Feature recognition using language models for autonomous systems and applications

Abstract

The invention relates to feature recognition using language models for autonomous systems and applications. In various examples, feature recognition using a language model for autonomous and semi-autonomous systems and applications is described herein. Systems and methods described herein may use a language model to determine information associated with features (e.g., surface markers) within an environment. For example, the sensor data may be used to generate one or more images or other sensor data representations corresponding to the environment. The image may then be processed to generate input data (e.g., input tags) that is applied to the language model. Based at least on the processing of the input data, the language model may be trained to output data (e.g., output tags) representing information associated with one or more features. Further, information associated with features within the environment may be determined using the output data, where the information may then be used to update a map and / or navigate one or more machines within the environment.

Description

Background Art

[0001] In order for a vehicle (e.g., an autonomous vehicle, a semi-autonomous vehicle, a robot, etc.) to navigate safely within an environment, the vehicle must be able to effectively perform various vehicle maneuvers, such as lane keeping, lane changes, lane splitting, turning, and stopping and starting at intersections, crosswalks, and the like. For example, in order for a vehicle to navigate on surface streets (e.g., city streets, side streets, neighborhood streets, etc.) and highways (e.g., multi-lane roads), the vehicle must navigate between one or more road partitions or boundaries (e.g., lanes, intersections, crosswalks, boundaries, etc.), which are typically marked by road markings (e.g., road lines). Therefore, it is important for the vehicle to detect road markings in the environment so that the vehicle can determine how to navigate based on the rules associated with the road markings.

[0002] To detect road markings, a vehicle may use, at least in part, a map corresponding to the environment in which the vehicle is traveling. For example, a map may indicate the locations of important features that the vehicle needs to identify while traveling, such as the road surface and road markings. Traditional methods of determining the locations of road markings for these maps include using a convolutional neural network to process image data generated using an image sensor of a vehicle traveling in the environment. For example, the image data may represent an image depicting road markings within the environment. Thus, the system is capable of processing the image data, such as by using one or more image processing techniques (e.g., object detection, object recognition, etc.) using a convolutional neural network, to detect the locations of road markings in the image. The system may then use the locations of the road markings from the image to determine the corresponding locations of the road markings in the map.

[0003] While these systems are able to determine the location of road signs in an environment using convolutional neural networks, the accuracy and precision of these systems still needs to be improved. Therefore, techniques for improving the accuracy and precision of road sign location results can provide vehicles with better maps, which can also improve the vehicle's driving capabilities. Summary of the Invention

[0004] Embodiments of the present disclosure relate to feature recognition using language models for autonomous and semi-autonomous systems and applications. For example, the systems and methods described herein can use one or more language models (e.g., large language models (LLMs)) to determine information associated with features within an environment, such as road markings (e.g., lane markings, road boundaries, crosswalks, yield lines, bike lanes, etc.). For example, sensor data (e.g., image data, LiDAR data, RADAR data, ultrasonic data, etc.) can be used to generate one or more images (or other sensor data representations, such as point clouds) corresponding to the environment, such as intensity images, color images, and / or height images. The images can then be processed to generate input data (e.g., tokenized representations of feature information), which is applied to, for example, a language model, and processed by, the language model. Based at least on the processing of the input data, the language model can be trained to output data representing one or more attributes associated with one or more features (e.g., tokenized representations of feature or attribute information corresponding to the input data). For example, the output data can represent a location, color, type, shape, orientation, and / or any other attribute associated with the feature. Thus, the output data can be used to determine information associated with features within the environment, which can then be used to update a map of the environment and / or navigate one or more machines within the environment. Thus, the processes described herein can be used for offline map building or updating, and / or can be deployed to help navigate or control one or more autonomous or semi-autonomous machines.

[0005] Compared to conventional systems, in some embodiments, the systems described herein are able to more precisely and accurately determine information associated with features within an environment (e.g., road signs). This is because the current systems can use a language model (more specifically, a language model trained to determine attributes associated with such features) to determine information associated with the features, which may be more accurate than using a convolutional neural network (CNN) alone for image processing. For example, the accuracy of the language model can be improved based at least on training the language model using a specific type of input (e.g., a tokenized representation of detected feature information), such that the language model is able to generate a specific type of output, such as a tokenized representation associated with various attributes corresponding to the detected features. These outputs (which may be more accurate and precise than outputs using CNN techniques alone) can then be used to determine information associated with the features. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] The present system and method for feature recognition using language models for autonomous and semi-autonomous systems and applications are described in detail below with reference to the accompanying drawings, wherein:

[0007] Figure 1An example data flow diagram illustrating a process for feature recognition using one or more language models according to some embodiments of the present disclosure is shown;

[0008] Figure 2 shows an example of an environment including features according to some embodiments of the present disclosure;

[0009] Figures 3A to 3C shows an example of an image that may be used to identify information associated with features located within an environment according to some embodiments of the present disclosure;

[0010] Figure 4 shows an example of using one or more models to generate input data for one or more language models according to some embodiments of the present disclosure;

[0011] Figure 5 shows an example of one or more language models generating output data representing attributes associated with features located within an environment according to some embodiments of the present disclosure;

[0012] 6A to 6D shows an example of generating information associated with features located within an environment according to some embodiments of the present disclosure;

[0013] Figure 7A shows an example of updating a map to indicate information associated with features located within an environment in accordance with some embodiments of the present disclosure;

[0014] Figure 7B An example of controlling one or more machines based at least on information associated with features located within an environment is shown in accordance with some embodiments of the present disclosure;

[0015] Figure 8 shows a data flow diagram illustrating a process for training one or more language models to generate output data representing information associated with features located within an environment, according to some embodiments of the present disclosure;

[0016] Figure 9 A flow chart illustrating a method for determining a road line location using one or more language models according to some embodiments of the present disclosure is shown;

[0017] Figure 10 A flow chart illustrating a method for determining information associated with features located within an environment using one or more language models is shown, according to some embodiments of the present disclosure;

[0018] Figure 11A is an illustration of an example autonomous vehicle according to some embodiments of the present disclosure;

[0019] Figure 11BAccording to some embodiments of the present disclosure Figure 11A Examples of camera positions and fields of view for autonomous vehicles;

[0020] Figure 11C According to some embodiments of the present disclosure Figure 11A a block diagram of an example system architecture for an example autonomous vehicle;

[0021] Figure 11D According to some embodiments of the present disclosure, a method for Figure 11A System diagram of an example of communication between autonomous vehicles;

[0022] Figure 12 is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure; and

[0023] Figure 13 is a block diagram of an example data center suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0024] Systems and methods are disclosed related to feature recognition using language models for autonomous and semi-autonomous systems and applications. Although the present disclosure may be directed to an example autonomous or semi-autonomous vehicle or machine 1100 (referred to herein interchangeably as "vehicle 1100," "host vehicle 1100," "host machine 1100," or "machine 1100"), its examples are directed to 11A to 11D For example, the systems and methods described herein may be used by, but are not limited to, non-autonomous vehicles or machines, semi-autonomous vehicles or machines (e.g., in one or more adaptive driver assistance systems (ADAS)), autonomous vehicles or machines, manned and unmanned robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, flying boats, ships, space shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, construction vehicles, underwater vehicles, drones, and / or other vehicle types. Furthermore, while the present disclosure may be described with respect to feature detection in autonomous or semi-autonomous applications, this is not intended to be limiting, and the systems and methods described herein may be used in augmented reality, virtual reality, mixed reality, robotics, security and surveillance, autonomous or semi-autonomous machine applications, simulation applications, and / or any other technology field in which object or feature detection and / or map creation may be used.

[0025] For example, the system may receive sensor data generated by one or more sensors of one or more machines traveling in the environment. As described herein, the sensor data may include, but is not limited to, LiDAR data generated using one or more LiDAR sensors, image data generated using one or more image sensors (e.g., one or more cameras), RADAR data generated using one or more RADAR sensors, ultrasonic data generated using one or more ultrasonic sensors, and / or any other type of sensor data generated using any other type of sensor mode. The system may then be configured to use the sensor data to generate one or more images or other sensor data representations (e.g., point clouds) associated with the environment. As described herein, images may include, but are not limited to, intensity images (e.g., LiDAR intensity images), range images, projection images, stereo images, color images, height images, top-down images (or bird's-eye view (BEV)) images, and / or any other type of image or sensor data representation associated with the environment. In some examples, images may be captured from a particular perspective, and / or may be converted or translated to different perspectives so that the images used in the process correspond to the desired view or perspective—e.g., stereo, top-down, BEV, etc.

[0026] The system can then process the image data representing the image and generate input data, such as one or more input tags, based at least on the processing. For example, as described herein, the system can process the image data using one or more machine learning models, one or more neural networks, one or more transformers, one or more components, one or more modules, etc., which are trained to generate the input data based at least on the sensor data. The system can then apply the input data (e.g., in a labeled format) to one or more language models (e.g., one or more large language models, etc.), which are trained to determine one or more attributes, characteristics, etc. associated with one or more features located in the environment. For example, if the feature includes a road or other surface line, the attributes or characteristics can include, but are not limited to, one or more locations associated with the road line, one or more colors associated with the road line, one or more types associated with the road line, one or more shapes associated with the road line, and / or any other type of attribute associated with the road or surface line.

