 900 Figure 10 Example computing device 1000 and / or Figure 11 The components, features, and / or functions of the example data center 1100 are similar to those of other components, features, and / or functions used to perform this task.

[0027] At a high level, process 100 may include one or more machine learning models 104 that receive one or more inputs, such as data representing objects and / or the environment detected in sensor data 102, and generate one or more outputs, such as classifying the inputs of one or more units of a unit model that can be used by similarity matrix 112. While sensor data 102 is primarily discussed in relation to image data representing images, this is not intended to be limiting, and sensor data 102 may include other types of sensor data for object detection or recognition, such as lidar data, ultrasonic or other sonar data, radar data, and / or similar data—for example, data generated by… Figures 9A-9D Example autonomous vehicle 900 generates one or more sensors.

[0028] Process 100 may include generating and / or receiving sensor data 102 from one or more sensors. As a non-limiting example, sensor data 102 may be generated from a machine (e.g., Figures 9A-9D The sensor data 102 is received from one or more sensors of the machine (900) as described herein. Sensor data 102 may include, but is not limited to, sensor data 102 from any sensor of the machine, including, for example, and referenced from… Figures 9A-9DThe sensor data 102 may include: a Global Navigation Satellite System (GNSS) sensor 958 (e.g., a Global Positioning System sensor), a RADAR sensor 960, an ultrasonic sensor 962, a LIDAR sensor 964, an Inertial Measurement Unit (IMU) sensor 966 (e.g., an accelerometer, gyroscope, magnetic compass, magnetometer, etc.), a microphone 996, a stereo camera 968, a wide-angle camera 970 (e.g., a fisheye camera), an infrared camera 972, a surround camera 974 (e.g., a 360-degree camera), a long-range and / or medium-range camera 998, a speed sensor 944 (e.g., for measuring the speed of vehicle 900), and / or other sensor types. As another example, sensor data 102 may include virtual (e.g., simulated, synthetic, or augmented) sensor data generated from any number of sensors of a virtual vehicle or other virtual object in a virtual (e.g., test) environment. In such an example, the virtual sensor may correspond to a virtual vehicle or other virtual object in a simulated environment (e.g., for testing, training, and / or validating the performance of a neural network), and the virtual sensor data may represent sensor data captured by a virtual sensor in a simulated or virtual environment. Therefore, by using virtual sensor data, simulated data from a simulated environment and / or augmented real-world data can be used to test, train, and / or validate the machine learning model 104 described herein. This may allow for testing in more extreme scenarios outside of real-world environments, where such testing might be less safe.

[0029] In some embodiments, sensor data 102 may include image data representing an image, image data representing video (e.g., a snapshot of a video), and / or sensor data representing a sensor's sensing field (e.g., a depth map of a LiDAR sensor, a value map of an ultrasonic sensor, etc.). Where sensor data 102 includes image data, any type of image data format may be used, such as, but not limited to, compressed images such as Joint Image Experts Group (JPEG) or Luminosity / Chroma (YUV) formats, compressed images derived from frames of compressed video formats (e.g., H.264 / Advanced Video Coding (AVC) or H.265 / High-Efficiency Video Coding (HEVC)), original images, such as those derived from Red-to-Blue Transparent (RCCB), Red-to-Cross (RCCC), or other types of imaging sensors and / or other formats. Furthermore, in some examples, sensor data 102 may be used within process 100 without any preprocessing (e.g., in raw or captured format), while in other examples, sensor data 102 may undergo preprocessing (e.g., noise balancing, demosaicing, scaling, cropping, enhancement, white balance, tone curve adjustment, etc., for example using a sensor data preprocessor (not shown)). As used herein, sensor data 102 may refer to unprocessed sensor data, preprocessed sensor data, or a combination thereof.

[0030] The sensor data 102 used to manage training data 120 may include raw sensor data representations (e.g., captured by one or more sensors), downsampled representations (e.g., downsampled images), upsampled representations, cropped or region of interest (ROI) representations, other augmented representations, and / or combinations thereof. The sensor data 102, training data, and / or other data may be used to train machine learning models 104, such that one or more machine learning models 104 are configured to compute classification results. Although machine learning models 104 can be trained for classification, detection, etc., when determining the similarity between units of a unit model, the similarity matrix 112 may use data (e.g., feature data) from one or more inner layers (e.g., the penultimate layer) of the machine learning model 102.

[0031] Reference object detector 110, which can be used to crop, annotate, and / or label the representation of sensor data 102. Object detector 110 can identify object position / pose / size / etc., and can be used to generate annotations in drawing programs (e.g., annotation programs), computer-aided design (CAD) programs, labeling programs, and another type of program suitable for generating annotations. In any example, annotations can be synthetically generated (e.g., generated from computer models or renderings), practically generated (e.g., designed and generated based on real-world data), machine automated (e.g., using feature analysis and learning to extract features from data, and then generating labels), human annotation (e.g., labelers or annotation experts, defining the location of labels), and / or combinations thereof (e.g., humans identify the center or origin and the dimensions of the area, machines generate labels for polygons and / or objects and / or lanes).

[0032] Object detector 110 can generate cropped images, annotations, or other label types corresponding to bounding shapes—e.g., polygons—that depict regions of interest in the environment represented by sensor data 102. In some examples, objects such as vehicles, cars, pedestrians, etc., can be depicted by one or more polygons corresponding to objects detected in sensor data 102—e.g., within the sensor data representation of sensor data 102. Polygons can be generated as bounding boxes, ellipses, and / or any other shapes that can be used to depict objects in sensor data 102. Bounding polygons can be generated randomly to capture portions of the environment represented by sensor data 102 (e.g., capturing the overall look and feel of the environment). Object detector 110 can generate annotations or other label types for each image (or other data representation) and / or for each of the one or more polygons in the images represented by sensor data 102, which can be used as input to machine learning model 104. For example, object detector 110 can generate a label associated with each polygon, the label providing information indicating a specific identifier to which the detected object corresponding to the bounding shape belongs (e.g., sign, mark, vehicle, car, truck, pedestrian, motorcycle, etc.). The identified objects can then be used to crop or otherwise preprocess the sensor data 102 so that it is in a form suitable for the machine learning model 104. In some embodiments, the sensor data representation can be modified to adjust the brightness, color, and / or contrast of the depicted sensor data representation. For example, the machine learning model 104 can be trained to classify signs into one or more units belonging to the unit model 114, and the object detector 110 can be used to identify signs represented by the sensor data 102 such that the sensor data representation (e.g., an image, point cloud, etc.) can be cropped and / or modified according to the location of the signs and then rescaled (if necessary) to the input resolution of the machine learning model 104.

[0033] Now for reference Figure 4 , Figure 4 An example of annotated image 400 according to some embodiments of the present invention is shown, wherein annotations are applied to sensor data to identify objects. As depicted herein, Figure 4 Objects 402 and 406 detected in the image are annotated with polygons or enclosing shapes to identify objects of interest in a specific image 400. For example, Figure 4Bounding shapes 404 and 408 are depicted for use in representing sensor data. Detected objects may represent specific objects (e.g., signs, vehicles, pedestrians, etc.), as in the example of detected object 402, and / or may represent a street or road environment including environmental and landscape elements that provide the “look and feel” of the environment, such as detected object 406. In some examples, the bounding shapes may be implemented as bounding boxes, as in bounding shapes 404 and 408, and / or may be implemented as any other shape or polygon, such as a bounding ellipse. Bounding shapes 404 and 408 may be generated using object detector 110 (e.g., one or more object detection networks, machine learning models, computer vision algorithms, etc.), and these bounding shapes may be used to preprocess (e.g., cropping, scaling, etc.) sensor data 102 for input into machine learning model 104.

[0034] Return to reference Figure 1 Reference frame marker 108, which can be used to mark, annotate, and / or generate metadata for frame sensor data 102 based on unit model 114, such as those referenced below. Figure 2 The frame tagger 108 can use information from the cell model 114 to tag frames of sensor data 102 with cell information (such as cell identifiers). For example, the frame tagger 108 can use the cell model 114 and GNSS locations associated with the sensor data 102 to tag frames of sensor data 102 with cell identifiers, where the GNSS locations are associated with the sensor data 102 and the cell identifiers may correspond to cells within the cell model 114 where the GNSS locations are located. Frames of sensor data 102 tagged by the frame tagger 108 can be used to train a machine learning model 104 to classify the sensor data 102 into one or more cells of the cell model 114.

[0035] Referring to unit model 114, unit model 114 can represent the division of a geographic region (e.g., the Earth's surface, continents, states, celestial bodies, etc.) into one or more units. Unit model 114 can be defined based on parameters such as the minimum and / or maximum number of units per country, the total number of units, road network density, and / or land area. For a non-limiting example, unit model 114 can be defined such that the Earth is divided into several distinct units (e.g., 400, 500, 650, 719, etc.). In some embodiments, countries, states, municipalities, regions, or other boundaries can be preserved in unit model 114, and regions corresponding to a particular country can be associated with one or more constituent units. For a non-limiting example, the region of the United States can be divided into multiple units in unit model 114, such as 20 units (e.g., ...). Figure 2 and 3The unit model can be defined as (as depicted in the text) or 74 units. A unit model can be defined as (e.g., allowing units to be divided into) one or more sub-units. For example, a unit can be divided into multiple units based on road network density and unit dimension using a binary space partitioning algorithm, which iteratively splits the units of unit model 114 into sub-units based on the density of the road network contained within the sub-units and / or the required number of units. For example, a unit can be divided into multiple units of different sizes based on the density of the road network within multiple sub-unit regions (e.g., highways, primary roads, secondary roads, tertiary roads, local roads, etc.).

[0036] Now for reference Figure 2 , Figure 2 An example of a cell model description 200 for segmenting a geographic region into clusters based on road similarity according to some embodiments of the present invention is shown. As depicted herein, the cell model may represent dividing a geographic region (e.g., the United States) into one or more cells, such as 20 cells in this example. Each cell may correspond to a unique identifier and / or value that can be used to distinguish a particular cell of the cell model from other cells and can be used by frame tagger 108 to annotate the sensor data representation with cell identifiers. For example, in this example, the 20 cells correspond to cell identifiers US01-US20. The cells defined by the cell model can be... Figure 1 The machine learning model 104 is used to classify sensor data representations into one or more units of the unit model. Based on the classification of sensor data representations into the units of the unit model, the similarity between the units can be determined. Figure 2 In the example, similarities between overall road environments (e.g., the appearance and feel of roads and / or locations) were determined. Based on the determined level of similarity, units can be grouped into clusters and / or sets of units, such as clusters represented by unit clusters 202A, 202B, 202C, 202D, and 202E. Each unit cluster can be represented by cluster representation 210. Cluster representation 210 depicts five clusters of the unit model—C01, C02, C03, C04, and C05—and units US01–US20 associated with each cluster. For example, units US15, US16, US19, and US20 are associated with cluster C04.

[0037] Now for reference Figure 3 , Figure 3An example of a cell model description 300 for segmenting geographic regions into clusters based on marker similarity according to some embodiments of the present invention is shown. Similar to the cell model description 200 described above, the cell model can represent dividing a geographic region (e.g., the United States) into one or more cells, which is also 20 cells in this example. The cells defined by the cell model can be... Figure 1 Machine learning model 104 is used to classify sensor data representations into one or more units of a unit model. Based on the units that classify the sensor data representations into, the similarity between units can be determined based on the detected labels represented in the sensor data. Figure 3 In the example, the similarity between markers associated with a cell is determined. Based on the determined similarity level, cells can be grouped into clusters and / or cell sets, such as clusters represented by cell clusters 302A, 302B, and 302C. In some embodiments, the clusters determined for a cell model based on marker similarity may differ from the clusters determined for the same cell model based on road environment similarity, as per [reference to...]. Figure 2 As described. Each cell cluster can be represented by a cluster representation 310. Cell representation 310 depicts three clusters of the cell model—C01, C02, and C03—and cells US01-US20 associated with each cluster. For example, cells US04, US14, US15, US16, US19, and US20 are associated with cluster C03.

[0038] Return to reference Figure 1 Once frames of sensor data 102 (e.g., with or without preprocessing and / or labeling) are provided to one or more machine learning models 104, feature vectors corresponding to candidate frames can be obtained as outputs from one or more machine learning models 104. For example, one or more machine learning models 104 can generate output feature vectors for each candidate frame of sensor data 102.

[0039] Feature vectors generated by one or more machine learning models 104 can be compared to calculate the similarity between feature vectors associated with frames of sensor data 102 and feature vectors associated with one or more other frames of sensor data 102. In some embodiments, similarity can be calculated based on comparing feature vectors with ground reality cell identification tags provided by frame tagger 108, the feature vectors being associated with one or more outputs of one or more machine learning models 104. In at least one embodiment, the similarity determined by one or more machine learning models 104 can be used to train and / or update one or more machine learning models 104 and / or one or more other machine learning models 104. For example, one or more parameters of one or more machine learning models 104 can be updated based on a comparison of the outputs of one or more machine learning models 104 with cell identification tags from frame tagger 108 (e.g., using one or more loss functions).