[0027] For example, a language model can be trained to generate output data, such as a tokenized representation, corresponding to different points associated with a feature. For example, if the feature again includes a road line, the output data can include a first set of tags associated with a first point on the road line, a second set of tags associated with a second point on the road line, a third set of tags associated with a third point on the road line, and so on. Furthermore, for a set of tags, a first tag can indicate the location of the point (e.g., x-coordinate, y-coordinate, and / or z-coordinate), a second tag can indicate the category of the point (e.g., a starting point indicating the beginning of a line segment, an intermediate point indicating a point that does not begin or end a line segment, an end point indicating a point at the end of a line segment, etc.), a third tag can indicate the type of the road line (e.g., solid line, dashed line, double line, etc.), a fourth tag can indicate the color of the road line (e.g., white, yellow, green, orange, etc.), a fifth tag can indicate the shape of the road line (e.g., straight line, curved line, etc.), and / or one or more additional tags can indicate one or more additional attributes or characteristics associated with the road line. The system can then use the output data to determine information associated with the features, such as location, type, color, shape, orientation, and / or the like, for example, by converting the tokenized representation back to image space, sensor space, and / or world space.

[0028] For example, if the output data is again associated with a road line, the system can use the first portion of the output data (e.g., a first set of tags) to determine information associated with a first point on the road line. For example, the first portion of the output data can indicate at least a first location of the first point, the first point including a starting point, the type of the road line at the first location, the color of the road line at the first point, and the shape of the road line at the first point. The system can then use the second portion of the output data (e.g., a second set of tags) to determine information associated with a second point on the road line. For example, the second portion of the output data can indicate at least a second location of the second point, the second point including an intermediate point, the type of the road line at the second location, the color of the road line at the second point, and the shape of the road line at the second point. In some examples, the language model generates the second portion of the output data after the first portion of the output data (e.g., sequentially), such that the second point is located after the first point along the road line. Therefore, the system can connect the second point to the first point using a line that includes the type associated with the second point, the color associated with the second point, and / or the shape associated with the second point.

[0029] The system can then continue to perform these processes to continue generating the road line as the output data continues to indicate additional intermediate points associated with the road line. For example, the system can continue these processes until processing a final portion of the output data indicating a final point for the road line. For example, the final portion of the output data can indicate at least the final location of the final point, including the final point, the type of road line at the final location, the color of the road line at the final point, and the shape of the road line at the final point. The system can then perform the processes described herein to connect the final point to the previous intermediate point, such that the entire line now represents the road line from the starting point to the ending point. Furthermore, the system can perform similar processes for one or more other road lines and / or other types of features located in the environment.

[0030] The system can then cause one or more actions to occur based at least on the generated information associated with the feature. For the first example, for example, if the feature includes a road line, the system can update the map to indicate the location, type, color, and / or shape of the road line. For the second example, for example, if the system is associated with a machine traveling in the environment, and the feature again includes a road line, the system can cause the machine to travel based at least on the location, type, color, and / or shape associated with the road line.

[0031] In some examples, the system (and / or another system) can train a language model to perform one or more processes described herein. For example, the system can use training sensor data to generate training images or other sensor data representations, where the training images can include intensity images, color images, height images, etc. The system can then process the training image data representing the training images to generate training input data. For example, the system can use one or more machine learning models, one or more neural networks, one or more transformers, one or more components, one or more modules, etc. to process the training image data, which are trained to generate training input data based at least on the training image data. As described herein, the training input data can also include input tags. The system can then apply the training input data to the language model and, based at least on applying the training input data, receive training output data from the language model. As described herein, the training output data can represent attributes associated with features represented by the training image data. For example, the training output data can represent output tags associated with the attributes.

[0032] The system (e.g., a training engine) can then compare the training output data with ground truth data representing ground truth attributes associated with the features. For example, the ground truth data can represent ground truth labels associated with the ground truth attributes. In some examples, the system generates the ground truth labels by processing one or more ground truth images indicating the ground truth position, type, color, and / or shape of the features. Based at least on comparing the training output data with the ground truth output data, the system can then update the language model. For example, the system can determine one or more losses based at least on comparing the training output data with the ground truth output data, and then use the losses to update the parameters and / or weights of the language model.

[0033] The systems and methods described herein can be used by, but are not limited to, non-autonomous vehicles or machines, semi-autonomous vehicles or machines (e.g., in one or more adaptive driver assistance systems (ADAS)), autonomous vehicles or machines, manned and unmanned robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles connected to one or more trailers, aircraft, boats, space shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, airplanes, construction vehicles, underwater vehicles, drones, and / or other vehicle types. In addition, the systems and methods described herein can be used for various 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 monitoring, simulation and digital twins, autonomous or semi-autonomous machine applications, deep learning, environmental simulation, object or participant simulation and / or digital twins, data center processing, conversational artificial intelligence (AI), light transport simulation (e.g., ray tracing, path tracing, etc.), collaborative content creation of 3D assets, cloud computing, and / or any other suitable application.

[0034] The disclosed embodiments can 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, systems for performing digital twin operations, systems implemented using edge devices, systems implementing large language models (LLMs), systems including one or more virtual machines (VMs), systems for performing synthetic data generation operations, systems implemented at least in part in a data center, systems for performing conversational AI operations, systems for performing light transport simulations, systems for performing collaborative content creation of 3D assets, systems for performing generative AI operations, systems implemented at least in part using cloud computing resources, and / or other types of systems.

[0035] refer to Figure 1, Figure 1 An example data flow diagram of a process 100 for feature recognition using a language model according to some embodiments of the present disclosure is shown. It should be understood that this and other arrangements described herein are presented as examples only. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of the arrangements and elements shown, and some elements may be omitted entirely. In addition, many of the elements described herein are functional entities that may be implemented as discrete or distributed components or in combination with other components, and in any suitable combination and location. The various functions performed by the entities described herein may be performed by hardware, firmware, and / or software. For example, the various functions may be performed by a processor executing instructions stored in a memory. In some embodiments, the systems, methods, and processes described herein may be implemented using hardware, firmware, and / or software. 11A to 11D The example autonomous vehicle 1100, the example computing device 1200 of FIG. 1200, and / or Figure 13 may perform similar components, features, and / or functions as the example data center 1300.

[0036] Process 100 may include a representation component 102 receiving sensor data 104 generated using one or more sensors 106 of one or more machines (e.g., one or more autonomous vehicles 1100). As described herein, sensor data 104 may include, but is not limited to, LiDAR data generated using one or more LiDAR sensors, image data generated using one or more image sensors (e.g., one or more cameras), RADAR data generated using one or more RADAR sensors, and / or any other type of sensor data generated using any other type of sensor. In some examples, sensor data 104 may be generated using one or more machines that have previously navigated an environment, such as when process 100 is associated with updating a map (such as for example, a map of a vehicle). Figure 7A In some examples, sensor data 104 may be generated by a machine currently traveling in an environment, such as when process 100 is associated with driving a machine (these are specific to Figure 7B In any example, sensor data 104 can represent an environment, such as features located within the environment. As described herein, features can include, but are not limited to, road features (e.g., road markings, such as road lines, roads, traffic signs, traffic signals, etc.), structures (e.g., buildings, etc.), objects (e.g., traffic signs, utility poles, traffic lights, warehouse objects, etc.), vehicles, pedestrians, and / or any other type of object or feature that may be located within the environment.

[0037] For example, Figure 2An example of an environment 202 including features according to some embodiments of the present disclosure is shown. As shown, the environment 202 includes at least a first road 204(1) and a second road 204(2), wherein the first road 204(1) is at least partially located above the second road 204(2) (e.g., the first road 204(1) includes an overpass). The first road 204(1) also includes at least road lines 206(1)-(5), while the second road 204(2) includes road lines 206(6)-(8). In some examples, the roads 204(1)-(2) (which may also be referred to as "roads 204" in the singular or "roads 204" in the plural) and / or the road lines 206(1)-(8) (which may also be referred to as "road lines 206" in the singular or "road lines 206" in the plural) may also be referred to as "features" of the environment 202. Additionally, although Figure 2 While the environment 202 in the example of FIG. 1 is shown only as including the road 204 and the road line 206 , in other examples, the environment 202 may include additional and / or alternative features (eg, road signs, road signals, other road markings, structures, vehicles, pedestrians, etc.).

[0038] Return Reference Figure 1 In an example of a sensor data component 104, process 100 may include a representation component 102 generating representation data 108 using at least a portion of sensor data 104, the representation data 108 representing one or more representations associated with an environment. As described herein, in some examples, a representation may include one or more images, such as an intensity image, a color image, a height image, and / or one or more other representations (e.g., a point cloud) that depict or otherwise correspond to information associated with the environment. Furthermore, in examples where the representation includes an image, the image may include a top-down image of the environment (e.g., a bird's-eye view), a stereoscopic view of the environment, and / or another perspective or view depending on the particular embodiment.

[0039] For example, such as when the sensor data 104 includes at least LiDAR data, the LiDAR data (and / or a point cloud associated with the LiDAR data) can represent an intensity associated with at least a portion of a point within an environment. For example, as described herein, a LiDAR sensor used to generate the LiDAR data can measure the intensity of the point as light returns to the LiDAR sensor. In some examples, the intensity can be represented using a number, such as a number between 0 and 256 (although other ranges can be used in other examples), where the number varies depending on the composition of the surface (e.g., color, texture, material, etc.) from which the light is reflected. For example, a low number can represent low reflectivity, while a high number represents high reflectivity. In some examples, the intensity can depend on other factors, such as the angle of arrival, the range of the point, the moisture content, etc.

[0040] Thus, the representation component 102 can process the LiDAR data to generate one or more intensity images using the intensities of the points, wherein the intensity images represent surfaces within the environment that reflect light. For example, the image can include a top-down (BEV) image representing at least one or more surfaces within the environment. In such an example, and as described herein, because the intensities of the points can vary based on one or more factors, such as the color of the surface associated with the point (e.g., the color of the surface that reflects light), the image can indicate the structure of the road line.

[0041] Furthermore, in some examples, such as when the sensor data 104 also includes image data, the representation component 102 can use the image data to determine color information associated with each point represented by the LiDAR data. The representation component 102 can then use the point positions and color information from the LiDAR data to generate one or more color images associated with the environment. Furthermore, in some examples, the representation component 102 can use the sensor data 104 (e.g., LiDAR data, RADAR data, image data, etc.) representing the distance to each point within the environment to determine height information associated with the point within the environment. The representation component 102 can then use the height information to generate one or more height images associated with the environment.