[0040] Therefore, the similarity calculated by machine learning model 104 can be used to generate similarity matrix 112. Similarity matrix 112 can represent a calculated probability distribution classifying sensor data frames into one or more labels, which correspond to units of unit model 114. For example, similarity matrix 112 can indicate the probability that an image classified into a first unit can also be classified into a second unit (within a threshold range). In some embodiments, similarity matrix 112 can be represented as a matrix indicating pairwise similarity between units of the unit model. For example, Figure 5 An example of a similarity matrix 500 is shown to represent the pairwise similarity of unit models.

[0041] The machine learning model 104 and / or other machine learning models described in this paper may include, but are not limited to, any type of machine learning model, such as machine learning models using linear regression, logistic regression, decision trees, support vector machines (SVM), Naive Bayes, k-nearest neighbors (Knn), K-means clustering, random forests, dimensionality reduction algorithms, gradient boosting algorithms, neural networks (e.g., autoencoders, convolutions, recurrent structures, perceptrons, long / short-term memory / LSTM, Hopfield, Boltzmann, deep belief, deconvolution, generative adversarial, liquid state machines, etc.), region of interest detection algorithms, computer vision algorithms, and / or other types of machine learning models.

[0042] As an example, in the case where the machine learning model includes a CNN, the machine learning model can include any number of layers. One or more layers can include an input layer. The input layer can store values ​​associated with the sensor data 102 (e.g., before or after post-processing). For example, when the sensor data 102 is an image, the input layer can store values ​​representing the raw pixel values ​​of the image as a volume (e.g., width, height, and color channels (e.g., RGB), such as 32x32x3).

[0043] One or more layers may include convolutional layers. Convolutional layers compute the outputs of neurons connected to local regions in the input layer, with each neuron computing the dot product between its weights and the cell domains it is connected to in the input volume. The result of a convolutional layer may be another volume, where one dimension is based on the number of filters applied (e.g., width, height, and the number of filters, such as 32x32x12 if 12 is the number of filters).

[0044] One or more layers may include rectified linear unit (ReLU) layers. For example, a ReLU layer may apply an element-wise activation function, such as max(0, x), with a threshold set to zero. The output volume of a ReLU layer may be the same as its input volume.

[0045] One or more layers may include pooling layers. Pooling layers may perform downsampling operations along spatial dimensions (e.g., height and width), which can result in a volume smaller than the input of the pooling layer (e.g., 16x16x12 from a 32x32x12 input volume).

[0046] One or more layers may include one or more fully connected layers. Each neuron in a fully connected layer can be connected to every neuron in the preceding body. In some examples, a CNN may include one or more fully connected layers such that the output of one or more layers of the CNN can be fed as input to one or more fully connected layers of the CNN. In some examples, one or more convolutional streams may be implemented by one or more machine learning models, and some or all of the convolutional streams may include corresponding fully connected layers.

[0047] In some non-limiting embodiments, one or more machine learning models may include a series of convolutional layers and max pooling layers to facilitate image feature extraction, followed by multi-scale dilated convolutional and upsampling layers to facilitate global contextual feature extraction.

[0048] While this article discusses input layers, convolutional layers, pooling layers, ReLU layers, and fully connected layers in relation to one or more machine learning models, this is not intended to be restrictive. For example, additional or alternative layers, such as normalization layers, SoftMax layers, and / or other layer types, can be used in one or more machine learning models.

[0049] Regional clusters 120 can be determined based on pairwise similarity between units in unit model 114. Units can be clustered into regional clusters 120 using distance-based clustering operations and / or algorithms (e.g., k-median or k-mean clustering). Regional clusters 120 can represent sets of units in a unit model that are determined to be similar, unaffected by geographical boundaries and / or geographical proximity. In some embodiments, clusters of one or more units in the unit model can be presented separately, for example... Figure 2 and 3 The cluster representations are 210 and 310.

[0050] Now for reference Figure 6-8 Each block of methods 600, 700, and 800 described herein includes a computational process that can be performed using at least one or a combination of hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in memory. Methods can also be embodied as computer-usable instructions stored on a computer storage medium. Methods can be provided by a standalone application, service, or managed service (independently or in combination with another managed service) or a plug-in to another product, to name a few. Furthermore, by way of example, regarding... Figure 1 Process 100 describes methods 600, 700, and 800. However, these methods may be performed additionally or alternatively by any system or any combination of systems, including but not limited to those described herein.

[0051] Figure 6 This is a flowchart illustrating a method 600 for generating a cluster of cells using a neural network according to some embodiments of the present disclosure. At block 602, method 600 includes: determining a first cell classification for a first sensor data instance, the first cell classification mapping the first sensor data instance to a first cell of a cell model. For example, one or more machine learning models 104 may determine a classification of sensor data frame 102, which classifies the sensor data frame to a specific cell of cell model 114.

[0052] In box B604, method 600 includes determining a second cell classification for a second sensor data instance, the second cell classification mapping the second sensor database instance to a second cell of a cell model. For example, one or more machine learning models 104 may determine cells of cell model 114 in which frames of sensor data 102 are classified.

[0053] In box B606, method 600 includes determining the similarity between the first unit and the second unit. For example, based on the classification provided by one or more machine learning models 104, the similarity between units of unit model 114 can be determined and used to generate a similarity matrix 112.

[0054] In box B608, method 600 includes generating a cluster comprising a first unit and a second unit, at least in part, based on similarity. For example, based on the similarity between units in unit model 114, the units can be clustered into a region cluster 120.

[0055] Now for reference Figure 7 , Figure 7 This is a flowchart illustrating a method 700 for generating a cluster of cells using a neural network according to some embodiments of the present disclosure. In block B702, method 700 includes determining, for a sensor dataset, a mapping from one or more sensor data instances in the sensor dataset to one or more cells of a cell model. For example, machine learning model 104 could classify a representation of sensor data 102 into one or more cells of cell model 114.

[0056] In box B704, method 700 includes determining a similarity distribution based at least on a mapping of one or more sensor data instances to one or more unit cells, wherein the similarity distribution indicates pairwise similarity between each unit in one or more units of the unit model. For example, one or more machine learning models 104 can be used to generate a similarity matrix 112.

[0057] In box B706, method 700 includes generating at least one cluster based on at least a similarity distribution, the at least one cluster comprising one or more units of a unit model. For example, region cluster 120 can be generated based on the similarity between units of unit model 114 indicated by similarity matrix 112.

[0058] refer to Figure 8 , Figure 8This is a flowchart illustrating a method 800 for generating cell clusters using a neural network according to some embodiments of the present disclosure. In block B802, method 800 includes determining a cell identifier for each sensor data instance of the sensor dataset using location data associated with the sensor dataset, wherein the cell identifier is associated with a cell model comprising the cell set. For example, frame tagger 108 may use GNSS location data associated with sensor data 102 to determine a tag corresponding to a specific cell of cell model 114.

[0059] In box B804, method 800 includes training a machine learning model using the cell identifier determined for each instance of sensor data to classify the sensor data into cells in the cell set. For example, using sensor data 102 and labels determined by frame tagger 108, machine learning model 104 can be trained to classify sensor data 102 into one or more cells of cell model 114.

[0060] In box B806, method 800 includes classifying a sensor dataset into a set of cells using a machine learning model. For example, machine learning model 104 can classify instances of sensor data 102 into cells of cell model 114.

[0061] In box B808, method 800 includes determining a similarity distribution based on classifying the sensor dataset into the unit set, the similarity distribution indicating similarity values ​​between multiple units of the unit model. For example, machine learning model 104 may output similarity values ​​between units of unit model 114 to generate a similarity matrix 112.

[0062] In box B810, method 800 includes using the similarity distribution to determine at least one cluster of units in the unit set. For example, the similarity between units of unit model 114, indicated by similarity matrix 112, can be used to determine region cluster 120, which groups the units of unit model 114 into one or more similar unit clusters.

[0063] The systems and methods described herein can be used, but are not limited to, non-autonomous vehicles, semi-autonomous vehicles (e.g., in one or more adaptive driver assistance systems (ADAS)), manned and unmanned robots or robot-using platforms, warehouse vehicles, off-road vehicles, vehicles connected to one or more trailers, airships, boats, shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, engineering vehicles, underwater vehicles, drones, and / or other vehicle types. Furthermore, the systems and methods described herein can be used for a variety of purposes, such as, but not limited to, machine control, machine motion, machine driving, synthetic data generation, model training, perception, augmented reality, virtual reality, mixed reality, robotics, security and surveillance, autonomous or semi-autonomous machine applications, deep learning, environmental simulation, object simulation and digital twins, data center processing, conversational artificial intelligence, optical transmission simulation (e.g., ray tracing, path tracing, etc.), collaborative content creation of 3D assets, cloud computing, and / or any other suitable application for creating collaborative content.

[0064] The disclosed embodiments may be included in a variety of different systems, such as automotive systems (e.g., control systems for autonomous or semi-autonomous machines, perception systems for autonomous or semi-autonomous machines), systems implemented using robots, aviation systems, medical systems, boating systems, smart area monitoring systems, systems for performing deep learning operations, systems for performing simulation operations (including environmental and / or object simulation and digital twins), systems implemented using edge devices, systems that merge one or more virtual machines (VMs), systems for performing synthetic data generation operations, systems implemented at least partially in a data center, systems for performing conversational AI operations, systems for performing optical transmission simulations, systems for performing collaborative content creation of 3D assets, systems implemented at least partially using cloud computing resources, and / or other types of systems.

[0065] Example autonomous vehicles

[0066] Figure 9AThis is an illustration of an example autonomous vehicle 900 according to some embodiments of this disclosure. The autonomous vehicle 900 (also referred to herein as “vehicle 900”) may include, but is not limited to, passenger vehicles such as automobiles, trucks, buses, ambulances, shuttles, electric or motorized bicycles, motorcycles, fire trucks, police cars, ambulances, boats, engineering vehicles, underwater vehicles, drones, trailer-mounted vehicles, and / or other types of vehicles (e.g., driverless and / or capable of accommodating one or more passengers). Autonomous vehicles are typically described according to the level of automation defined by the National Highway Traffic Safety Administration (NHTSA) of the U.S. Department of Transportation and the Society of Automotive Engineers (SAE) in their standard “Classification and Definition of Terms Related to Driving Automation Systems for Road Motor Vehicles” (Standard No. J3016-201806, published June 15, 2018; Standard No. J3016-201609, published September 30, 2016; and previous and future versions of this standard). Vehicle 900 may be capable of having one or more of the functions of Level 3 to Level 5 according to the level of autonomous driving. Vehicle 900 may be able to function according to one or more of Levels 1 through 5 of autonomous driving. For example, vehicle 900 may be able to provide driver assistance (Level 1), partial automation (Level 2), conditional automation (Level 3), high automation (Level 4), and / or full automation (Level 5), depending on the implementation. As used herein, the term “autonomy” may include any and / or all types of autonomy for vehicle 900 or other machines, such as full autonomy, high autonomy, conditional autonomy, partial autonomy, providing assisted autonomy, semi-autonomy, primary autonomy, or other names.

[0067] Vehicle 900 may include components such as chassis, body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other vehicle components. Vehicle 900 may include a propulsion system 950, such as an internal combustion engine, a hybrid power plant, an all-electric motor, and / or another type of propulsion system. Propulsion system 950 may be connected to the drivetrain of vehicle 900, which may include a transmission, to enable propulsion of vehicle 900. Propulsion system 950 may be controlled in response to receiving a signal from throttle / accelerator 952.

[0068] A steering system 954, which may include a steering wheel, can be used to steer the vehicle 900 (e.g., along a desired path or route) when the propulsion system 950 is operating (e.g., when the vehicle is in motion). The steering system 954 may receive signals from the steering actuator 956. For fully automatic (level 5) functions, the steering wheel may be optional.

[0069] The brake sensor system 946 can be used to operate the vehicle brakes in response to receiving signals from the brake actuator 948 and / or the brake sensor.

[0070] It can include one or more System-on-a-Chip (SoC) 904 ( Figure 9C One or more controllers 936, including one or more GPUs, can provide signals (e.g., signals representing commands) to one or more components and / or systems of vehicle 900. For example, one or more controllers can send signals to operate vehicle brakes via one or more brake actuators 948, to operate steering system 954 via one or more steering actuators 956, and to operate propulsion system 950 via one or more throttles / accelerators 952. One or more controllers 936 may include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operating commands (e.g., signals representing commands) to enable autonomous driving and / or assist a human driver in driving vehicle 900. One or more controllers 936 may include a first controller 936 for autonomous driving functions, a second controller 936 for functional safety functions, a third controller 936 for artificial intelligence functions (e.g., computer vision), a fourth controller 936 for infotainment functions, a fifth controller 936 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 936 can handle two or more of the functions described above, two or more controllers 936 can handle a single function, and / or any combination thereof.