[0042] For example, Figures 3A to 3B 2 shows an example of an image that may be used to identify information associated with features located within environment 202 according to some embodiments of the present disclosure. Figure 3A As shown in the example of , representation component 102 can generate an intensity image 302 associated with environment 202. In some examples, because intensities 304 of points associated with road line 206 within environment 202 (although only one is labeled for clarity) can be different from intensities 306 of points associated with other portions of environment 202 (although again only one is labeled for clarity), intensity image 302 can indicate locations associated with road line 206 within environment 202. For example, points of intensity image 302 associated with road line 206 within environment 202 can include a different color than points associated with other portions of environment 202.

[0043] like Figure 3B As shown in the example of , the representation component 102 can generate a color image 308 associated with the environment 202. As described herein, the representation component 102 can generate the color image 308 using LiDAR data and / or image data. Figure 3CAs shown in the example of , the representation component 102 can generate a height map 310 associated with the environment 202. As described herein, the representation component 102 can generate the height map 310 using LiDAR data, RADAR data, image data, and / or any other type of distance data. Additionally, in some examples, such as Figure 3C As shown in the example of , because road 204 may include at least a slightly different elevation than surrounding portions of environment 202 , elevation map 310 may indicate at least the location of road 204 within environment 202 .

[0044] Return Reference Figure 1 In an example, process 100 may include processing representation data 108 using one or more models 110 and, based at least on the processing, generating input data 112 for one or more language models 114. As described herein, in some examples, input data 112 may represent labels (e.g., tokenized representations) corresponding to features located within an environment. For example, if process 100 is associated with determining information associated with road lines, the labels may be associated with information corresponding to roads, lanes, road lines, etc. within the environment. In some examples, model 110 may include any type of machine learning model, neural network, etc., configured to generate input data 112 based at least on processing representation data 108. For example, the model may include a convolutional neural network, a feedforward neural network, a spatially invariant artificial neural network, a recurrent neural network, a perceptron, a transformer, and / or any other type of network.

[0045] For example, Figure 4 An example of using model 110 to generate input data for language model 114 according to some embodiments of the present disclosure is shown. As shown, one or more convolutional neural networks (CNNs) 402 can process representation data 404 (which can represent and / or include representation data 108). Based at least on the processing, CNN 402 can extract features (e.g., roads, lanes, road lines, etc.), where the features can be represented by feature data 406. Furthermore, in some examples, based at least on the processing, CNN 402 can generate a heat map, which can be represented by heat map data 408, where the heat map indicates at least locations associated with the features. For example, if the features include road lines, the heat map (e.g., in the form of a segmentation mask) can be configured to indicate the locations of the road lines in the environment.

[0046] The transformer 410 can then process the feature data 406 and / or the heat map data 408. Based at least on this processing, the transformer 410 can generate one or more tags 412 (which can represent and / or include the input data 112) for the language model 114. As described herein, the tags 414 can represent information associated with the features, such as roads, lanes, road lines, etc. In addition, the tags 414 can be in a format that the language model 114 is trained to process.

[0047] Return Reference Figure 1 In an example, process 100 may include applying input data 112 (e.g., input tokens) to a language model 114. As described herein, language model 114 may include any type of language model, such as a statistical language model, a neural language model, a probabilistic language model, a large language model, and the like. Language model 114 may then be trained to process input data 112 and, based at least on the processing, generate output data 116 representing information associated with features corresponding to representation data 108. As described herein, in some examples, output data 116 may represent tokens (e.g., tokenized representations) corresponding to attributes or characteristics associated with the features, such as a location, color, type, orientation, size, and the like associated with the features. For example, if the features include road lines, the attributes may include, but are not limited to, location (e.g., two-dimensional location, three-dimensional location, and the like), color (e.g., white, yellow, orange, and the like), type (single solid line, dashed line, double solid line, double dashed line, and the like), shape (e.g., straight line, curved line, and the like), and / or any other type of attribute or characteristic associated with road lines.

[0048] For example, the language model 114 can be trained to generate output data 116 that represents a corresponding set of labels for one or more points (e.g., each point) associated with a road line. For example, the output data 116 can include a first set of labels associated with a first point on the road line, a second set of labels associated with a second point on the road line, a third set of labels associated with a third point on the road line, and so on. In addition, for a set of labels, the first label can indicate the location of the point (e.g., x-coordinate, y-coordinate, and / or z-coordinate), the second label can indicate the category of the point (e.g., starting point, middle point, end point, etc.), the third label can indicate the type of the road line (e.g., solid line, dashed line, double line, etc.), the fourth label can indicate the color (e.g., white, yellow, orange, etc.), the fifth label can indicate the shape of the road line (e.g., straight line, curved line, etc.), and / or one or more additional labels can indicate one or more additional attributes.

[0049] In these examples, language model 114 can generate output data 116 using a sequence. For example, language model 114 can generate a first set of tokens, then a second set of tokens, then a third set of tokens, and so on. Furthermore, in some examples, language model 114 can be configured to generate a single token in different instances, such that tokens included in a set of tokens are generated at different time instances. Additionally, or alternatively, in some examples, language model 114 can be configured to generate a set of tokens in a single instance, such that different layers generate different tokens corresponding to different attributes of the set of tokens.

[0050] For example, Figure 5 An example of one or more language models 502 (which may represent and / or include language model 114) generating output data representing attributes associated with one or more features according to some embodiments of the present disclosure is shown. As shown, based at least on processing input tokens 412, language model 502 may include at least a location component 504 configured to generate one or more location tokens 506, a category component 508 configured to generate one or more category tokens 510, a color component 512 configured to generate one or more color tokens 514, a type component 516 configured to generate one or more type tokens 518, and a shape component 520 configured to generate one or more shape tokens 522. As described herein, components 504, 508, 512, 516, and / or 520 may include, but are not limited to, one or more transformer modules, one or more layers, one or more headers, and / or any other components associated with language model 502.

[0051] exist Figure 5 In the example of FIG5 , the language model 502 can be configured to generate a set of labels 506, 510, 514, 518, and / or 522 for points associated with a feature (e.g., a road line) in a single instance. For example, the language model 502 can generate a first set of labels for a first point associated with the feature, then a second set of labels for a second point associated with the feature, then a third set of labels for a third point associated with the feature, and so on. However, in other examples, the language model 502 can generate the labels 506, 510, 514, 518, and / or 522 separately. For example, the language model 502 can generate a location label 506 for a first set of labels for a first point associated with the feature, then a category label 510, then a color label 514, then a type label 518, and then a shape label 522. The language model 502 can then perform a similar process for a second set of labels for a second point associated with the feature, then a third set of labels for a third point associated with the feature, and so on.

[0052] In some examples, markers 506, 510, 514, 518, and / or 522 can be associated with various latent spaces. For example, position marker 506 can be associated with a first latent (or embedding) space, category marker 510 can be associated with a second latent space, color marker 514 can be associated with a third latent space, type marker 518 can be associated with a fourth latent space, and shape marker 522 can be associated with a fifth latent space. Thus, markers 506, 510, 514, 518, and / or 522 can be interpreted to determine one or more attributes described herein. For the first example, first category marker 510 can be associated with a starting point, second category marker 510 can be associated with an intermediate point, and third category marker 510 can be associated with an end point. For the second example, first color marker 514 can be associated with white, second color marker 514 can be associated with yellow, and so on.

[0053] Return Reference Figure 1 In the example of , process 100 may include one or more processing components 118 processing output data 116 and generating feature information 120 based at least on the processing. In some examples, feature information 120 may represent information associated with features within the environment. For example, if the features include road lines, feature information 120 may represent the location of the road lines, the type of the road lines, the color of the road lines, the shape of the road lines, and / or any other attributes associated with the road lines. As described herein, in some examples, processing component 118 may use a sequence to process output data 116 as language model 114 generates it.

[0054] For example, if the output data 116 is again associated with a road line, the processing component 118 can use the first portion of the output data 116 (e.g., the first set of markers) to determine information associated with a first point on the road line. For example, the first portion of the output data 116 can indicate at least a first location of the first point, the first point including a starting point, the type of road line at the first location, the color of the road line at the first point, and the shape of the road line at the first point. The processing component 118 can then use the second portion of the output data 116 (e.g., the second set of markers) to determine information associated with a second point on the road line. For example, the second portion of the output data 116 can indicate at least a second location of the second point, the second point including a midpoint, the type of road line at the second location, the color of the road line at the second point, and the shape of the road line at the second point. Thus, the processing component 118 can then connect the second point to the first point using a line that includes the type, color, and / or shape associated with the second point and / or the first point.

[0055] Then, because the output data 116 continues to indicate additional intermediate points associated with the road line, the processing component 118 can continue to perform these processes to continue generating the road line. For example, the processing component 118 can continue these processes until processing a final portion of the output data 116 (e.g., a final set of markers) indicating a final point of the road line. For example, the final portion of the output data 116 can indicate at least the final location of the final point, including the end point, the type of road line at the point location, the color of the road line at the final point, and the shape of the road line at the final point. The processing component 118 can then perform the processes described herein to connect the final point to the previous intermediate point, such that the entire line now represents a road line from the start point to the end point. Furthermore, the processing component 118 can perform similar processes for one or more other road lines located within the environment.

[0056] For example, 6A to 6D 2 shows an example of generating information 602 associated with a road line 206 located within an environment 202 according to some embodiments of the present disclosure. Figure 6A As shown in the example of , the processing component 118 can receive first output data (e.g., a first set of labels) associated with a first point 604(1) corresponding to a first road line 206(1). As described herein, the first output data can indicate a first location of the first point 604(1), the first point 604(1) including a starting point, a type of the first road line 206(1) at the first location, a color of the first road line 206(1) at the first point 604(1), and a shape of the first road line 206(1) at the first point 604(1). Thus, the processing component 118 can begin by inputting information associated with the first point 604(1). 6A to 6D In the example of , the white circle point may indicate the starting point of the road line 206 .