[0071] One or more controllers 936 may provide signals for controlling one or more components and / or systems of vehicle 900 in response to sensor data (e.g., sensor inputs) received from one or more sensors. Sensor data may be received from, for example, but not limited to, a global navigation satellite system sensor 958 (e.g., a Global Positioning System sensor), a RADAR sensor 960, an ultrasonic sensor 962, a LIDAR sensor 964, an inertial measurement unit (IMU) sensor 966 (e.g., an accelerometer, gyroscope, magnetic compass, magnetometer, etc.), a microphone 996, a stereo camera 968, a wide-angle camera 970 (e.g., a fisheye camera), an infrared camera 972, a surround camera 974 (e.g., a 360-degree camera), a long-range and / or medium-range camera 998, a speed sensor 944 (e.g., for measuring the rate of vehicle 900), a vibration sensor 942, a steering sensor 940, a braking sensor (e.g., as part of a braking sensor system 946), and / or other sensor types.

[0072] One or more controllers 936 may receive inputs (e.g., represented by input data) from the instrument cluster 932 of the vehicle 900 and provide outputs (e.g., represented by output data, display data, etc.) via a human-machine interface (HMI) display 934, an auditory signaling device, a speaker, and / or via other components of the vehicle 900. These outputs may include information such as vehicle speed, rate, time, map data (e.g., [missing information]). Figure 9C Information such as the HD map 922, location data (e.g., the location of vehicle 900 on the map), direction, and the location of other vehicles (e.g., occupying a grid), as well as information about objects and their states perceived by the controller 936, etc. For example, the HMI display 934 can display information about the existence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information about driving maneuvers that the vehicle has made, is making, or will make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0073] The vehicle 900 further includes a network interface 924, which can communicate via one or more networks using one or more wireless antennas 926 and / or a modem. For example, the network interface 924 may be able to communicate via LTE, WCDMA, UMTS, GSM, CDMA2000, etc. The one or more wireless antennas 926 may also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using one or more local area networks such as Bluetooth, Bluetooth LE, Z-Wave, ZigBee, etc., and / or one or more low-power wide area networks (LPWANs) such as LoRaWAN, SigFox, etc.

[0074] Figure 9B For use in accordance with some embodiments of this disclosure Figure 9A This is an example of the camera position and field of view of an example autonomous vehicle 900. The camera and its respective field of view are an example embodiment and are not intended to be limiting. For example, additional and / or replaceable cameras may be included, and / or these cameras may be located at different positions on the vehicle 900.

[0075] The camera type used for the camera may include, but is not limited to, a digital camera suitable for use with components and / or systems of vehicle 900. The camera may operate at Automotive Safety Integrity Level (ASIL) B and / or another ASIL. The camera type may have any image capture rate, such as 60 frames per second (fps), 120 fps, 240 fps, etc., depending on the embodiment. The camera may be able to use a rolling shutter, a global shutter, another type of shutter, or a combination thereof. In some examples, the color filter array may include a red-transparent (RCCC) color filter array, a red-transparent-blue (RCCB) color filter array, a red-blue-green (RBGC) color filter array, a Foveon X3 color filter array, a Bayer sensor (RGGB) color filter array, a monochrome sensor color filter array, and / or another type of color filter array. In some embodiments, a high-resolution camera, such as a camera with RCCC, RCCB, and / or RBGC color filter arrays, may be used in efforts to improve light sensitivity.

[0076] In some examples, one or more of the cameras can be used to perform advanced driver assistance system (ADAS) functions (e.g., as part of a redundant or fail-safe design). For example, a multi-function monocular camera can be installed to provide functions including lane departure warning, traffic sign assistance, and intelligent headlight control. One or more of the cameras (e.g., all cameras) can simultaneously record and provide image data (e.g., video).

[0077] One or more of the cameras can be mounted in mounting components such as custom-designed (3-D printed) parts to cut off stray light and reflections from inside the vehicle (e.g., reflections from the dashboard in the windshield mirror) that may interfere with the camera's image data capture capabilities. Regarding the wing mirror mounting components, the wing mirror components can be custom-3-D printed so that the camera mounting plate matches the shape of the wing mirror. In some examples, one or more cameras can be integrated into the wing mirror. For side-view cameras, one or more cameras can also be integrated into the four pillars at each corner of the cab.

[0078] A camera with a field of view that includes the environment in front of the vehicle 900 (e.g., a front-facing camera) can be used for surround view to help identify forward paths and obstacles, and, with the assistance of one or more controllers 936 and / or control SoCs, to provide information crucial for generating an occupancy grid and / or determining a preferred vehicle path. The front-facing camera can be used to perform many of the same ADAS functions as LiDAR, including emergency braking, pedestrian detection, and collision avoidance. The front-facing camera can also be used in ADAS functions and systems, including Lane Departure Warning (“LDW”), Autonomous Cruise Control (“ACC”), and / or other functions such as traffic sign recognition.

[0079] A variety of cameras can be used in front-facing configurations, including, for example, monocular camera platforms including CMOS (Complementary Metal-Oxide-Semiconductor) color imagers. Another example could be the wide-angle camera 970, which can be used to perceive objects entering the field of view from the periphery (such as pedestrians, traffic at intersections, or bicycles). Although Figure 9B The middle image shows only one wide-angle camera, but any number of wide-angle cameras 970 can be present on the vehicle 900. Furthermore, a remote camera 998 (e.g., a pair of long-view stereo cameras) can be used for depth-based object detection, especially for objects for which a neural network has not yet been trained. The remote camera 998 can also be used for object detection and classification, as well as basic object tracking.

[0080] One or more stereo cameras 968 may also be included in a front-mounted configuration. The stereo camera 968 may include an integrated control unit comprising a scalable processing unit that can provide a multi-core microprocessor and programmable logic (FPGA) with an integrated CAN or Ethernet interface on a single chip. Such a unit can be used to generate a 3D map of the vehicle environment, including distance estimates for all points in the image. Alternative stereo cameras 968 may include a compact stereo vision sensor that may include two camera lenses (one on each side) and an image processing chip capable of measuring the distance from the vehicle to a target object and using the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning functions. Other types of stereo cameras 968 may be used in addition to those described herein, or alternatively.

[0081] Cameras with a field of view including the side portion of the vehicle 900 (e.g., side-view cameras) can be used for surround view, providing information for creating and updating occupancy grids and generating side-impact collision warnings. For example, surround camera 974 (e.g., ... Figure 9B The four surround cameras 974 shown can be mounted on the vehicle 900. The surround cameras 974 can include wide-angle cameras 970, fisheye cameras, 360-degree cameras, and / or the like. Four examples are provided; the four fisheye cameras can be positioned at the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle can use three surround cameras 974 (e.g., left, right, and rear) and can utilize one or more other cameras (e.g., forward-facing cameras) as a fourth surround-view camera.

[0082] A camera with a field of view that includes the environment behind the vehicle 900 (e.g., a rear-view camera) can be used for parking assistance, surround view, rear collision warning, and creating and updating occupancy grids. A wide variety of cameras can be used, including but not limited to those also suitable as front-facing cameras as described herein (e.g., long-range and / or mid-range camera 998, stereo camera 968, infrared camera 972, etc.).

[0083] Figure 9C For use in accordance with some embodiments of this disclosure Figure 9A The example autonomous vehicle 900 is illustrated in the block diagram of an example system architecture. It should be understood that this arrangement, and other arrangements described herein, are merely illustrative. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities, which may be implemented as discrete or distributed components or in combination with other components, and in any suitable combination and location. The various functions described herein as being performed by these entities can be implemented via hardware, firmware, and / or software. For example, the various functions can be implemented by a processor executing instructions stored in memory.

[0084] Figure 9C Each component, feature, and system in vehicle 900 is illustrated as being connected via bus 902. Bus 902 may include a Controller Area Network (CAN) data interface (or, alternatively, referred to herein as the "CAN bus"). CAN may be a network within vehicle 900 used to assist in the control of various features and functions of vehicle 900, such as the actuation of brakes, acceleration, braking, steering, windshield wipers, etc. The CAN bus can be configured to have dozens or even hundreds of nodes, each with its own unique identifier (e.g., CAN ID). The CAN bus can be read to find steering wheel angle, ground speed, engine speed per minute (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0085] Although bus 902 is described herein as a CAN bus, this is not intended to be limiting. For example, FlexRay and / or Ethernet may be used in addition to or alternatively to a CAN bus. Furthermore, although bus 902 is represented by a single line, this is not intended to be limiting. For example, any number of buses 902 may exist, which may include one or more CAN buses, one or more FlexRay buses, one or more Ethernet buses, and / or one or more other types of buses using different protocols. In some examples, two or more buses 902 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 902 may be used for collision avoidance functions, and a second bus 902 may be used for drive control. In any example, each bus 902 may communicate with any component of vehicle 900, and two or more buses 902 may communicate with the same component. In some examples, each SoC 904, each controller 936, and / or each computer within the vehicle may have access to the same input data (e.g., input from sensors of vehicle 900) and may be connected to a common bus such as a CAN bus.

[0086] Vehicle 900 may include one or more controllers 936, such as those described herein. Figure 9A The controllers described. Controller 936 can be used for a wide variety of functions. Controller 936 can be coupled to any other different components and systems of vehicle 900 and can be used for the control of vehicle 900, artificial intelligence of vehicle 900, infotainment and / or the like for vehicle 900.

[0087] Vehicle 900 may include one or more System-on-Chip (SoC) 904s. SoC 904 may include a CPU 906, GPU 908, processor 910, cache 912, accelerator 914, data storage 916, and / or other components and features not shown. SoC 904 can be used to control vehicle 900 across a wide variety of platforms and systems. For example, one or more SoCs 904s may be combined with an HD map 922 in a system (e.g., the system of vehicle 900), the HD map being transmitted via a network interface 924 from one or more servers (e.g., [server name missing]). Figure 9D One or more servers (978) receive map refresh and / or updates.

[0088] The CPU 906 may include a CPU cluster or a CPU complex (or, alternatively, referred to herein as "CCPLEX"). The CPU 906 may include multiple cores and / or L2 cache. For example, in some embodiments, the CPU 906 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 906 may include four dual-core clusters, each with a dedicated L2 cache (e.g., 2MB L2 cache). The CPU 906 (e.g., CCPLEX) may be configured to support simultaneous cluster operation, such that any combination of CPU 906 clusters can be active at any given time.

[0089] The CPU 906 can implement power management capabilities including one or more of the following features: automatic clock gating of hardware blocks when idle to conserve dynamic power; clock gating of each core when the core is not actively executing instructions due to the execution of WFI / WFE instructions; independent power gating of each core; independent clock gating of each core cluster when all cores are clock-gated or power-gated; and / or independent power gating of each core cluster when all cores are power-gated. The CPU 906 can further implement enhanced algorithms for managing power states, where allowed power states and desired wake-up times are specified, and the hardware / microcode determines the optimal power state to enter for the core, cluster, and CCPLEX. The processing core can support simplified power state entry sequences in software, with this work offloaded to the microcode.

[0090] The GPU 908 may include an integrated GPU (or, alternatively, referred to herein as an "iGPU"). The GPU 908 may be programmable and efficient for parallel workloads. In some examples, the GPU 908 may use an enhanced tensor instruction set. The GPU 908 may include one or more streaming microprocessors, each of which may include an L1 cache (e.g., an L1 cache with at least 96KB of storage capacity), and two or more of these streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB of storage capacity). In some embodiments, the GPU 908 may include at least eight streaming microprocessors. The GPU 908 may use a computation application programming interface (API). Furthermore, the GPU 908 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0091] In automotive and embedded applications, the GPU 908 can be power-optimized for optimal performance. For example, the GPU 908 can be fabricated on FinFETs. However, this is not intended to be limiting, and the GPU 908 can be fabricated using other semiconductor manufacturing processes. Each streaming microprocessor can combine several mixed-precision processing cores divided into multiple blocks. For example, and without limitation, 64 PF32 cores and 32 PF64 cores can be divided into four processing blocks. In such an example, each processing block can be allocated 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two mixed-precision NVIDIA Tensor cores for deep learning matrix arithmetic, an L0 instruction cache, a warp scheduler, dispatch units, and / or a 64KB register file. Furthermore, the streaming microprocessor can include independent parallel integer and floating-point data paths to leverage the mixture of computation and addressing computations for efficient execution of workloads. The streaming microprocessor can include independent thread scheduling capabilities to allow for finer-grained synchronization and cooperation between parallel threads. Streaming microprocessors can include a combination of L1 data cache and shared memory units to improve performance while simplifying programming.

[0092] The GPU 908 may include, in some examples, a high-bandwidth memory (HBM) and / or a 16GB HBM2 memory subsystem providing peak memory bandwidth of approximately 900GB / s. In some examples, in addition to HBM memory or alternatively, synchronous graphics random access memory (SGRAM), such as fifth-generation graphics double data rate synchronous random access memory (GDDR5), may be used.

[0093] The GPU 908 may include unified memory technology, which includes access counters to allow memory pages to be migrated more precisely to the processors that access them most frequently, thereby improving the efficiency of shared memory ranges between processors. In some examples, Address Translation Service (ATS) support can be used to allow the GPU 908 to directly access the CPU 906 page tables. In such examples, when the GPU 908 Memory Management Unit (MMU) experiences a miss, an address translation request can be transferred to the CPU 906. In response, the CPU 906 can look up the virtual-physical mapping for the address in its page tables and transfer the translation back to the GPU 908. Thus, unified memory technology can allow a single unified virtual address space for the memory of both the CPU 906 and GPU 908, simplifying GPU 908 programming and porting applications to the GPU 908.

[0094] In addition, the GPU 908 may include access counters that track how frequently the GPU 908 accesses the memory of other processors. These access counters help ensure that memory pages are moved to the physical memory of the processor that accesses those pages most frequently.

[0095] SoC 904 may include any number of caches 912, including those described herein. For example, cache 912 may include an L3 cache available to both CPU 906 and GPU 908 (e.g., it is connected to both CPU 906 and GPU 908). Cache 912 may include a write-back cache, which can track the state of rows, for example, using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). Depending on the embodiment, the L3 cache may include 4MB or more, but a smaller cache size may also be used.

[0096] SoC 904 may include an arithmetic logic unit (ALU), which can be used to perform processing of any of a variety of tasks or operations related to vehicle 900—such as processing a DNN. Additionally, SoC 904 may include a floating-point unit (FPU)—or other mathematical coprocessor or digital coprocessor type—for performing mathematical operations within the system. For example, SoC 104 may include one or more FPUs integrated as execution units within CPU 906 and / or GPU 908.

[0097] SoC 904 may include one or more accelerators 914 (e.g., hardware accelerators, software accelerators, or a combination thereof). For example, SoC 904 may include a hardware acceleration cluster, which may include optimized hardware accelerators and / or large on-chip memory. This large on-chip memory (e.g., 4MB SRAM) can enable the hardware acceleration cluster to accelerate neural networks and other computations. The hardware acceleration cluster can be used to supplement GPU 908 and offload some tasks from GPU 908 (e.g., freeing up more cycles of GPU 908 to perform other tasks). As an example, accelerator 914 can be used for targeted workloads (e.g., perception, convolutional neural networks (CNNs), etc.) that are stable enough to be easily controlled for acceleration. When used herein, the term "CNN" can include all types of CNNs, including region-based or region convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0098] Accelerator 914 (e.g., hardware acceleration clusters) may include a Deep Learning Accelerator (DLA). A DLA may include one or more Tensor Processing Units (TPUs) that can be configured to provide an additional 10 trillion operations per second for deep learning applications and inference. TPUs may be accelerators configured to perform image processing functions (e.g., for CNNs, RCNNs, etc.) and optimized for performing image processing functions. DLAs may be further optimized for a specific set of neural network types and floating-point operations as well as inference. DLAs are designed to provide higher performance per millimeter than general-purpose GPUs and significantly outperform CPUs. TPUs can perform several functions, including single-instance convolution functions, support for INT8, INT16, and FP16 data types for both features and weights, and post-processor functions.

[0099] DLA can execute neural networks, especially CNNs, quickly and efficiently on processed or unprocessed data for any function across a wide variety of applications, such as, but not limited to: CNNs for object recognition and detection using data from camera sensors; CNNs for distance estimation using data from camera sensors; CNNs for emergency vehicle detection and recognition using data from microphones; CNNs for face recognition and vehicle owner recognition using data from camera sensors; and / or CNNs for safety and / or safety-related events.

[0100] The DLA can perform any function of the GPU 908, and by using inference accelerators, for example, designers can target either the DLA or the GPU 908 for any function. For instance, a designer can focus the CNN processing and floating-point operations on the DLA and leave other functions to the GPU 908 and / or other accelerators 914.

[0101] Accelerator 914 (e.g., a hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may alternatively be referred to herein as a computer vision accelerator. A PVA can be designed and configured to accelerate computer vision algorithms for advanced driver assistance systems (ADAS), autonomous driving, and / or augmented reality (AR) and / or virtual reality (VR) applications. A PVA can provide a balance between performance and flexibility. For example, each PVA may include, for example, but not limited to, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0102] RISC cores can interact with image sensors (such as the image sensor of any camera described herein), image signal processors, and / or the like. Each of these RISC cores may include any amount of memory. Depending on the embodiment, the RISC core may use any of several protocols. In some examples, the RISC core may execute a real-time operating system (RTOS). RISC cores may be implemented using one or more integrated circuit devices, application-specific integrated circuits (ASICs), and / or memory devices. For example, a RISC core may include an instruction cache and / or tightly coupled RAM.

[0103] DMA enables PVA components to access system memory independently of the CPU 906. DMA can support any number of features to provide optimizations to the PVA, including but not limited to support for multidimensional addressing and / or circular addressing. In some examples, DMA can support addressing in up to six or more dimensions, which can include block width, block height, block depth, horizontal block step, vertical block step, and / or depth step.

[0104] A vector processor can be a programmable processor designed to efficiently and flexibly execute programming for computer vision algorithms and provide signal processing capabilities. In some examples, a PVA may include a PVA core and two vector processing subsystem partitions. The PVA core may include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem may operate as the main processing engine of the PVA and may include a vector processing unit (VPU), an instruction cache, and / or a vector memory (e.g., a VMEM). The VPU core may include a digital signal processor, such as, for example, a Single Instruction Multiple Data (SIMD) or Very Long Instruction Word (VLIW) digital signal processor. The combination of SIMD and VLIW can enhance throughput and speed.

[0105] Each of the vector processors may include an instruction cache and may be coupled to dedicated memory. Consequently, in some examples, each of the vector processors may be configured to execute independently of other vector processors. In other examples, the vector processors included in a particular PVA may be configured to employ data parallelism. For example, in some embodiments, multiple vector processors included in a single PVA may execute the same computer vision algorithm, but on different regions of an image. In other examples, vector processors included in a particular PVA may execute different computer vision algorithms simultaneously on the same image, or even different algorithms on a sequence of images or portions of an image. Among other things, any number of PVAs may be included in a hardware-accelerated cluster, and any number of vector processors may be included in each of these PVAs. Furthermore, the PVA may include additional error-correcting code (ECC) memory to enhance overall system security.

[0106] Accelerator 914 (e.g., a hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high-bandwidth, low-latency SRAM for accelerator 914. In some examples, on-chip memory may include at least 4MB of SRAM consisting of, for example, but not limited to, eight field-configurable memory blocks accessible by both PVA and DLA. Each pair of memory blocks may include an Advanced Peripheral Bus (APB) interface, configuration circuitry, a controller, and a multiplexer. Any type of memory can be used. PVA and DLA can access memory via a backbone that provides high-speed memory access to PVA and DLA. The backbone may include (e.g., using an APB) an on-chip computer vision network that interconnects PVA and DLA to memory.

[0107] On-chip computer vision networks can include interfaces that ensure both the PVA and DLA provide ready and valid signals before transmitting any control signals / addresses / data. Such interfaces can provide separate phases and channels for transmitting control signals / addresses / data, as well as burst communication for continuous data transmission. This type of interface can conform to ISO 26262 or IEC 61508 standards, but other standards and protocols can also be used.

[0108] In some examples, the SoC 904 may include, for example, a real-time ray tracing hardware accelerator as described in U.S. Patent Application No. 16 / 101,232, filed August 10, 2018. This real-time ray tracing hardware accelerator can be used to quickly and efficiently determine the location and extent of objects (e.g., within a world model) to generate real-time visualization simulations for RADAR signal interpretation, sound propagation synthesis and / or analysis, SONAR system simulation, general wave propagation simulation, comparison with LiDAR data for localization and / or other functional purposes, and / or for other uses. In some embodiments, one or more Tree Traversal Units (TTUs) may be used to perform one or more ray tracing-related operations.

[0109] Accelerator 914 (e.g., hardware accelerator clusters) has broad applications in autonomous driving. PVAs can be programmable vision accelerators used in critical processing stages of ADAS and autonomous vehicles. The capabilities of PVAs are a good match for algorithmic domains requiring predictable processing, low power, and low latency. In other words, PVAs perform well in semi-dense or dense rule computation, even on small datasets requiring predictable runtimes with low latency and low power. Therefore, in the context of platforms for autonomous vehicles, PVAs are designed to run classical computer vision algorithms because they are efficient in object detection and integer mathematical operations.

[0110] For example, according to one embodiment of this technology, PVA is used to perform computer stereo vision. In some examples, semi-global matching-based algorithms may be used, but this is not intended to be limiting. Many applications for Level 3-5 autonomous driving require instantaneous motion estimation / stereo matching (e.g., from moving structures, pedestrian recognition, lane detection, etc.). PVA can perform computer stereo vision functions on input from two monocular cameras.

[0111] In some examples, PVA can be used to perform intensive optical flow, providing processed RADAR data from the raw RADAR data (e.g., using 4D Fast Fourier Transform). In other examples, PVA is used for time-of-flight depth processing, which, for example, involves processing raw time-of-flight data to provide processed time-of-flight data.

[0112] DLA can be used to run any type of network to enhance control and driving safety, including, for example, neural networks that output a confidence metric for each object detection. Such a confidence value can be interpreted as a probability or as providing a relative “weight” for each detection compared to other detections. This confidence value allows the system to make further decisions about which detections should be considered true positives rather than false positives. For example, the system can set a threshold for the confidence and only consider detections exceeding the threshold as true positives. In an Automatic Emergency Braking (AEB) system, false positives can cause the vehicle to automatically perform emergency braking, which is clearly undesirable. Therefore, only the most confident detections should be considered as triggers for AEB. DLA can run a neural network to regress the confidence value. This neural network can take at least a subset of parameters as input, such as bounding box dimensions, ground plane estimates (e.g., from another subsystem), inertial measurement unit (IMU) sensor 966 outputs related to vehicle orientation and distance, 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LiDAR sensor 964 or RADAR sensor 960), etc.

[0113] The SoC 904 may include one or more data storage units 916 (e.g., memory). The data storage unit 916 may be on-chip memory of the SoC 904, which may store neural networks to be executed on the GPU and / or DLA. In some examples, for redundancy and security, the data storage unit 916 may be large enough to store multiple instances of the neural network. The data storage unit 912 may include an L2 or L3 cache 912. References to the data storage unit 916 may include references to memory associated with the PVA, DLA, and / or other accelerators 914 as described herein.

[0114] The SoC 904 may include one or more processors 910 (e.g., embedded processors). Processor 910 may include a startup and power management processor, which may be a dedicated processor and subsystem for handling startup power and management functions, as well as safety implementation. The startup and power management processor may be part of the SoC 904 startup sequence and may provide runtime power management services. The startup power and management processor may provide clock and voltage programming, auxiliary system low-power state transitions, SoC 904 thermal and temperature sensor management, and / or SoC 904 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to the temperature, and the SoC 904 may use the ring oscillator to detect the temperature of the CPU 906, GPU 908, and / or accelerator 914. If it is determined that the temperature exceeds a threshold, the startup and power management processor may enter a temperature fault routine and place the SoC 904 into a lower power state and / or place the vehicle 900 into a driver-safe parking mode (e.g., safely stopping the vehicle 900).

[0115] The processor 910 may further include a set of embedded processors that can be used as an audio processing engine. The audio processing engine can be an audio subsystem that allows for full hardware support for multi-channel audio via multiple interfaces, as well as a wide and flexible range of audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core with a digital signal processor and dedicated RAM.

[0116] The processor 910 may further include an always-on-processor engine that can provide the necessary hardware features to support low-power sensor management and wake-up use cases. This always-on-processor engine may include a processor core, tightly coupled RAM, support for peripherals (such as timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0117] The processor 910 may further include a secure cluster engine, which includes a dedicated processor subsystem for handling security management for automotive applications. The secure cluster engine may include two or more processor cores, tightly coupled RAM, support for peripheral devices (e.g., timers, interrupt controllers, etc.), and / or routing logic. In secure mode, the two or more cores may operate in lockstep mode and function as a single core with comparison logic that detects any differences between their operations.

[0118] The processor 910 may further include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0119] The processor 910 may further include a high dynamic range signal processor, which may include an image signal processor, which is a hardware engine that is part of the camera processing pipeline.

[0120] Processor 910 may include a video image compositer, which may be (e.g., implemented on a microprocessor) a processing block, implementing video post-processing functions required by the video playback application to generate the final image for the player window. The video image compositer may perform lens distortion correction on the wide-angle camera 970, the surround camera 974, and / or the in-cabin monitoring camera sensor. The in-cabin monitoring camera sensor is preferably monitored by a neural network running on another instance of an advanced SoC, configured to recognize in-cabin events and respond accordingly. The in-cabin system may perform lip reading to activate mobile phone services and make calls, dictate emails, change vehicle destinations, activate or change the vehicle's infotainment system and settings, or provide voice-activated web browsing. Some functions are only available to the driver when the vehicle is operating in autonomous mode and are disabled in other situations.

[0121] Video image compositers can include enhanced temporal denoising for both spatial and temporal noise reduction. For example, in the case of motion in the video, denoising appropriately weights spatial information, reducing the noise of neighboring elements.

[0122] The weighting of information provided by frames. When an image or part of an image does not contain motion, the temporal noise reduction performed by the video image compositer can use information from previous images to reduce noise in the current image.