[0057] Next, if Figure 6B As shown in the example of , the processing component 118 can receive second output data (e.g., a second set of markers) associated with a second point 604(2) corresponding to the first road line 206(1). As described herein, the second output data can indicate a second location of the second point 604(2), that the second point 604(2) includes an intermediate point, the type of the first road line 206(1) at the second location, the color of the first road line 206(1) at the second point 604(2), and the shape of the first road line 206(1) at the second point 604(2). Thus, the processing component 118 can input information associated with the second point 604(2). 6A to 6DIn the example of FIG, the gray circle point can indicate the middle point of the road line 206. In addition, the processing component 118 can connect the second point 604(2) to the first point 604(1) using the first line 606(1). As described herein, the first line 606(1) can include a shape based on the second output data, such as a straight line, a curve (e.g., a Bezier curve), and / or any other shape of line. The first line 606(1) can also include a color and / or type indicated by the second output data.

[0058] Next, if Figure 6C As shown in the example of FIG. 6 , processing component 118 can continue to perform these processes using additional output data associated with additional intermediate points to input points 604 (3)-(7) and lines 606 (2)-(6) connecting points 604 (2)-(7). As described herein, processing component 118 can continue to perform these processes on the intermediate points until output data associated with an endpoint is received.

[0059] For example, the processing component 118 can receive eighth output data (e.g., an eighth set of markings) associated with an eighth point 604(8) corresponding to the first road line 206(1). As described herein, the eighth output data can indicate an eighth position of the eighth point 604(8), the eighth point 604(8) including an end point, a type of the first road line 206(1) at the eighth position, a color of the first road line 206(1) at the eighth point 604(8), and a shape of the first road line 206(1) at the eighth point 604(8). Thus, the processing component 118 can input information associated with the eighth point 604(8). 6A to 6D In the example of FIG, the black circle point can indicate the end point of the road line 206. In addition, the processing component 118 can connect the eighth point 604(8) to the seventh point 604(7) using the seventh line 606(7). Thus, the processing component 118 can generate a line associated with the first road line 206(1).

[0060] Next, if Figure 6D As shown in the example of , processing component 118 can continue to perform these processes to generate lines 608 ( 1 )-( 8 ) (also referred to as “line 608 ” or “lines 608 ”) for each road line 206 . As shown, based at least on performing the processes described herein, lines 608 can be similar to road lines 206 .

[0061] In some examples, language model 114 can be configured to generate a different number of points associated with information 602. For example, the grid size associated with language model 114 can be configured such that language model 114 generates a greater number of points as the grid size increases. Figure 6DAs shown in the example of FIG, lines 606(6)-(8) associated with second road 204(2) at least partially located below first road 204(1) indicate content that road lines 206(6)-(8) may include even though they are below an overpass. In such an example, language model 114 may be trained to generate portions of lines 606(6)-(8) even when these portions are occluded.

[0062] Return Reference Figure 1 For example, although Figure 1 The example of FIG10 shows model 110 and processing component 118 as separate from language model 114, but in other examples, model 110 and / or processing component 118 may include at least a portion of language model 110. For example, model 110 and / or processing component 118 may include one or more layers of language model 114.

[0063] As described herein, the feature information 120 may then be used to perform one or more processes. For example, Figure 7A An example of updating a map 702 to indicate information associated with one or more features located within an environment 202 is shown in accordance with some embodiments of the present disclosure. As shown, a mapping component 704 can update a map 702 associated with the environment 202 using feature information 120. As shown, the map 702 can be updated to indicate at least road line representations 706(1)-(8) (also referred to as "road line representation 706" in the singular and / or "road line representations 706" in the plural) that are associated with the road line 206, respectively. By performing the processes described herein, the road line representations 706 can include the same type, color, shape, and / or any other attributes associated with the road line 206.

[0064] also, Figure 7B An example of controlling one or more machines based at least on information associated with one or more features located within environment 202 according to some embodiments of the present disclosure is shown. As shown, a driving stack 708 associated with a machine 710 (e.g., autonomous vehicle 1100) can use feature information 120 to determine a trajectory 712 for the machine 710 to drive within environment 202. For example, the driving stack 708 can determine the trajectory 712 so that the machine 710 complies with one or more road rules regarding road lines 206. The driving stack 708 can then cause the machine 710 to drive according to the trajectory 712.

[0065] For more details on determining information associated with features, a goal of process 100 may be to detect lane entities in a given input scene and discern topological relationships between lane entities. Thus, input associated with process 100 may include one or more representations (e.g., images) described herein. For example, a lane map may be generated by Definition, where V represents a set of lanes in the scene {l1,l2,…,l L Then, each lane l i A set of ordered 3D vertices Composition, of which and V i Yes i In addition, each lane l i and vertex v i Can have additional properties attached to it.

[0066] therefore, represents the edge set, symbolizing the topological relationship between lanes. Then the topology can be represented by the adjacency matrix, where only when lane l i The end point is connected to lane l j The entry (i, j) is set to 1 only when the starting point is reached.

[0067] Therefore, in order to generate a lane map using the transformer decoder, the lane map can be represented as a sequence of discrete markers. To define the markers, the lane graph can be decomposed into its component lanes {l1,l2,…,l L}, and each lane l i Represented as a sequence of key points Here, N i is the number of keypoints in the lane, and Thus, the key points of a lane may include its endpoints and their intersections with fixed equidistant grid lines along the X and Y axes. Furthermore, the key points may be sorted along the direction of the lane.

[0068] Next, you can The lanes in are serialized into a single sequence. This can be achieved by concatenating the keypoints of one or more (e.g., all) lanes together, resulting in a sequence of length N, where Additionally, a class label can be introduced for each keypoint to act as a lane delimiter. This is probably because a method is needed to decompose the sequence back into individual lanes. Therefore, two class labels are added, <eol>represents the last keypoint in each lane, <vtx>Represents each other key point.

[0069] The tokenization scheme described in this paper can produce the sequence [s1,s2,…,s N ], where s i =(c i ,x i ,y i ,z i ) represents a key point in the lane graph serialization. Therefore, c i ,x i ,y i , Represent the category label and coordinate label respectively. Therefore, the sequence can be represented by the matrix express.

[0070] In some examples, a random depth search traversal of the lane directed graph can be used to sort the lanes in the concatenated sequence. More specifically, a lane can be randomly sampled and its keypoints added to the sequence. Next, a depth-first traversal can be performed starting from that lane, and the keypoints of each visited lane can be added in the traversal order. This process can then be repeated until all lanes are visited. All lanes in the.

[0071] Regarding the decoder architecture, given a scene I and one or more (e.g., all) previously generated keypoint labels S [1：t-1] =[s1,s2,…,s t-1 ], the decoder can be trained to predict the tag s of the next keypoint i =(c i ,x i ,y i ,z i ) and s t-1 With s t Bezier control points of the segments between Here, C is the number of Bezier control points. Therefore, the predicted probability at step t can be decomposed into conditional probabilities:

[0072] p(c t ,x t ,y t ,z t ,b t |S [1：t-1] ,I)=p(b t |c t ,x t ,y t ,z t ,S [1：t-1] ,I)

[0073] ×p(z t |c t ,x t ,y t ,S [1：t-1] ,I)

[0074] ×p(y t |c t ,x t ,S [1：t-1] ,I)

[0075] ×p(x|c t ,S [1：t-1] ,I)

[0076] ×p(c t |S [1：t-1] ,I)

[0077] In this cascaded decoder architecture, each of these conditional probabilities can be modeled by a transformer decoder. In addition, each transformer decoder cross-attentions the input scene code F I , followed by the MLP predictor head.

[0078] In some examples, language model 114 can be trained to perform one or more of the processes described herein. For example, Figure 8 A data flow diagram illustrating a process 800 for training one or more language models 802 (which may represent and / or include language model 114 and / or language model 502 ) is shown, in accordance with some embodiments of the present disclosure.

[0079] As shown, language model 802 can be trained using input data 804. In some examples, input data 804 can be similar to input data 112, representation data 108, and / or tags 412. For example, input data 804 can represent tags corresponding to features within an environment. For example, input data 804 can be generated using one or more processes similar to input data 112 and / or tags 412 described herein (e.g., using representation component 102).

[0080] The language model 802 can be trained using training input data 804 and corresponding ground truth data 806. In some examples, the ground truth data 806 can include labels corresponding to various attributes associated with the features. For example, the ground truth data 806 can include a label indicating the location 808 of a point, a label indicating the type 810 associated with the point (e.g., straight line, dashed line, double line, etc.), a label indicating the color 812 associated with the point (e.g., white, yellow, orange, etc.), a label indicating the shape 814 associated with the point (e.g., straight line, curved line, etc.), and a label indicating the category 816 associated with the point (e.g., starting point, middle point, end point, etc.). As described herein, the ground truth data 806 can be synthetically generated (e.g., generated from a computer model or rendering), realistically generated (e.g., designed and generated from real-world data), machine-automated (e.g., using feature analysis and learning to extract features from the data and then generate labels), manually annotated (e.g., annotators or annotation experts define the location of the labels), and / or a combination thereof. In some examples, for each instance of the input data 804, there can be corresponding ground truth data 806.