[0123] The video image compositer can also be configured to perform stereo correction on the input stereo camera frames. This is useful when the operating system desktop is in use and the GPU 908 does not need to continuously render new frames.

[0124] In addition, the video-image compositer can be further used in user interface components. Even when the GPU 908 is powered on and activated for 3D rendering, the video-image compositer can be used to reduce the load on the GPU 908 to improve performance and responsiveness.

[0125] The SoC 904 may further include a Mobile Industry Processor Interface (MIPI) camera serial interface for receiving video and input from the camera, a high-speed interface, and / or interfaces that can be used for the camera and related pixels.

[0126] The SoC 904 may further include a video input block with input functionality. This can be software-controlled and used to receive I / O signals not assigned to a specific role.

[0127] The SoC 904 can further include a wide range of peripheral interfaces to enable communication with peripheral devices, audio codecs, power management and / or other devices. The SoC 904 can be used to process signals from cameras (via Gigabit Multimedia Serial Link and Ethernet connections), sensors (e.g., those that can communicate via…)

[0128] Data from LIDAR sensors 964, RADAR sensors 960, etc. (connected via Ethernet), data from bus 902 (e.g., vehicle speed, steering wheel position, etc.), and data from GNSS sensors 958 (connected via Ethernet or CAN bus). SoC 904 can then...

[0129] One step includes a dedicated high-performance, high-capacity storage controller, which may include its own DMA engine and can be used to free up the CPU 906 from routine data management tasks.

[0130] The SoC 904 can be an end-to-end platform with a flexible architecture that spans automation 3-5.

[0131] This provides a comprehensive functional safety architecture for a platform that leverages and efficiently utilizes computer vision and ADAS technologies to achieve diversity and redundancy, along with deep learning tools, to deliver a flexible and reliable driving software stack. The SoC 904 can be faster, more reliable, and even more energy- and space-efficient than conventional systems. For example, when combined with the CPU 906, GPU 908, and data storage 916, the accelerator 914 can provide a fast and efficient platform for Level 3–5 autonomous vehicles.

[0132] Therefore, this technology offers capabilities and functionalities that cannot be achieved through conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using high-level programming languages ​​such as C to execute a wide variety of processing algorithms across a diverse range of visual data. However, CPUs often cannot meet the performance requirements of many computer vision applications, such as those related to execution time and power consumption. In particular, many CPUs cannot execute complex object detection algorithms in real time, which is a requirement for automotive ADAS applications and practical Level 3-5 autonomous vehicles.

[0133] In contrast to conventional systems, the techniques described in this paper, by providing CPU complexes, GPU complexes, and hardware acceleration clusters, allow multiple neural networks to be executed simultaneously and / or sequentially, and the results to be combined to achieve Level 3–5 autonomous driving capabilities. For example, a CNN executed on a DLA or dGPU (e.g., GPU 920) could include text and word recognition, allowing a supercomputer to read and understand traffic signs, including those for which neural networks have not yet been specifically trained. The DLA could further include a neural network capable of recognizing, interpreting, and providing semantic understanding of the signs, and passing that semantic understanding to a path planning module running on a CPU complex.

[0134] As another example, multiple neural networks can operate simultaneously, as required for Level 3, 4, or 5 driving. For instance, a warning sign consisting of "Caution: Flashing lights indicate icy conditions" along with a light can be interpreted independently or jointly by several neural networks. The sign itself can be recognized as a traffic sign by a deployed first neural network (e.g., a trained neural network), and the text "Flashing lights indicate icy conditions" can be interpreted by a deployed second neural network that informs the vehicle's path planning software (preferably executing on a CPU complex) that icy conditions exist when the flashing lights are detected. The flashing lights can be identified by a deployed third neural network operating across multiple frames, informing the vehicle's path planning software of the presence (or absence) of the flashing lights. All three neural networks can operate simultaneously, for example, within a DLA and / or on a GPU908.

[0135] In some examples, the CNN used for facial recognition and owner identification can use data from camera sensors to identify the presence of an authorized driver and / or owner of vehicle 900. A processing engine always on the sensors can be used to unlock the vehicle and turn on the lights when the owner approaches the driver's door, and in safe mode, to disable the vehicle when the owner leaves. In this way, SoC 904 provides security against theft and / or carjacking.

[0136] In another example, the CNN used for emergency vehicle detection and identification can use data from microphone 996 to detect and identify emergency vehicle siren. In contrast to conventional systems that use a general classifier to detect siren and manually extract features, SoC 904 uses a CNN to classify environmental and urban sounds as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative shut-off rate of emergency vehicles (e.g., by using the Doppler effect). The CNN can also be trained to identify emergency vehicles specific to the localized area in which the vehicle operates, as identified by GNSS sensor 958. Thus, for example, when operating in Europe, the CNN will seek to detect European siren, and when operating in the United States, the CNN will seek to identify siren only in North America. Once an emergency vehicle is detected, with the assistance of ultrasonic sensor 962, the control program can be used to execute emergency vehicle safety routines, causing the vehicle to slow down, pull over to the side of the road, stop, and / or idle until the emergency vehicle passes.

[0137] The vehicle may include a CPU 918 (e.g., a discrete CPU or dCPU) that can be coupled to the SoC 904 via a high-speed interconnect (e.g., PCIe). The CPU 918 may include, for example, an x86 processor. The CPU 918 can be used to perform any of a wide variety of functions, including, for example, arbitrating the results of potential inconsistencies between ADAS sensors and the SoC 904, and / or monitoring the status and health of the controller 936 and / or the infotainment SoC 930.

[0138] Vehicle 900 may include a GPU 920 (e.g., a discrete GPU or dGPU) that can be coupled to SoC 904 via a high-speed interconnect (e.g., NVIDIA's NVLINK). GPU 920 may provide additional artificial intelligence capabilities, for example by executing redundant and / or different neural networks, and can be used to train and / or update neural networks based on inputs from sensors of vehicle 900 (e.g., sensor data).

[0139] Vehicle 900 may further include a network interface 924, which may include one or more wireless antennas 926 (e.g., one or more wireless antennas for different communication protocols, such as cellular antennas, Bluetooth antennas, etc.). Network interface 924 can be used to enable wireless connectivity via the Internet to the cloud (e.g., with server 978 and / or other network devices), with other vehicles, and / or with computing devices (e.g., a passenger's client device). For communication with other vehicles, a direct link can be established between the two vehicles, and / or an indirect link can be established (e.g., across a network and via the Internet). A direct link can be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link can provide vehicle 900 with information about vehicles approaching vehicle 900 (e.g., vehicles in front, to the side, and / or behind vehicle 900). This functionality can be part of vehicle 900's cooperative adaptive cruise control function.

[0140] Network interface 924 may include a SoC that provides modulation and demodulation functions and enables controller 936 to communicate over a wireless network. Network interface 924 may include an RF front-end for up-conversion from baseband to RF and down-conversion from RF to baseband. Frequency conversion can be performed using known processes and / or using a superheterodyne process. In some examples, the RF front-end functionality may be provided by a separate chip. The network interface may include wireless functions for communication via LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-Wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0141] Vehicle 900 may further include data storage 928, which may include off-chip (e.g., outside of SoC 904) storage devices. Data storage 928 may include one or more storage elements, including RAM, SRAM, DRAM, VRAM, flash memory, hard disk, and / or other components and / or devices capable of storing at least one bit of data.

[0142] Vehicle 900 may further include a GNSS sensor 958. The GNSS sensor 958 (e.g., GPS, assisted GPS sensor, differential GPS (DGPS) sensor, etc.) is used for auxiliary mapping, sensing, occupancy grid generation, and / or path planning functions. Any number of GNSS sensors 958 can be used, including, for example, but not limited to, GPS using a USB connector with an Ethernet-to-serial (RS-232) bridge.

[0143] Vehicle 900 may further include a RADAR sensor 960. The RADAR sensor 960 can be used by vehicle 900 for remote vehicle detection even in dark and / or inclement weather conditions. The RADAR functional safety level may be ASIL B. The RADAR sensor 960 can use CAN and / or bus 902 (e.g., to transmit data generated by the RADAR sensor 960) for control and access to object tracking data, and in some examples, Ethernet access for accessing raw data. A wide variety of RADAR sensor types can be used. For example, and without limitation, the RADAR sensor 960 can be adapted for front, rear, and side RADAR use. In some examples, a pulse Doppler RADAR sensor is used.

[0144] The RADAR sensor 960 can include different configurations, such as long-range with a narrow field of view, short-range with a wide field of view, short-range side coverage, etc. In some examples, the long-range RADAR can be used for adaptive cruise control functions. The long-range RADAR system can provide a wide field of view (e.g., within 250m) achieved through two or more independent scans. The RADAR sensor 960 can help distinguish between stationary and moving objects and can be used by ADAS systems for emergency braking assist and forward collision warning. The long-range RADAR sensor can include a single-site multi-mode RADAR with multiple (e.g., six or more) fixed RADAR antennas and high-speed CAN and FlexRay interfaces. In an example with six antennas, the four central antennas can create a focused beam pattern designed to record the vehicle 900's surroundings at a higher rate with minimal traffic interference from adjacent lanes. The other two antennas can extend the field of view, enabling rapid detection of vehicles entering or leaving the vehicle 900's lane.

[0145] As an example, a mid-range RADAR system can include a range of up to 960m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 950 degrees (rear). Short-range RADAR systems can include, but are not limited to, RADAR sensors designed to be mounted at both ends of the rear bumper. When mounted at both ends of the rear bumper, such a RADAR sensor system can create two beams that continuously monitor blind spots behind and beside the vehicle.

[0146] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0147] Vehicle 900 may further include ultrasonic sensors 962. Ultrasonic sensors 962, which may be positioned at the front, rear, and / or sides of vehicle 900, can be used for parking assistance and / or creating and updating occupancy grids. A wide variety of ultrasonic sensors 962 can be used, and different ultrasonic sensors 962 can be used for different detection ranges (e.g., 2.5m, 4m). Ultrasonic sensors 962 can operate at functional safety level ASIL B.

[0148] Vehicle 900 may include a LIDAR sensor 964. The LIDAR sensor 964 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The LIDAR sensor 964 may be of functional safety level ASIL B. In some examples, vehicle 900 may include multiple LIDAR sensors 964 (e.g., two, four, six, etc.) that can use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0149] In some examples, the LiDAR sensor 964 may be able to provide a list of objects and their distances within a 360-degree field of view. Commercially available LiDAR sensors 964 may have an advertising range of, for example, approximately 900m, with an accuracy of 2cm-3cm, and support 900Mbps Ethernet connectivity. In some examples, one or more non-protruding LiDAR sensors 964 may be used. In such examples, the LiDAR sensor 964 may be implemented as a small device that can be embedded in the front, rear, sides, and / or corners of a vehicle 900. In such examples, the LiDAR sensor 964 may provide a horizontal field of view of up to 120 degrees and a vertical field of view of 35 degrees, even for low-reflectivity objects, with a range of 200m. Front-mounted LiDAR sensors 964 may be configured for a horizontal field of view between 45 and 135 degrees.

[0150] In some examples, LiDAR technologies such as 3D flash LiDAR can also be used. 3D flash LiDAR uses a flash of laser light as the emission source to illuminate the vehicle's surroundings up to approximately 200 meters. A flash LiDAR unit includes a receiver that records the laser pulse propagation time and reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LiDAR allows for the generation of highly accurate and distortion-free images of the surrounding environment using each laser flash. In some examples, four flash LiDAR sensors can be deployed, one on each side of the vehicle. Available 3D flash LiDAR systems include solid-state 3D staring array LiDAR cameras (e.g., non-scanning LiDAR devices) without moving parts other than a fan. Flash LiDAR devices can use 5 nanosecond Class I (eye-safe) laser pulses per frame and can capture reflected laser light in the form of a 3D range point cloud and co-registered intensity data. By using a flash LiDAR, and because a flash LiDAR is a solid-state device with no moving parts, the LiDAR sensor 964 is less susceptible to motion blur, vibration, and / or shock.

[0151] The vehicle may further include an IMU sensor 966. In some examples, the IMU sensor 966 may be located at the center of the rear axle of the vehicle 900. The IMU sensor 966 may include, for example, but not limited to, an accelerometer, a magnetometer, a gyroscope, a magnetic compass, and / or other sensor types. In some examples, such as in a six-axis application, the IMU sensor 966 may include an accelerometer and a gyroscope, while in a nine-axis application, the IMU sensor 966 may include an accelerometer, a gyroscope, and a magnetometer.

[0152] In some embodiments, the IMU sensor 966 can be implemented as a miniature, high-performance GPS-assisted inertial navigation system (GPS / INS) that combines a microelectromechanical system (MEMS) inertial sensor, a high-sensitivity GPS receiver, and an advanced Kalman filter algorithm to provide estimates of position, velocity, and attitude. Thus, in some examples, the IMU sensor 966 can enable the vehicle 900 to estimate heading by directly observing and correlating velocity changes from GPS to the IMU sensor 966 without input from a magnetic sensor. In some examples, the IMU sensor 966 and the GNSS sensor 958 can be combined into a single integrated unit.

[0153] The vehicle may include a microphone 996 placed in and / or around the vehicle 900. Among other things, the microphone 996 may be used for emergency vehicle detection and identification.

[0154] The vehicle may further include any number of camera types, including stereo camera 968, wide-angle camera 970, infrared camera 972, surround camera 974, long-range and / or mid-range camera 998, and / or other camera types. These cameras can be used to capture image data around the entire perimeter of the vehicle 900. The camera types used depend on the embodiment and the requirements of the vehicle 900, and any combination of camera types can be used to provide the necessary coverage around the vehicle 900. Furthermore, the number of cameras may vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. As an example and without limitation, these cameras may support Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the cameras is described herein with respect to... Figure 9A and Figure 9B It was described in more detail.

[0155] Vehicle 900 may further include vibration sensor 942. Vibration sensor 942 can measure vibrations of vehicle components such as axles. For example, changes in vibration can indicate changes in the road surface. In another example, when two or more vibration sensors 942 are used, differences between vibrations can be used to determine friction or slippage on the road surface (e.g., when there is a vibration difference between a power drive shaft and a free-rotating shaft).

[0156] Vehicle 900 may include ADAS system 938. In some examples, ADAS system 938 may include SoC. ADAS system 938 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward collision warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keeping assist (LKA), blind spot warning (BSW), rear cross traffic warning (RCTW), collision warning system (CWS), lane centering (LC) and / or other features and functions.

[0157] The ACC system can use a RADAR sensor 960, a LIDAR sensor 964, and / or a camera. The ACC system can include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to vehicles immediately in front of vehicle 900 and automatically adjusts the vehicle speed to maintain a safe distance. Lateral ACC performs distance holding and, if necessary, advises vehicle 900 to change lanes. Lateral ACC is associated with other ADAS applications such as LCA and CWS.

[0158] CACC uses information from other vehicles, which can be received indirectly from other vehicles via a wireless link or network connection (e.g., via the Internet) through network interface 924 and / or wireless antenna 926. Direct links can be provided by vehicle-to-vehicle (V2V) communication links, while indirect links can be infrastructure-to-vehicle (I2V) communication links. Typically, the V2V communication concept provides information about vehicles immediately ahead (e.g., vehicles immediately in front of vehicle 900 and in the same lane), while the I2V communication concept provides information about traffic further ahead. A CACC system can include either or both of these I2V and V2V information sources. Given information about vehicles ahead of vehicle 900, CACC can be more reliable, and it has the potential to improve traffic flow and reduce road congestion.

[0159] The Forward-Looking Warning (FCW) system is designed to alert the driver to hazards, enabling the driver to take corrective action. The FCW system uses a front-facing camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components. The FCW system can provide warnings in the form of, for example, audible, visual, haptic, and / or rapid braking pulses.

[0160] An AEB (Autonomous Emergency Braking) system detects an impending forward collision with another vehicle or other object and can automatically apply the brakes if the driver does not take corrective action within a specified time or distance parameter. The AEB system can use a front-facing camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA, and / or ASIC. When the AEB system detects a hazard, it typically first alerts the driver to take corrective action to avoid a collision, and if the driver does not take corrective action, the AEB system can automatically apply the brakes to attempt to prevent or at least mitigate the effects of the predicted collision. The AEB system may include technologies such as dynamic brake support and / or collision proximity braking.

[0161] The Lane Departure Warning (LDW) system provides visual, auditory, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle crosses lane markings. When the driver indicates intentional lane departure, the LDW system is deactivated by activating a turn signal. The LDW system can utilize a front-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0162] The LKA system is a variation of the LDW system. If vehicle 900 begins to leave the lane, the LKA system provides steering input or braking to correct vehicle 900.

[0163] The BSW system detects and warns the driver of vehicles in the blind spot. The BSW system can provide visual, auditory, and / or tactile alerts to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses turn signals. The BSW system can utilize a rear-facing camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0164] RCTW systems can provide visual, auditory, and / or tactile notifications when an object is detected outside the range of a rear-view camera while the vehicle is reversing. Some RCTW systems include AEB (Autonomous Emergency Braking) to ensure the application of the vehicle's brakes to avoid a collision. RCTW systems may use one or more rear-view RADAR sensors 960 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0165] Conventional ADAS systems can be prone to false positives, which can be annoying and distracting for the driver, but typically not catastrophic, as they alert the driver and allow them to determine whether a safe condition truly exists and take appropriate action. However, in an autonomous vehicle 900, in the event of conflicting results, the vehicle 900 itself must decide whether to heed the results from the main computer or auxiliary computer (e.g., the first controller 936 or the second controller 936). For example, in some embodiments, the ADAS system 938 may be a backup and / or auxiliary computer used to provide perception information to a backup computer rationality module. The backup computer rationality monitor may run redundant and varied software on hardware components to detect faults in perception and dynamic driving tasks. Outputs from the ADAS system 938 may be provided to a supervisory MCU. If the outputs from the main computer and the auxiliary computer conflict, the supervisory MCU must determine how to reconcile the conflict to ensure safe operation.

[0166] In some examples, the master computer can be configured to provide a confidence score to the supervisory MCU, indicating the master computer's confidence level in the selected result. If the confidence score exceeds a threshold, the supervisory MCU can follow the master computer's direction regardless of whether the auxiliary computer provides conflicting or inconsistent results. If the confidence score does not meet the threshold and the master and auxiliary computers indicate different results (e.g., conflict), the supervisory MCU can arbitrate between these computers to determine the appropriate result.

[0167] The supervisory MCU can be configured to run a neural network trained and configured to determine the conditions under which the auxiliary computer provides a false alarm based on outputs from both the host and auxiliary computers. Thus, the neural network in the supervisory MCU can learn when the output of the auxiliary computer can be trusted and when it cannot. For example, when the auxiliary computer is a RADAR-based FCW system, the neural network in the supervisory MCU can learn when the FCW system is identifying a metallic object that is not actually dangerous, such as a drain grid or manhole cover that triggers an alarm. Similarly, when the auxiliary computer is a camera-based LDW system, the neural network in the supervisory MCU can learn to ignore the LDW when a cyclist or pedestrian is present and lane departure is actually the safest strategy. In embodiments that include a neural network running on the supervisory MCU, the supervisory MCU may include at least one of a DLA or GPU suitable for running the neural network using associated memory. In a preferred embodiment, the supervisory MCU may include components of and / or be included as components of the SoC 904.

[0168] In other examples, ADAS system 938 may include an auxiliary computer that performs ADAS functions using conventional computer vision rules. This allows the auxiliary computer to use classic computer vision rules (if-then), and the presence of neural networks in the supervising MCU can improve reliability, safety, and performance. For example, diverse implementations and intentional non-identity make the entire system more fault-tolerant, especially for failures caused by software (or software-hardware interface) functionality. For instance, if a software vulnerability or bug exists in the software running on the host computer and non-identical software code running on the auxiliary computer provides the same overall result, the supervising MCU can be more confident that the overall result is correct and that the vulnerability in the software or hardware on the host computer does not cause a substantial error.

[0169] In some examples, the output of the ADAS system 938 can be fed to the perception block and / or the dynamic driving task block of the main computer. For example, if the ADAS system 938 issues a forward collision warning because an object is immediately in front, the perception block can use this information when identifying the object. In other examples, the assistance computer can have its own neural network, which is trained and thus reduces the risk of false positives as described herein.

[0170] Vehicle 900 may further include an infotainment SoC 930 (e.g., an in-vehicle infotainment system (IVI)). Although illustrated and described as an SoC, the infotainment system may not be an SoC and may include two or more discrete components. The infotainment SoC 930 may include a combination of hardware and software that can be used to provide vehicle 900 with audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming media, etc.), telephone (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.) and / or information services (e.g., navigation system, rear parking assistance, radio data system, vehicle-related information such as fuel level, total coverage distance, brake fuel level, fuel level, door opening / closing, air filter information, etc.). For example, the infotainment SoC 930 may include a radio, disc player, navigation system, video player, USB and Bluetooth connectivity, in-vehicle computer, in-vehicle entertainment, Wi-Fi, steering wheel audio controls, hands-free voice controls, head-up display (HUD), HMI display 934, telematics device, control panel (e.g., for controlling and / or interacting with various components, features, and / or systems) and / or other components. The infotainment SoC 930 may further be used to provide information (e.g., visual and / or auditory) to the vehicle's users, such as information from the ADAS system 938, autonomous driving information such as planned vehicle maneuvers, trajectories, surrounding environment information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0171] The infotainment SoC 930 may include GPU functionality. The infotainment SoC 930 can communicate with other devices, systems, and / or components of the vehicle 900 via bus 902 (e.g., CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 930 may be coupled to a supervisory MCU, allowing the GPU of the infotainment system to perform some autonomous driving functions in the event of a failure of the main controller 936 (e.g., the primary and / or backup computer of the vehicle 900). In such an example, the infotainment SoC 930 may place the vehicle 900 into a driver-safe parking mode as described herein.

[0172] Vehicle 900 may further include instrument cluster 932 (e.g., digital instrument panel, electronic instrument cluster, digital instrument panel, etc.). Instrument cluster 932 may include a controller and / or supercomputer (e.g., a discrete controller or supercomputer). Instrument cluster 932 may include a set of instruments such as speedometer, fuel level, oil pressure, tachometer, odometer, turn indicator, shift position indicator, seatbelt warning light, parking brake warning light, engine malfunction indicator, airbag (SRS) system information, lighting controls, safety system controls, navigation information, etc. In some examples, information may be displayed and / or shared between infotainment SoC 930 and instrument cluster 932. In other words, instrument cluster 932 may be included as part of infotainment SoC 930, or vice versa.

[0173] Figure 9D For cloud-based servers and according to some embodiments of this disclosure Figure 9A The diagram illustrates a system for communication between example autonomous vehicles 900. System 976 may include server 978, network 990, and vehicles including vehicle 900. Server 978 may include multiple GPUs 984(A)-984(H) (collectively referred to herein as GPU 984), PCIe switches 982(A)-982(H) (collectively referred to herein as PCIe switch 982), and / or CPUs 980(A)-980(B) (collectively referred to herein as CPU 980). GPUs 984, CPUs 980, and PCIe switches may be interconnected with high-speed interconnects and / or PCIe connections 986, such as, but not limited to, NVLink interface 988 developed by NVIDIA. In some examples, GPUs 984 are connected via NVLink and / or NVSwitch SoCs, and GPUs 984 and PCIe switches 982 are connected via PCIe interconnects. Although eight GPUs 984, two CPUs 980, and two PCIe switches are shown in the diagram, this is not intended to be limiting. Depending on the embodiment, each of the servers 978 may include any number of GPUs 984, CPUs 980, and / or PCIe switches. For example, each of the servers 978 may include eight, sixteen, thirty-two, and / or more GPUs 984.

[0174] Server 978 can receive image data from vehicles via network 990, representing images of unexpected or changed road conditions such as recently commenced roadworks. Server 978 can also transmit neural network 992, updated neural network 992, and / or map information 994, including information about traffic and road conditions, to vehicles via network 990. Updates to map information 994 may include updates to HD map 922, such as information about construction sites, potholes, bends, floods, or other obstacles. In some examples, neural network 992, updated neural network 992, and / or map information 994 may have been generated from new training and / or data received from any number of vehicles in the environment, and / or based on experience gained from training performed at a data center (e.g., using server 978 and / or other servers).

[0175] Server 978 can be used to train machine learning models (e.g., neural networks) based on training data. Training data can be generated by the vehicle and / or generated in a simulation (e.g., using a game engine). In some examples, the training data is labeled (e.g., where the neural network benefits from supervised learning) and / or undergoes other preprocessing, while in other examples, the training data is not labeled and / or preprocessed (e.g., where the neural network does not require supervised learning). Training can be performed according to any one or more categories of machine learning techniques, including but not limited to: categories such as supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, joint learning, transfer learning, feature learning (including principal component and cluster analysis), multilinear subspace learning, manifold learning, representation learning (including alternative dictionary learning), rule-based machine learning, anomaly detection, and any variations or combinations thereof. Once the machine learning model is trained, it can be used by the vehicle (e.g., transmitted to the vehicle via network 990), and / or the machine learning model can be used by Server 978 to remotely monitor the vehicle.

[0176] In some examples, server 978 can receive data from vehicles and apply that data to state-of-the-art real-time neural networks for real-time intelligent inference. Server 978 may include a deep learning supercomputer powered by GPU 984 and / or a dedicated AI computer, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 978 may include a deep learning infrastructure in a data center that uses only CPU power.

[0177] The deep learning infrastructure of server 978 is capable of rapid, real-time inference and can be used to assess and verify the health of the processor, software, and / or associated hardware in vehicle 900. For example, the deep learning infrastructure can receive periodic updates from vehicle 900, such as image sequences and / or objects located within those image sequences by vehicle 900 (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure can run its own neural network to identify objects and compare them to those identified by vehicle 900. If the results do not match and the infrastructure concludes that the AI ​​in vehicle 900 has malfunctioned, server 978 can transmit a signal to vehicle 900 instructing its fail-safe computer to take control, notify passengers, and complete a safe stopping operation.