[0081] like Figure 8 As further shown, training engine 818 can use one or more loss functions to measure the loss (e.g., error) of output 820 compared to ground truth data 806. In some examples, output 818 can be similar to output data 116 and / or labels 506, 510, 514, 518, and / or 522. For example, output 818 can include labels corresponding to various attributes associated with the features. Any type of loss function can be used, such as cross-entropy loss, mean squared error, mean absolute error, mean deviation error, and / or other loss function types. In some examples, different outputs 818 can have different loss functions. For example, position can include a first loss function, type can include a second loss function, color can include a third loss function, shape can include a fourth loss function, and category can include a fifth loss function. In such examples, the loss functions can be combined to form a total loss, and the total loss can be used to train language model 802 (e.g., to update parameters of language model 802). In any example, a backward pass calculation can be performed to recursively calculate the gradient of the loss function with respect to the training parameters. In some examples, these gradients can be calculated using the weights and biases of language model 802 .

[0082] For example, the loss can be decomposed into at least class labeling, coordinate labeling, and Bessel coefficient regression loss, where:

[0083]

[0084] Here, λ C ,λ X ,λ Y ,λ Z and λ reg Can include scalar loss values, and The conventional cross entropy loss for labeled classification tasks can be included. In addition, An L1 loss may be included for lane segment regression. In some examples, to compute P equidistant points can be sampled along the original target lane segment and the predicted Bezier curve.

[0085] Now refer to Figure 9 and Figure 10 , each block of methods 900 and 1000 described herein includes a computing process that can be performed using any combination of hardware, firmware, and / or software. For example, the various functions can be performed by a processor executing instructions stored in a memory. Methods 900 and 1000 can also be embodied as computer-usable instructions stored on a computer storage medium. Methods 900 and 1000 can be provided by a standalone application, a service or a hosted service (standalone or in combination with another hosted service), or a plug-in to another product, to name a few. In addition, methods 900 and 1000 are relative to Figure 1 However, these methods 900 and 1000 may additionally or alternatively be performed by any one system or any combination of systems, including but not limited to the systems described herein.

[0086] Figure 9 A flow chart of a method 900 for determining the location of a road line using one or more language models, according to some embodiments of the present disclosure, is shown. Method 900, at block B902, may include generating one or more representations corresponding to an environment based at least on sensor data. For example, representation component 102 may generate representation data 108 using at least sensor data 104. As described herein, representation data 108 may represent one or more representations corresponding to the environment, such as one or more images representing the environment. In some examples, the images may include at least an intensity image, a color image, a height image, and / or any other type of image.

[0087] At block B904, method 900 may include generating one or more input tags based at least on the one or more representations, the input tags representing one or more road lines associated with the environment. For example, model 110 may process representation data 108 and, based at least on the processing, generate input data 112 representing the one or more input tags. As described herein, the input tags may be associated with road lines corresponding to representation data 108.

[0088] At block B906, method 900 may include processing one or more input tokens based at least on one or more language models to generate one or more output tokens representing one or more attributes associated with one or more road lines. For example, language model 114 may process input data 112 and, based at least on the processing, generate output data 116 representing one or more output tokens. As described herein, language model 114 may generate a corresponding set of tokens for one or more (e.g., each) point associated with a road line. Furthermore, within a set of tokens, each token may be associated with one or more attributes. For example, a first token may indicate the location of the point (e.g., x-coordinate, y-coordinate, and / or z-coordinate), a second token may indicate the category of the point (e.g., starting point, middle point, end point, etc.), a third token may indicate the type of road line (e.g., solid line, dashed line, double line, etc.), a fourth token may indicate a color (e.g., white, yellow, orange, etc.), a fifth token may indicate a shape (e.g., straight line, curved line, etc.), and / or one or more additional tokens may indicate one or more additional attributes.

[0089] At block B908, method 900 may include determining one or more locations associated with one or more road lines within the environment based at least on the one or more output markers. For example, processing component 118 may use at least output data 116 to determine locations associated with the road lines. As described herein, processing component 118 may determine locations for sequential points associated with output data 116. Furthermore, in some examples, processing component 118 may determine additional information associated with the road lines, such as the type, color, and / or shape of the road lines.

[0090] Figure 10 A flow chart of a method 1000 for determining information associated with features located within an environment using one or more language models is shown in accordance with some embodiments of the present disclosure. Method 1000 may include, at block B1002, generating one or more representations corresponding to the environment based at least on sensor data. For example, representation component 102 may generate representation data 108 using at least sensor data 104. As described herein, representation data 108 may represent one or more representations corresponding to the environment, such as one or more images representing the environment. In some examples, the images may include at least an intensity image, a color image, a height image, and / or any other type of image.

[0091] Method 1000 may include, at block B1004, generating input data corresponding to one or more features associated with the environment based at least on the one or more representations. For example, model 110 may process representation data 108 and, based at least on the processing, generate input data 112. As described herein, in some examples, input data 112 may represent one or more input tags associated with the features.

[0092] Method 1000 may include, at block B1006, processing input data based on one or more language models to generate output data representing one or more attributes associated with one or more features. For example, language model 114 may process input data 112 and, based at least on the processing, generate output data 116 representing attributes associated with the features. As described herein, output data 116 may represent one or more output tokens associated with the attributes.

[0093] Method 1000 may include determining information associated with one or more features based at least on the output data at block B1008. For example, processing component 118 may determine information associated with a feature using at least output data 116. As described herein, the information may include one or more locations of the feature, one or more types associated with the feature, one or more colors associated with the feature, one or more shapes associated with the feature, and / or any other information associated with the feature.

[0094] Example autonomous vehicle

[0095] Figure 11A is an illustration of an example autonomous vehicle 1100 according to some embodiments of the present disclosure. Autonomous vehicle 1100 (alternatively referred to herein as "vehicle 1100") may include, but is not limited to, passenger vehicles such as cars, trucks, buses, first responder vehicles, shuttles, electric or motorized bicycles, motorcycles, fire trucks, police vehicles, ambulances, boats, construction vehicles, submarines, robotic vehicles, drones, airplanes, vehicles coupled to trailers (e.g., semi-trailer trucks for hauling cargo), and / or other types of vehicles (e.g., vehicles that are unmanned and / or accommodate one or more passengers). Autonomous vehicles are generally described in terms of levels of automation as defined by the National Highway Traffic Safety Administration (NHTSA), a division of the U.S. Department of Transportation, and the Society of Automotive Engineers (SAE), "Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles" (Standard No. J3016-201806, issued on June 15, 2018, Standard No. J3016-201609, issued on September 30, 2016, and previous and future versions of the same). The vehicle 1100 may be capable of implementing one or more functions consistent with levels 3-5 of autonomous driving. The vehicle 1100 may be capable of implementing one or more functions according to levels 1-5 of autonomous driving. For example, depending on the embodiment, the vehicle 1100 may be capable of driver assistance (Level 1), partial automation (Level 2), conditional automation (Level 3), high automation (Level 4), and / or full automation (Level 5). As used herein, the term "autonomy" may include any and / or all types of autonomy of a vehicle 1100 or other machine, such as fully autonomous, highly autonomous, conditionally autonomous, partially autonomous, assisted autonomy, semi-autonomous, primarily autonomous, or other designations.

[0096] Vehicle 1100 may include components such as a chassis, a body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other vehicle components. Vehicle 1100 may include a propulsion system 1150, such as an internal combustion engine, a hybrid power plant, an all-electric engine, and / or another type of propulsion system. Propulsion system 1150 may be connected to a drivetrain of vehicle 1100, which may include a transmission, to achieve propulsion of vehicle 1100. Propulsion system 1150 may be controlled in response to receiving a signal from throttle / accelerator 1152.

[0097] A steering system 1154, which may include a steering wheel, may be used to steer the vehicle 1100 (e.g., along a desired path or route) when the propulsion system 1150 is operating (e.g., when the vehicle is in motion). The steering system 1154 may receive signals from a steering actuator 1156. For fully automated (Level 5) functionality, a steering wheel may be optional.

[0098] Brake sensor system 1146 may be used to operate vehicle brakes in response to receiving signals from brake actuator 1148 and / or brake sensors.

[0099] May include one or more system on chip (SoC) 1104 ( Figure 11C ) and / or one or more GPUs can provide signals (e.g., representing commands) to one or more components and / or systems of the vehicle 1100. For example, the one or more controllers can send signals to operate the vehicle brakes via one or more brake actuators 1148, to operate the steering system 1154 via one or more steering actuators 1156, and to operate the propulsion system 1150 via one or more throttles / accelerators 1152. The one or more controllers 1136 may include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operational commands (e.g., signals representing commands) to enable autonomous driving and / or assist a human driver in driving the vehicle 1100. The one or more controllers 1136 may include a first controller 1136 for autonomous driving functions, a second controller 1136 for functional safety functions, a third controller 1136 for artificial intelligence functions (e.g., computer vision), a fourth controller 1136 for infotainment functions, a fifth controller 1136 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 1136 may handle two or more of the above functions, two or more controllers 1136 may handle a single function, and / or any combination thereof.

[0100] The one or more controllers 1136 may provide signals for controlling one or more components and / or systems of the vehicle 1100 in response to sensor data (e.g., sensor inputs) received from one or more sensors. The sensor data may be received from, for example and without limitation, a global navigation satellite system ("GNSS") sensor 1158 (e.g., a global positioning system sensor), a RADAR sensor 1160, an ultrasonic sensor 1162, a LIDAR sensor 1164, an inertial measurement unit (IMU) sensor 1166 (e.g., an accelerometer, a gyroscope, a magnetic compass, a magnetometer, etc.), a microphone 1196, a stereo camera 1168, a wide-angle camera 1170 (e.g., a fisheye camera), an infrared camera 1172, a surround camera 1174 (e.g., a 360-degree camera), a long-range and / or mid-range camera 1198, a speed sensor 1144 (e.g., for measuring the velocity of the vehicle 1100), a vibration sensor 1142, a steering sensor 1140, a brake sensor (e.g., as part of a brake sensor system 1146), and / or other sensor types.