[0178] For inference, the server 978 can include a GPU 984 and one or more programmable inference accelerators (such as NVIDIA's TensorRT). The combination of a GPU-powered server and inference acceleration enables real-time response. In other examples, such as where performance is less critical, CPU, FPGA, and other processor-powered servers can be used for inference.

[0179] Example computing device

[0180] Figure 10 This is a block diagram suitable for implementing some embodiments of the present disclosure of an example computing device 1000. The computing device 1000 may include an interconnect system 1002 directly or indirectly coupled to the following devices: a memory 1004, one or more central processing units (CPUs) 1006, one or more graphics processing units (GPUs) 1008, a communication interface 1010, input / output (I / O) ports 1012, input / output components 1014, a power supply 1016, one or more presentation components 1018 (e.g., displays), and one or more logic units 1020. In at least one embodiment, one or more computing devices 1000 may include one or more virtual machines (VMs), and / or any component thereof may include virtual components (e.g., virtual hardware components). For a non-limiting example, one or more GPUs 1008 may include one or more vGPUs, one or more CPUs 1006 may include one or more vCPUs, and / or one or more logic units 1020 may include one or more virtual logic units. Thus, one or more computing devices 1000 may include discrete components (e.g., a full GPU dedicated to computing device 1000), virtual components (e.g., a portion of the GPU dedicated to computing device 1000), or combinations thereof.

[0181] although Figure 10 The various blocks are shown as connected via an interconnect system 1002 with wiring, but this is not intended to be limiting and is merely for clarity. For example, in some embodiments, a presentation component 1018, such as a display device, may be considered an I / O component 1014 (e.g., if the display is a touchscreen). As another example, the CPU 1006 and / or GPU 1008 may include memory (e.g., memory 1004 may represent a storage device other than the memory of the GPU 1008, CPU 1006, and / or other components). In other words, Figure 10 The computing devices mentioned are merely illustrative. No distinction is made between categories such as "workstation," "server," "laptop," "desktop," "tablet," "client device," "mobile device," "handheld device," "game console," "electronic control unit (ECU)," "virtual reality system," and / or other device or system types, as all of these are considered within the same category. Figure 10 Within the scope of computing devices.

[0182] Interconnect system 1002 may represent one or more links or buses, such as address buses, data buses, control buses, or combinations thereof. Interconnect system 1002 may include one or more bus or link types, such as Industry Standard Architecture (ISA) bus, Extended Industry Standard Architecture (EISA) bus, Video Electronics Standards Association (VESA) bus, Peripheral Component Interconnect (PCI) bus, Peripheral Component Interconnect Fast (PCIe) bus, and / or another type of bus or link. In some embodiments, there is a direct connection between components. For example, CPU 1006 may be directly connected to memory 1004. Furthermore, CPU 1006 may be directly connected to GPU 1008. In cases where there is a direct or point-to-point connection between components, interconnect system 1002 may include a PCIe link to perform the connection. In these examples, a PCI bus is not required in computing device 1000.

[0183] The memory 1004 may include any of a wide variety of computer-readable media. Computer-readable media can be any available medium that can be accessed by the computing device 1000. Computer-readable media may include volatile and non-volatile media, as well as removable and non-removable media. For example and without limitation, computer-readable media may include computer storage media and communication media.

[0184] Computer storage media may include volatile and non-volatile media and / or removable and non-removable media, implemented in any method or technique for storing information such as computer-readable instructions, data structures, program modules, and / or other data types. For example, memory 1004 may store computer-readable instructions (e.g., representing programs and / or program elements, such as an operating system). Computer storage media may include, but is not limited to, RAM, ROM, EEPROM, flash memory or other storage technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage devices, magnetic tape cassettes, magnetic tape, disk storage devices or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by computing device 1000. As used herein, computer storage media does not include the signal itself.

[0185] Computer storage media may include computer-readable instructions, data structures, program modules, and / or other data types in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium. The term "modulated data signal" may refer to a signal whose characteristics are set or altered in a manner that encodes information into that signal. For example and without limitation, computer storage media may include wired media such as wired networks or direct wired connections, and wireless media such as sound, RF, infrared, and other wireless media. Any combination of the above should also be included within the scope of computer-readable media.

[0186] CPU 1006 may be configured to execute at least some of computer-readable instructions to control one or more components of computing device 1000 to perform one or more of the methods and / or processes described herein. Each of CPU 1006 may include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of processing a large number of software threads simultaneously. CPU 1006 may include any type of processor and may include different types of processors depending on the type of computing device 1000 implemented (e.g., processors with fewer cores for mobile devices and processors with more cores for servers). For example, depending on the type of computing device 1000, the processor may be an advanced RISC machine (ARM) processor implemented using Reduced Instruction Set Computing (RISC) or an x86 processor implemented using Complex Instruction Set Computing (CISC). In addition to one or more microprocessors or supplementary coprocessors such as math coprocessors, computing device 1000 may also include one or more CPUs 1006.

[0187] In addition to or as a replacement for CPU 1006, one or more GPUs 1008 may be configured to execute at least some of computer-readable instructions to control one or more components of computing device 1000 to perform one or more of the methods and / or processes described herein. One or more GPUs 1008 may be integrated GPUs (e.g., having one or more CPUs 1006) and / or one or more GPUs 1008 may be discrete GPUs. In embodiments, one or more GPUs 1008 may be coprocessors of one or more CPUs 1006. Computing device 1000 may use GPUs 1008 to render graphics (e.g., 3D graphics) or perform general-purpose computing. For example, one or more GPUs 1008 may be used for general-purpose computing on a GPU (GPGPU). One or more GPUs 1008 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. GPUs 1008 may generate pixel data for outputting an image in response to rendering commands (e.g., rendering commands received from CPU 1006 via a host interface). GPU 1008 may include graphics memory, such as display memory, for storing pixel data or any other suitable data, such as GPGPU data. Display memory may be included as part of memory 1004. One or more GPUs 1008 may include two or more GPUs operating in parallel (e.g., via a link). The link may be directly connected to the GPUs (e.g., using NVLINK) or connected via a switch (e.g., using NVSwitch). When combined, each GPU 1008 may generate pixel data or GPGPU data for different portions of the output or different outputs (e.g., the first GPU for the first image, the second GPU for the second image). Each GPU may include its own memory or may share memory with other GPUs.

[0188] In addition to or as an alternative to CPU 1006 and / or GPU 1008, logic unit 1020 may be configured to execute at least some of computer-readable instructions to control one or more components of computing device 1000 to perform one or more of the methods and / or processes described herein. In embodiments, CPU 1006, GPU 1008, and / or logic unit 1020 may execute any combination of methods, processes, and / or portions thereof, discretely or jointly. One or more logic units 1020 may be part of and / or integrated into one or more of CPU 1006 and / or GPU 1008, and / or one or more logic units 1020 may be discrete components or otherwise separate from CPU 1006 and / or GPU 1008. In embodiments, one or more logic units 1020 may be coprocessors of one or more CPUs 1006 and / or one or more GPUs 1008.

[0189] Examples of logic unit 1020 include one or more processing cores and / or components thereof, such as data processing unit (DPU), tensor core (TC), tensor processing unit (TPU), pixel vision core (PVC), vision processing unit (VPU), graphics processing cluster (GPC), texture processing cluster (TPC), streaming multiprocessor (SM), tree traversal unit (TTU), artificial intelligence accelerator (AIA), deep learning accelerator (DLA), arithmetic logic unit (ALU), application-specific integrated circuit (ASIC), floating-point unit (FPU), input / output (I / O) element, peripheral component interconnect (PCI) or peripheral component interconnect fast (PCIe) element, etc.

[0190] The communication interface 1010 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 1000 to communicate with other computing devices via electronic communication networks, including wired and / or wireless communications. The communication interface 1010 may include components and functions that enable communication via any of several different networks, such as wireless networks (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), wired networks (e.g., communication via Ethernet or InfiniBand), low-power wide area networks (e.g., LoRaWAN, SigFox, etc.), and / or the Internet. In one or more embodiments, one or more logic units 1020 and / or the communication interface 1010 may include one or more data processing units (DPUs) to directly transmit data received via a network and / or via interconnect system 1002 to (e.g., memory) one or more GPUs 1008.

[0191] I / O port 1012 enables computing device 1000 to be logically coupled to other devices, including I / O component 1014, presentation component 1018, and / or other components, some of which may be built into (e.g., integrated into) computing device 1000. Illustrative I / O component 1014 includes microphones, mice, keyboards, joysticks, game pads, game controllers, satellite dish antennas, scanners, printers, wireless devices, and so on. I / O component 1014 can provide a Natural User Interface (NUI) for processing user-generated air gestures, voice, or other physiological input. In some instances, the input may be transmitted to appropriate network elements for further processing. The NUI can implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, on-screen and adjacent-screen gesture recognition, air gestures, head and eye tracking, and touch recognition associated with the display of computing device 1000 (described in more detail below). Computing device 1000 may include depth cameras such as stereo camera systems, infrared camera systems, RGB camera systems, touchscreen technology, and combinations thereof for gesture detection and recognition. In addition, the computing device 1000 may include an accelerometer or gyroscope that enables motion detection (e.g., as part of an inertial measurement unit (IMU)). In some examples, the output of the accelerometer or gyroscope may be used by the computing device 1000 to render immersive augmented reality or virtual reality.

[0192] The power supply 1016 may include a hard-wired power supply, a battery power supply, or a combination thereof. The power supply 1016 may supply power to the computing device 1000 so that the components of the computing device 1000 can operate.

[0193] The presentation component 1018 may include a display (such as a monitor, touch screen, television screen, head-up display (HUD), other display types, or combinations thereof), speakers, and / or other presentation components. The presentation component 1018 may receive data from other components (such as GPU 1008, CPU 1006, DPU, etc.) and output that data (such as as images, videos, sounds, etc.).

[0194] Example Data Center

[0195] Figure 11 An example data center 1100 that may be used in at least one embodiment of this disclosure is shown. The data center 1100 may include a data center infrastructure layer 1110, a framework layer 1120, a software layer 1130, and / or an application layer 1140.

[0196] like Figure 11As shown, the data center infrastructure layer 1110 may include a resource coordinator 1112, grouped computing resources 1114, and node computing resources (“nodes CR”) 1116(1)-1116(N), where “N” represents any complete positive integer. In at least one embodiment, nodes CR 1116(1)-1116(N) may include, but are not limited to, any number of central processing units (“CPUs”) or other processors (including accelerators, field-programmable gate arrays (FPGAs), graphics processing units or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memory), storage devices (e.g., solid-state or disk drives), network input / output (“NW I / O”) devices, network switches, virtual machines (“VMs”), power modules and / or cooling modules, etc. In some embodiments, one or more node CRs from nodes CR 1116(1)-1116(N) may correspond to servers having one or more of the aforementioned computing resources. In addition, in some embodiments, nodes CR1116(1)-11161(N) may include one or more virtual components, such as vGPU, vCPU, etc., and / or one or more of nodes CR1116(1)-1116(N) may correspond to virtual machines (VMs).

[0197] In at least one embodiment, the grouped computing resources 1114 may include individual groups of nodes CR1116 housed within one or more racks (not shown), or multiple racks housed within a data center at different geographical locations (also not shown). Individual groups of nodes CR1116 within the grouped computing resources 1114 may include grouped computing, networking, memory, or storage resources that can be configured or allocated to support one or more workloads. In at least one embodiment, several nodes CR1116, including CPUs, GPUs, DPUs, and / or other processors, may be grouped within one or more racks to provide computing resources to support one or more workloads. One or more racks may also include any number of power modules, cooling modules, and / or network switches in any combination.

[0198] Resource coordinator 1122 may be configured or otherwise control one or more nodes CR1116(1)-1116(N) and / or grouped computing resources 1114. In at least one embodiment, resource coordinator 1122 may include a Software Design Infrastructure (“SDI”) management entity for data center 1100. Resource coordinator 1122 may include hardware, software, or some combination thereof.

[0199] In at least one embodiment, such as Figure 11As shown, framework layer 1120 may include a job scheduler 1133, a configuration manager 1134, a resource manager 1136, and / or a distributed file system 1138. Framework layer 1120 may include a framework for software 1132 supporting software layer 1130 and / or one or more applications 1142 of application layer 1140. Software 1132 or application 1142 may respectively contain web-based service software or applications, such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. Framework layer 1120 may be, but is not limited to, free and open-source software web application frameworks (such as Apache Spark) that can utilize distributed file system 1138 for large-scale data processing (e.g., "big data"). TM (Hereinafter referred to as "Spark") is a type of resource. In at least one embodiment, the job scheduler 1133 may include Spark drivers to facilitate the scheduling of workloads supported by different layers of data center 1100. The configuration manager 1134 may be able to configure different layers, such as software layer 1130 and framework layer 1120 (which includes Spark and distributed file system 1138 for supporting large-scale data processing). The resource manager 1136 may be able to manage compute resources mapped to or allocated to clusters of distributed file system 1138 and job scheduler 1133 or to support clusters of distributed file system 1138 and job scheduler 1133. In at least one embodiment, the clustered or grouped compute resources may include grouped compute resources 1114 in data center infrastructure layer 1110. The resource manager 1136 may coordinate with resource coordinator 1112 to manage these mapped or allocated compute resources.