[0101] One or more of the controllers 1136 may receive input (e.g., represented by input data) from the instrument cluster 1132 of the vehicle 1100 and provide output (e.g., represented by output data, display data, etc.) via a human machine interface (HMI) display 1134, an audible annunciator, a speaker, and / or via other components of the vehicle 1100. These outputs may include information such as vehicle speed, velocity, time, map data (e.g., Figure 11C The HMI display 1134 may display information regarding the presence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information regarding driving maneuvers that the vehicle has made, is making, or will make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0102] The vehicle 1100 also includes a network interface 1124 that can communicate over one or more networks using one or more wireless antennas 1126 and / or a modem. For example, the network interface 1124 can be capable of communicating over Long Term Evolution ("LTE"), Wideband Code Division Multiple Access ("WCDMA"), Universal Mobile Telecommunications System ("UMTS"), Global System for Mobile Communications ("GSM"), IMT-CDMA Multi-Carrier ("CDMA2000"), and the like. The one or more wireless antennas 1126 can also enable communication between objects in the environment (e.g., vehicles, mobile devices, and the like) using one or more local area networks such as Bluetooth, Bluetooth Low Energy ("LE"), Z-Wave, ZigBee, and the like, and / or one or more low power wide area networks (LPWANs) such as LoRaWAN, SigFox, and the like.

[0103] Figure 11B For use according to some embodiments of the present disclosure Figure 11A 1 . An example of camera positions and fields of view for autonomous vehicle 1100 is shown. The cameras and respective fields of view are an example embodiment and are not intended to be limiting. For example, additional and / or alternative cameras may be included and / or located at different locations on vehicle 1100.

[0104] The camera type used for the camera may include, but is not limited to, a digital camera that may be suitable for use with components and / or systems of the vehicle 1100. The camera may operate at Automotive Safety Integrity Level (ASIL) B and / or at another ASIL. The camera type may have any image capture rate, such as 60 frames per second (fps), 120fps, 240fps, and the like, depending on the embodiment. The camera may be capable of using 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-white-white-white (RCCC) color filter array, a red-white-white-blue (RCCB) color filter array, a red-blue-green-white (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 clear pixel camera such as a camera with an RCCC, RCCB, and / or RBGC color filter array may be used in an effort to improve light sensitivity.

[0105] 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 of the cameras) can simultaneously record and provide image data (e.g., video).

[0106] One or more of the cameras can be mounted in a mounting assembly, such as a custom-designed (three-dimensional ("3D") printed) assembly, to cut out stray light and reflections from within the car (e.g., reflections from the dashboard reflected in the windshield mirror) that might interfere with the camera's ability to capture image data. With respect to the wing mirror mounting assembly, the wing mirror assembly can be custom 3D 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 cabin.

[0107] A camera (e.g., a front-facing camera) having a field of view that includes a portion of the environment in front of the vehicle 1100 can be used for surround vision to help identify the forward path and obstacles, as well as assist in providing information critical to generating an occupancy grid and / or determining a preferred vehicle path with the help of one or more controllers 1136 and / or control SoCs. 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 for ADAS functions and systems, including lane departure warning ("LDW"), autonomous cruise control ("ACC"), and / or other functions such as traffic sign recognition.

[0108] A variety of cameras may be used in the front-facing configuration, including, for example, a monocular camera platform including a complementary metal oxide semiconductor ("CMOS") color imager. Another example may be a wide-angle camera 1170, which may be used to sense objects entering the field of view from the periphery (e.g., pedestrians, traffic at an intersection, or bicycles). Although Figure 11B The figure shows only one wide-angle camera, but there can be any number (including zero) of wide-angle cameras 1170 on the vehicle 1100. In addition, any number of one or more remote cameras 1198 (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. One or more remote cameras 1198 can also be used for object detection and classification and basic object tracking.

[0109] Any number of stereo cameras 1168 may also be included in the front configuration. In at least one embodiment, one or more stereo cameras 1168 may include an integrated control unit including a scalable processing unit that may provide a multi-core microprocessor and programmable logic ("FPGA") with an integrated controller area network ("CAN") or Ethernet interface on a single chip. Such a unit may be used to generate a 3D map of the vehicle's environment, including distance estimates for all points in the image. Alternative stereo cameras 1168 may include a compact stereo vision sensor that may include two camera lenses (one on the left and one on the right) and an image processing chip that may measure the distance from the vehicle to a target object and use the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning features. Other types of stereo cameras 1168 may be used in addition to or alternatively to those described herein.

[0110] Cameras with a field of view that includes portions of the environment to the sides of the vehicle 1100 (e.g., side-view cameras) can be used for surround vision, providing information used to create and update occupancy grids and generate side impact collision warnings. For example, surround cameras 1174 (e.g., Figure 11B Four surround cameras 1174 (shown in FIG) can be placed on the vehicle 1100. The surround cameras 1174 can include a wide-angle camera 1170, a fisheye camera, a 360-degree camera, and / or the like. For example, the four fisheye cameras can be placed on the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle can use three surround cameras 1174 (e.g., left, right, and rear), and can utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround-view camera.

[0111] A camera having a field of view that includes a portion of the environment behind the vehicle 1100 (e.g., a rearview camera) can be used to assist with parking, surround view, rear collision warning, and creating and updating an occupancy grid. A variety of cameras can be used, including but not limited to cameras that are also suitable as front-facing cameras as described herein (e.g., long-range and / or mid-range cameras 1198, stereo cameras 1168, infrared cameras 1172, etc.).

[0112] Figure 11C For use according to some embodiments of the present disclosure Figure 11A 100. It will be understood that this arrangement and other arrangements described herein are set forth merely as examples. 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. Further, many of the elements described herein are functional entities that may be implemented as discrete or distributed components or in conjunction with other components, and in any appropriate combination and location. The various functions described herein as being performed by entities may be implemented by hardware, firmware, and / or software. For example, the various functions may be implemented by a processor executing instructions stored in memory.

[0113] Figure 11C Each of the components, features, and systems of the vehicle 1100 is illustrated as being connected via a bus 1102. The bus 1102 may include a controller area network (CAN) data interface (alternatively referred to herein as a "CAN bus"). The CAN may be a network internal to the vehicle 1100 that assists in controlling various features and functions of the vehicle 1100, such as actuation of brakes, acceleration, braking, steering, windshield wipers, and the like. The CAN bus may be configured to have tens or even hundreds of nodes, each with its own unique identifier (e.g., a CAN ID). The CAN bus may be read to find steering wheel angle, ground speed, engine revolutions per minute (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0114] Although bus 1102 is described here as a CAN bus, this is not intended to be limiting. For example, FlexRay and / or Ethernet may be used in addition to or in lieu of a CAN bus. Furthermore, although bus 1102 is represented by a single line, this is not intended to be limiting. For example, there may be any number of buses 1102, 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 1102 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 1102 may be used for collision avoidance functionality, and a second bus 1102 may be used for drive control. In any example, each bus 1102 may communicate with any component of vehicle 1100, and two or more buses 1102 may communicate with the same component. In some examples, each SoC 1104 , each controller 1136 , and / or each computer within the vehicle may have access to the same input data (e.g., input from sensors of the vehicle 1100 ) and may be connected to a common bus such as a CAN bus.

[0115] The vehicle 1100 may include one or more controllers 1136, such as those described herein. Figure 11A Controller 1136 may be used for a variety of functions. Controller 1136 may be coupled to any of the various other components and systems of vehicle 1100 and may be used for control of vehicle 1100, artificial intelligence of vehicle 1100, infotainment for vehicle 1100, and / or the like.

[0116] The vehicle 1100 may include one or more system-on-chips (SoCs) 1104. The SoC 1104 may include a CPU 1106, a GPU 1108, a processor 1110, a cache 1112, an accelerator 1114, a data store 1116, and / or other components and features not shown. The SoC 1104 may be used to control the vehicle 1100 in a variety of platforms and systems. For example, the one or more SoCs 1104 may be combined with an HD map 1122 in a system (e.g., a system of the vehicle 1100), which may be downloaded from one or more servers (e.g., a server) via a network interface 1124. Figure 11D one or more servers 1178) to obtain map refreshes and / or updates.

[0117] The CPU 1106 may include a CPU cluster or CPU complex (alternatively referred to herein as a "CCPLEX"). The CPU 1106 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU 1106 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 1106 may include four dual-core clusters, each with a dedicated L2 cache (e.g., a 2MB L2 cache). The CPU 1106 (e.g., CCPLEX) may be configured to support simultaneous cluster operations, such that any combination of the CPU 1106 clusters can be active at any given time.

[0118] CPU 1106 may implement power management capabilities including one or more of the following features: each hardware block may be automatically clock gated when idle to conserve dynamic power; each core clock may be gated when the core is not actively executing instructions due to the execution of WFI / WFE instructions; each core may be independently power gated; each core cluster may be independently clock gated when all cores are clock gated or power gated; and / or each core cluster may be independently power gated when all cores are power gated. CPU 1106 may further implement an enhanced algorithm for managing power states, in which allowed power states and expected wakeup times are specified, and hardware / microcode determines the optimal power state to enter for the core, cluster, and CCPLEX. The processing core may support a simplified power state entry sequence in software, with this work being offloaded to the microcode.

[0119] The GPU 1108 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The GPU 1108 may be programmable and efficient for parallel workloads. In some examples, the GPU 1108 may use an enhanced tensor instruction set. The GPU 1108 may include one or more streaming microprocessors, wherein each streaming microprocessor may include an L1 cache (e.g., an L1 cache with at least 96KB of storage capacity), and two or more of the streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB of storage capacity). In some embodiments, the GPU 1108 may include at least eight streaming microprocessors. The GPU 1108 may use a computing application programming interface (API). In addition, the GPU 1108 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0120] In the case of automotive and embedded use, GPU 1108 can be power optimized to achieve optimal performance. For example, GPU 1108 can be manufactured on fin field effect transistors (FinFETs). However, this is not intended to be limiting, and GPU 1108 can be manufactured using other semiconductor manufacturing processes. Each streaming microprocessor can incorporate 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, L0 instruction cache, warp scheduler, dispatch unit and / or 64KB register file. In addition, the streaming microprocessor may include independent parallel integer and floating point data paths to provide efficient execution of workloads using a mix of computation and addressing calculations. The streaming microprocessor may include independent thread scheduling capabilities to allow for finer-grained synchronization and cooperation between parallel threads. The streaming microprocessor may include a combined L1 data cache and shared memory unit to improve performance while simplifying programming.