[0200] In at least one embodiment, the software 1132 included in software layer 1130 may include software used in at least a portion of the nodes CRs 1116(1)-1116(N), the grouped computing resources 1114, and / or the distributed file system 1138 of framework layer 1120. One or more types of software may include, but are not limited to, internet web search software, email virus scanning software, database software, and streaming video content software.

[0201] In at least one embodiment, the application 1142 included in the application layer 1140 may include one or more types of applications used at least in part by nodes CR1116(1)-1116(N), grouped computing resources 1114, and / or the distributed file system 1138 of the framework layer 1120. One or more types of applications may include, but are not limited to, any number of genomics applications, cognitive computing, and machine learning applications, including training or inference software, machine learning framework software (e.g., PyTorch, TensorFlow, Caffe, etc.), and / or other machine learning applications used in combination with one or more embodiments.

[0202] In at least one embodiment, any of the configuration manager 1134, resource manager 1136, and resource coordinator 1112 can implement any number and type of self-modification actions based on any amount and type of data acquired in any technically feasible manner. Self-modification actions can free data center operators of data center 1100 from making potentially poor configuration decisions and may prevent underutilization and / or poor performance of the data center.

[0203] According to one or more embodiments described herein, data center 1100 may include tools, services, software, or other resources to train one or more machine learning models or to use one or more machine learning models to predict or infer information. For example, one or more machine learning models may be trained by using the software and / or computing resources described above with respect to data center 1100 to compute weight parameters according to a neural network architecture. In at least one embodiment, a trained or deployed machine learning model corresponding to one or more neural networks may be used to infer or predict information using the resources described above with respect to data center 1100 by using weight parameters computed through one or more training techniques, such as, but not limited to, those described herein.

[0204] In at least one embodiment, the data center 1100 may use a CPU, application-specific integrated circuit (ASIC), GPU, FPGA, and / or other hardware (or corresponding virtual computing resources) to perform training and / or inference using the aforementioned resources. Furthermore, one or more of the software and / or hardware resources described above may be configured to allow a user to train or perform services that infer information, such as image recognition, speech recognition, or other artificial intelligence services.

[0205] Example network environment

[0206] A network environment suitable for implementing embodiments of this disclosure may include one or more client devices, servers, network-attached storage (NAS), other backend devices, and / or other device types. Client devices, servers, and / or other device types (e.g., each device) may... Figure 10 This can be implemented on one or more instances of computing device 1000—for example, each device may include similar components, features, and / or functions of computing device 1000. Furthermore, in the case of implementing backend devices (e.g., servers, NAS, etc.), the backend devices may be included as part of data center 1100, examples of which are described in this document. Figure 11 To describe in more detail.

[0207] Components of a network environment can communicate with each other via a network, which can be wired, wireless, or both. A network can include multiple networks, or one of multiple networks. For example, a network can include one or more wide area networks (WANs), one or more local area networks (LANs), one or more public networks such as the Internet and / or the Public Switched Telephone Network (PSTN), and / or one or more private networks. In cases where the network includes a wireless telecommunications network, components such as base stations, communication towers, or even access points (and other components) can provide wireless connectivity.

[0208] A compatible network environment may include one or more peer-to-peer network environments (in which case the server may not be included in the network environment) and one or more client-server network environments (in which case one or more servers may be included in the network environment). In a peer-to-peer network environment, the server functionality described herein can be implemented on any number of client devices.

[0209] In at least one embodiment, the network environment may include one or more cloud-based network environments, distributed computing environments, combinations thereof, etc. A cloud-based network environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more servers, which may include one or more core network servers and / or edge servers. The framework layer may include a framework for supporting software at the software layer and / or one or more applications at the application layer. The software or applications may respectively include network-based service software or applications. In embodiments, one or more client devices may use network-based service software or applications (e.g., by accessing the service software and / or applications via one or more application programming interfaces (APIs)). The framework layer may be, but is not limited to, a type of free and open-source software network application framework, such as one that can use a distributed file system for large-scale data processing (e.g., "big data").

[0210] A cloud-based network environment can provide cloud computing and / or cloud storage for any combination of the computing and / or data storage functions (or one or more portions thereof) described herein. Any of these various functions can be distributed across multiple locations, such as a central or core server (e.g., distributed across one or more data centers at the state, region, country, global, etc.). If the connection to a user (e.g., a client device) is relatively close to an edge server, the core server can assign at least a portion of the functionality to the edge server. A cloud-based network environment can be private (e.g., limited to a single organization), public (e.g., available to many organizations), and / or a combination thereof (e.g., a hybrid cloud environment).

[0211] Client devices may include those described in this article. Figure 10 The example computing device 1000 described includes at least some components, features, and functions. By way of example and not limitation, the client device may be a personal computer (PC), laptop computer, mobile device, smartphone, tablet computer, smartwatch, wearable computer, personal digital assistant (PDA), MP3 player, virtual reality headset, global positioning system (GPS) or device, video player, camera, surveillance equipment or system, vehicle, boat, aircraft, virtual machine, drone, robot, handheld communication device, hospital equipment, gaming device or system, entertainment system, in-vehicle computer system, embedded system controller, remote control, electrical appliance, consumer electronics device, workstation, edge device, any combination of these devices described, or any other suitable device.

[0212] This disclosure can be described in the general context of machine-usable instructions or computer code, including computer-executable instructions such as program modules, which are executed by a computer or other machine such as a personal digital assistant or other handheld device. Typically, a program module, including routines, programs, objects, components, data structures, etc., refers to code that performs a specific task or implements a specific abstract data type. This disclosure can be practiced in a wide variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. This disclosure can also be practiced in distributed computing environments where tasks are performed by remote processing devices linked via a communication network.

[0213] As used herein, the phrase "and / or" relating to two or more elements should be interpreted as referring to only one element or a combination of elements. For example, "element A, element B, and / or element C" could include only element A, only element B, only element C, element A and element B, element A and element C, element B and element C, or elements A, B, and C. Furthermore, "at least one of element A or element B" could include at least one of element A, at least one of element B, or at least one of element A and at least one of element B. Further, "at least one of element A and element B" could include at least one of element A, at least one of element B, or at least one of element A and at least one of element B.

[0214] This document describes in detail the subject matter of this disclosure to satisfy legal requirements. However, the description itself is not intended to limit the scope of this disclosure. Rather, the inventors have envisioned that the claimed subject matter may also be embodied in other ways to include steps different from or similar combinations of steps described herein in conjunction with other current or future techniques. Moreover, although the terms "step" and / or "block" may be used herein to imply different elements of the method employed, these terms should not be construed as implying any particular order among or between the various steps disclosed herein, unless the order of the steps is explicitly described.

Claims

1. A processor, comprising: One or more circuits are used for: At least based on applying a first sensor data instance to one or more first machine learning models, when the first sensor data instance has uncertainty about whether it is classified into a first unit in the spatial unit model of the environment or into a second unit in the spatial unit model, output data is determined, the output data indicating the aggregate probability that the first sensor data instance can be classified into both the first unit and the second unit; Based at least on the aggregation probability, a cluster including the first unit and the second unit is generated; as well as At least based on associating a second sensor data instance with the cluster, the second sensor data instance is applied to one or more second machine learning models, wherein the first sensor data instance and the second sensor data instance include at least one of image data, LiDAR data, sonar data, or RADAR data.

2. The processor of claim 1, wherein the aggregation probability is the probability that the one or more first machine learning models classify sensor data depicting the second unit as depicting the first unit.

3. The processor of claim 1, wherein the aggregation probability is included in a similarity matrix, each cell in the similarity matrix corresponding to a corresponding pairwise aggregation probability, the pairwise aggregation probability being: the probability that sensor data depicting one or more first cells in the spatial cell model will be classified according to the one or more first machine learning models as corresponding to one or more second cells in the spatial cell model.

4. The processor according to claim 1, wherein the spatial unit model represents a method of dividing a geographic region into multiple units, each unit being associated with a different part of the geographic region.

5. The processor of claim 4, wherein the division is determined based on the road network density associated with the geographic region.

6. The processor of claim 1, wherein the first unit corresponds to a first country, and the second unit corresponds to a second country, the second country being different from the first country.

7. The processor of claim 1, wherein, The cluster includes a semantic region of the environment, and the second sensor data instance is applied to the one or more second machine learning models, based at least on the second sensor data instance being assigned to the semantic region.

8. The processor of claim 1, wherein the processor is included in at least one of the following: Control systems for autonomous or semi-autonomous machines; Sensing systems for autonomous or semi-autonomous machines; A system used to perform simulation operations; Systems used to perform digital twin operations; A system for performing optical transmission simulation; A system for collaborative content creation of 3D assets; A system used to perform deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system for performing conversational AI operations; A system for generating synthetic data; A system that merges one or more virtual machines (VMs); A system that is at least partially implemented in a data center; or A system that utilizes cloud computing resources at least in part.

9. A system comprising: One or more sensors; as well as One or more processing units, each including processing circuitry, said processing circuitry being used to: At least based on applying a first sensor data instance to one or more first machine learning models, when the first sensor data instance has uncertainty about whether it is classified into a first unit in the spatial unit model of the environment or into a second unit in the spatial unit model, output data is determined, the output data indicating the aggregate probability that the first sensor data instance can be classified into both the first unit and the second unit; At least one cluster is generated based on the aggregation probability, the at least one cluster comprising the first unit and the second unit; as well as At least based on associating a second sensor data instance with the cluster, the second sensor data instance is applied to one or more second machine learning models, wherein the first sensor data instance and the second sensor data instance include at least one of image data, LiDAR data, sonar data, or RADAR data.

10. The system of claim 9, wherein the output data comprises: For each instance in the first sensor data instance, indicate the probability that the instance can be classified into both the first unit and each of the remaining units of the spatial unit model by the one or more first machine learning models.

11. The system of claim 9, wherein the spatial unit model indicates that a geographic region is divided into multiple units, each unit representing a different part of the geographic region.

12. The system of claim 11, wherein the division is determined based on the road network density associated with the geographic region.

13. The system according to claim 9, wherein, The cluster represents a semantic region of the environment, which is composed of the first unit and the second unit.

14. The system according to claim 9, wherein, The second sensor data instance depicts one or more road signs of the first type, which correspond to the first unit, and the one or more road signs are used to train one or more second machine learning models, at least based on the cluster, to detect road signs of the second type corresponding to the second unit.

15. The system of claim 9, wherein the system comprises at least one of the following: Control systems for autonomous or semi-autonomous machines; Sensing systems for autonomous or semi-autonomous machines; A system used to perform simulation operations; Systems used to perform digital twin operations; A system for performing optical transmission simulation; A system for collaborative content creation of 3D assets; A system used to perform deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system for performing conversational AI operations; A system for generating synthetic data; A system that merges one or more virtual machines (VMs); A system that is at least partially implemented in a data center; or A system that utilizes cloud computing resources at least in part.

16. A method comprising: Based at least in part on the output of the second machine learning model, the self-machine performs one or more operations in the first geographic region, wherein one or more parameters of the second machine learning model are updated based at least on the following: At least based on applying a first sensor data instance to a first machine learning model, when the first sensor data instance has uncertainty about whether it is classified into a first unit in the spatial unit model of the environment or into a second unit in the spatial unit model, output data is determined, the output data indicating the aggregate probability that the first sensor data instance can be classified into both the first unit and the second unit; Based at least on the aggregation probability, a cluster including the first unit and the second unit is generated; as well as At least based on associating a second sensor data instance with the cluster, the second sensor data instance is applied to the second machine learning model, wherein the first sensor data instance and the second sensor data instance include at least one of image data, LiDAR data, sonar data, or RADAR data.

17. The method of claim 16, wherein the spatial unit model indicates dividing one or more geographic regions into multiple units, each unit being associated with a different portion of the one or more geographic regions.

18. The method of claim 16, wherein the first unit and the second unit are represented as similar at least in part based on a similarity matrix comprising pairs of similarity values ​​corresponding to a plurality of units, the similarity matrix being filled with one or more outputs of the first machine learning model.

19. The method of claim 16, wherein the first geographic region and the second geographic region correspond to different countries.

20. The method of claim 16, wherein the method is performed using at least one of the following: Control systems for autonomous or semi-autonomous machines; Sensing systems for autonomous or semi-autonomous machines; A system used to perform simulation operations; Systems used to perform digital twin operations; A system for performing optical transmission simulation; A system for collaborative content creation of 3D assets; A system used to perform deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system for performing conversational AI operations; A system for generating synthetic data; A system that merges one or more virtual machines (VMs); A system that is at least partially implemented in a data center; or A system that utilizes cloud computing resources at least in part.

Images