[0121] GPU 1108 can include high bandwidth memory (HBM) and / or a 16GB HBM2 memory subsystem that provides a peak memory bandwidth of approximately 900 GB / s in some examples. In some examples, synchronous graphics random access memory (SGRAM), such as fifth generation graphics double data rate synchronous random access memory (GDDR5), can be used in addition to or in lieu of HBM memory.

[0122] The GPU 1108 may include unified memory technology that includes access counters to allow memory pages to be more accurately migrated to the processor that accesses them most frequently, thereby improving the efficiency of memory ranges shared between processors. In some examples, address translation services (ATS) support may be used to allow the GPU 1108 to directly access the CPU 1106 page tables. In such an example, when the GPU 1108 memory management unit (MMU) experiences a miss, an address translation request may be transmitted to the CPU 1106. In response, the CPU 1106 may look up the virtual-to-physical mapping for the address in its page table and transmit the translation back to the GPU 1108. In this way, unified memory technology may allow a single unified virtual address space to be used for memory of both the CPU 1106 and the GPU 1108, thereby simplifying GPU 1108 programming and porting applications to the GPU 1108.

[0123] Additionally, GPU 1108 may include access counters that can track how often GPU 1108 accesses the memory of other processors. Access counters can help ensure that memory pages are moved to the physical memory of the processor that accesses them most frequently.

[0124] SoC 1104 may include any number of caches 1112, including those described herein. For example, cache 1112 may include an L3 cache available to both CPU 1106 and GPU 1108 (e.g., connected to both CPU 1106 and GPU 1108). Cache 1112 may include a write-back cache that can track the state of lines, 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, although smaller cache sizes may also be used.

[0125] The SoC 1104 may include an arithmetic logic unit (ALU) that may be utilized in performing any of a variety of tasks or operations associated with the vehicle 1100, such as processing a DNN. Furthermore, the SoC 1104 may include a floating point unit (FPU) (or other math coprocessor or digital coprocessor type) for performing mathematical operations within the system. For example, the SoC 1104 may include one or more FPUs integrated as execution units within the CPU 1106 and / or GPU 1108.

[0126] SoC 1104 may include one or more accelerators 1114 (e.g., hardware accelerators, software accelerators, or a combination thereof). For example, SoC 1104 may include a hardware accelerator cluster, which may include optimized hardware accelerators and / or large on-chip memory. The large on-chip memory (e.g., 4MB SRAM) may enable the hardware accelerator cluster to accelerate neural networks and other calculations. The hardware accelerator cluster may be used to supplement GPU 1108 and offload some tasks of GPU 1108 (e.g., freeing up more cycles of GPU 1108 for performing other tasks). As an example, accelerator 1114 may be used for targeted workloads (e.g., perception, convolutional neural networks (CNNs), etc.) that are sufficiently stable to be easily controlled for acceleration. When used herein, the term "CNN" may include all types of CNNs, including region-based or regional convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0127] Accelerator 1114 (e.g., a hardware accelerator cluster) may include a deep learning accelerator (DLA). 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 reasoning. TPUs may be accelerators configured to perform image processing functions (e.g., for CNN, RCNN, etc.) and optimized for performing image processing functions. DLA may be further optimized for a specific set of neural network types and floating-point operations and reasoning. The design of DLA may provide higher performance per millimeter than a general-purpose GPU and far exceed the performance of a CPU. TPUs may 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.

[0128] DLA can quickly and efficiently execute neural networks, particularly CNNs, on processed or unprocessed data for any of a wide variety of functions, 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 and detection using data from microphones; CNNs for facial recognition and vehicle owner identification using data from camera sensors; and / or CNNs for safety and / or security-related events.

[0129] The DLA can perform any function of the GPU 1108, and by using an inference accelerator, for example, the designer can target any function to either the DLA or the GPU 1108. For example, the designer can focus the processing of CNNs and floating-point operations on the DLA and leave other functions to the GPU 1108 and / or other accelerators 1114.

[0130] The accelerator 1114 (e.g., a hardware accelerator cluster) may include a programmable vision accelerator (PVA), which may be alternatively referred to herein as a computer vision accelerator. The PVA may 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. The PVA may provide a balance between performance and flexibility. For example, each PVA may include, for example and without limitation, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0131] The RISC core can interact with an image sensor (e.g., an image sensor of any camera described herein), an image signal processor, and / or the like. Each of these RISC cores can include any amount of memory. Depending on the embodiment, the RISC core can use any of a number of protocols. In some examples, the RISC core can execute a real-time operating system (RTOS). The RISC core can be implemented using one or more integrated circuit devices, application specific integrated circuits (ASICs), and / or memory devices. For example, the RISC core can include an instruction cache and / or tightly coupled RAM.

[0132] The DMA can enable components of the PVA to access system memory independently of the CPU 1106. The DMA can support any number of features used to provide optimizations for the PVA, including but not limited to support for multi-dimensional addressing and / or circular addressing. In some examples, the DMA can support addressing in up to six or more dimensions, which can include block width, block height, block depth, horizontal block stride, vertical block stride, and / or depth stride.

[0133] A vector processor can be a programmable processor that can be designed to efficiently and flexibly execute programming for computer vision algorithms and provide signal processing capabilities. In some examples, the PVA can include a PVA core and two vector processing subsystem partitions. The PVA core can include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem can operate as the main processing engine of the PVA and can include a vector processing unit (VPU), an instruction cache, and / or a vector memory (e.g., VMEM). The VPU core can include a digital signal processor, such as, for example, a single instruction multiple data (SIMD), a very long instruction word (VLIW) digital signal processor. The combination of SIMD and VLIW can enhance throughput and speed.

[0134] Each of the vector processors can include an instruction cache and can be coupled to dedicated memory. As a result, in some examples, each of the vector processors can be configured to execute independently of the other vector processors. In other examples, the vector processors included in a particular PVA can be configured to employ data parallelism. For example, in some embodiments, multiple vector processors included in a single PVA can execute the same computer vision algorithm, but on different regions of an image. In other examples, the vector processors included in a particular PVA can execute different computer vision algorithms simultaneously on the same image, or even execute different algorithms on sequential images or portions of images. Among other things, any number of PVAs can be included in a hardware accelerator cluster, and any number of vector processors can be included in each of these PVAs. In addition, the PVAs can include additional error correction code (ECC) memory to enhance overall system security.

[0135] The accelerator 1114 (e.g., a hardware accelerator cluster) may include an on-chip computer vision network and SRAM to provide high bandwidth, low latency SRAM for the accelerator 1114. In some examples, the on-chip memory may include at least 4MB of SRAM consisting of, for example and without limitation, eight field-configurable memory blocks that can be accessed by both the PVA and the 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 may be used. The PVA and DLA may access the memory via a backbone that provides high-speed memory access to the PVA and DLA. The backbone may include an on-chip computer vision network that interconnects the PVA and DLA to the memory (e.g., using APB).

[0136] The on-chip computer vision network can include an interface that ensures that both the PVA and DLA provide ready and valid signals before transmitting any control signals / addresses / data. Such an interface can provide separate phases and separate channels for transmitting control signals / addresses / data, as well as burst-based communication for continuous data transmission. This type of interface can comply with ISO 26262 or IEC 61508 standards, but other standards and protocols can also be used.

[0137] In some examples, SoC 1104 may include a real-time ray tracing hardware accelerator such as that described in U.S. patent application Ser. No. 16 / 101,232 filed on Aug. 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the position and extent of objects (e.g., within a world model) to generate real-time visualization simulations for RADAR signal interpretation, for sound propagation synthesis and / or analysis, for SONAR system simulations, for general wave propagation simulations, for comparison with LIDAR data for positioning and / or other functions, 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.

[0138] The accelerator 1114 (e.g., a hardware accelerator cluster) has a wide range of uses in autonomous driving. The PVA can be a programmable vision accelerator that can be used in key processing stages in ADAS and autonomous vehicles. The capabilities of the PVA are a good match for algorithmic domains that require predictable processing, low power, and low latency. In other words, the PVA performs well on semi-intensive or intensive rule computations, and even on small data sets that require predictable runtimes with low latency and low power. Therefore, in the context of platforms for autonomous vehicles, the PVA is designed to run classic computer vision algorithms because they are efficient at object detection and integer math operations.

[0139] For example, according to one embodiment of the technology, PVA is used to perform computer stereo vision. In some examples, a semi-global matching-based algorithm can be used, but this is not intended to be limiting. Many applications for Level 3-5 autonomous driving require on-the-fly motion estimation / stereo matching (e.g., structure from motion, pedestrian recognition, lane detection, etc.). PVA can perform computer stereo vision functions on input from two monocular cameras.

[0140] In some examples, PVA can be used to perform dense optical flow, by processing raw RADAR data (e.g., using a 4D Fast Fourier Transform) to provide processed RADAR. In other examples, PVA is used for time-of-flight depth processing, by processing raw time-of-flight data to provide processed time-of-flight data.

[0141] DLA can be used to run any type of network to enhance control and driving safety, including, for example, a neural network that outputs a confidence measure for each object detection. Such confidence values can be interpreted as probabilities, or as providing a relative "weight" of each detection compared to other detections. This confidence value enables the system to make further decisions about which detections should be considered true positive detections rather than false positive detections. For example, the system can set a threshold for confidence and only consider detections that exceed the threshold as true positive detections. In an automatic emergency braking (AEB) system, a false positive detection would cause the vehicle to automatically perform emergency braking, which is obviously undesirable. Therefore, only the most confident detections should be considered triggers for AEB. DLA can run a neural network for regressing confidence values. The neural network may take as its input at least some subset of parameters, such as bounding box dimensions, a ground plane estimate obtained (e.g., from another subsystem), inertial measurement unit (IMU) sensor 1166 output related to the orientation and distance of the vehicle 1100, 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LIDAR sensor 1164 or RADAR sensor 1160), etc.

[0142] SoC 1104 may include one or more data stores 1116 (e.g., memory). Data store 1116 may be on-chip memory of SoC 1104 that may store neural networks to be executed on the GPU and / or DLA. In some examples, for redundancy and safety, data store 1116 may be large enough to store multiple instances of the neural network. Data store 1112 may include an L2 or L3 cache 1112. References to data store 1116 may include references to memory associated with the PVA, DLA, and / or other accelerators 1114 as described herein.

[0143] SoC 1104 may include one or more processors 1110 (e.g., embedded processors). Processors 1110 may include a boot and power management processor, which may be a dedicated processor and subsystem for handling boot power and management functions and related safety implementations. The boot and power management processor may be part of the SoC 1104 boot sequence and may provide runtime power management services. The boot power and management processor may provide clock and voltage programming, auxiliary system low-power state transitions, SoC 1104 thermal and temperature sensor management, and / or SoC 1104 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and the SoC 1104 may use the ring oscillator to detect the temperature of the CPU 1106, GPU 1108, and / or accelerator 1114. If the temperature is determined to exceed a threshold, the boot and power management processor may enter a temperature fault routine and place the SoC 1104 in a lower power state and / or place the vehicle 1100 in a driver safety parking mode (e.g., to safely park the vehicle 1100).

[0144] The processor 1110 may also 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 through multiple interfaces and a wide range of flexible audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core having a digital signal processor with dedicated RAM.

[0145] The processor 1110 may also include an always-on processor engine that can provide the necessary hardware features to support low-power sensor management and wake-up use cases. The always-on processor engine may include a processor core, tightly coupled RAM, supporting peripherals (such as timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0146] Processor 1110 may also include a safety cluster engine, which includes a dedicated processor subsystem that handles safety management of automotive applications. The safety cluster engine may include two or more processor cores, tightly coupled RAM, supporting peripherals (such as timers, interrupt controllers, etc.), and / or routing logic. In safety mode, the two or more cores can operate in lockstep mode and act as a single core with comparison logic to detect any differences between their operations.

[0147] Processor 1110 may also include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0148] Processor 1110 may also 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.

[0149] The processor 1110 may include a video image compositer, which may be a processing block (e.g., implemented on a microprocessor) that implements the video post-processing functions required by the video playback application to produce the final image for the player window. The video image compositer may perform lens distortion correction for the wide-angle camera 1170, the surround camera 1174, and / or for the in-cab monitoring camera sensor. The in-cab monitoring camera sensor is preferably monitored by a neural network running on another instance of the advanced SoC, configured to recognize in-cab events and respond accordingly. The in-cab system may perform lip reading to activate mobile phone service and place calls, dictate emails, change vehicle destinations, activate or change the vehicle's infotainment system and settings, or provide voice-activated web surfing. Certain functions are only available to the driver when the vehicle is operating in autonomous mode and are disabled in other circumstances.

[0150] The video image compositer can include enhanced temporal noise reduction for both spatial and temporal noise reduction. For example, in the presence of motion in the video, the noise reduction appropriately weights spatial information and downweights information provided by neighboring frames. In the case where an image or portion of an image does not include motion, the temporal noise reduction performed by the video image compositer can use information from previous images to reduce noise in the current image.

[0151] The video image compositor can also be configured to perform stereo rectification on the input stereo footage frames. The video image compositor can further be used for user interface composition when the operating system desktop is in use and the GPU 1108 does not need to continuously render new surfaces. Even when the GPU 1108 is powered on and active for 3D rendering, the video image compositor can be used to offload the GPU 1108 to improve performance and responsiveness.

[0152] The SoC 1104 may also include a Mobile Industry Processor Interface (MIPI) camera serial interface, a high-speed interface for receiving video and input from a camera, and / or a video input block that may be used for camera and related pixel input functions. The SoC 1104 may also include an input / output controller that may be controlled by software and may be used to receive I / O signals that are not assigned to a specific role.

[0153] The SoC 1104 may also include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management, and / or other devices. The SoC 1104 may be used to process data from cameras (connected via Gigabit multimedia serial links and Ethernet), sensors (e.g., LIDAR sensor 1164, RADAR sensor 1160, etc., which may be connected via Ethernet), data from the bus 1102 (e.g., vehicle 1100 speed, steering wheel position, etc.), and data from the GNSS sensor 1158 (connected via Ethernet or a CAN bus). The SoC 1104 may also include dedicated high-performance mass storage controllers, which may include their own DMA engines and which may be used to free the CPU 1106 from routine data management tasks.

[0154] SoC 1104 can be an end-to-end platform with a flexible architecture that spans levels 3-5 of automation, providing a comprehensive functional safety architecture that leverages and efficiently uses computer vision and ADAS technologies for diversity and redundancy, along with deep learning tools to provide a platform for a flexible and reliable driving software stack. SoC 1104 can be faster, more reliable, and even more energy- and space-efficient than conventional systems. For example, when combined with CPU 1106, GPU 1108, and data storage 1116, accelerator 1114 can provide a fast and efficient platform for Level 3-5 autonomous vehicles.

[0155] This technology therefore provides capabilities and functionality that cannot be achieved with conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using high-level programming languages such as the C programming language to execute a wide variety of processing algorithms across a wide variety of visual data. However, CPUs often fail to meet the performance requirements of many computer vision applications, such as those related to execution time and power consumption. In particular, many CPUs are unable to execute complex object detection algorithms in real time, a requirement for in-vehicle ADAS applications and practical Level 3-5 autonomous vehicles.

[0156] In contrast to conventional systems, by providing a CPU complex, a GPU complex, and a cluster of hardware accelerators, the technology described herein allows 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 1120) can include text and word recognition, allowing the supercomputer to read and understand traffic signs, including signs for which a neural network has not been specifically trained. The DLA can also include a neural network that can recognize, interpret, and provide semantic understanding of the signs, and pass that semantic understanding to a path planning module running on the CPU complex.

[0157] As another example, as required for Level 3, 4, or 5 driving, multiple neural networks can run simultaneously. For example, a warning sign consisting of "Caution: Flashing lights indicate icing conditions" along with a light can be interpreted by several neural networks, either independently or collectively. The sign itself can be identified as a traffic sign by a first neural network deployed (e.g., a trained neural network), and the text "Flashing lights indicate icing conditions" can be interpreted by a second neural network deployed, which informs the vehicle's path planning software (preferably executing on a CPU complex) that icing conditions exist when the flashing lights are detected. The flashing lights can be identified by operating a third neural network deployed over multiple frames, which informs the vehicle's path planning software of the presence (or absence) of the flashing lights. All three neural networks can run simultaneously, for example, within the DLA and / or on GPU 1108.

[0158] In some examples, a CNN for facial recognition and owner recognition can use data from a camera sensor to identify the presence of an authorized driver and / or owner of the vehicle 1100. The always-on sensor processing engine can be used to unlock the vehicle and turn on the lights when the owner approaches the driver's door, and in security mode, disable the vehicle when the owner leaves the vehicle. In this way, the SoC 1104 provides security against theft and / or carjacking.

[0159] In another example, a CNN for emergency vehicle detection and identification can use data from microphone 1196 to detect and identify emergency vehicle sirens. In contrast to conventional systems that use general classifiers to detect sirens and manually extract features, SoC 1104 uses CNNs to classify environmental and urban sounds and to classify visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative closing speed of emergency vehicles (for example, by using the Doppler effect). The CNN can also be trained to identify emergency vehicles specific to the local area in which the vehicle is operating, as identified by the GNSS sensor 1158. Thus, for example, when operating in Europe, the CNN will seek to detect European sirens, and when in the United States, the CNN will seek to identify only North American sirens. Once an emergency vehicle is detected, with the assistance of the ultrasonic sensor 1162, the control program can be used to execute emergency vehicle safety routines, slowing the vehicle, pulling to the side of the road, stopping the vehicle, and / or idling the vehicle until the emergency vehicle passes.

[0160] The vehicle may include a CPU 1118 (e.g., a discrete CPU or dCPU) that may be coupled to the SoC 1104 via a high-speed interconnect (e.g., PCIe). The CPU 1118 may include, for example, an X86 processor. The CPU 1118 may be used to perform any of a variety of functions, including, for example, arbitrating potentially inconsistent results between ADAS sensors and the SoC 1104, and / or monitoring the status and health of the controller 1136 and / or the infotainment SoC 1130.

[0161] The vehicle 1100 may include a GPU 1120 (e.g., a discrete GPU or dGPU) that may be coupled to the SoC 1104 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU 1120 may provide additional artificial intelligence functionality, for example, by executing redundant and / or different neural networks, and may be used to train and / or update the neural network based at least in part on input from sensors of the vehicle 1100 (e.g., sensor data).

[0162] The vehicle 1100 may also include a network interface 1124, which may include one or more wireless antennas 1126 (e.g., one or more wireless antennas for different communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 1124 can be used to enable wireless connections to the cloud (e.g., to a server 1178 and / or other network devices), to other vehicles, and / or to computing devices (e.g., a passenger's client device) via the Internet. To communicate 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 through the Internet). The direct link can be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link can provide the vehicle 1100 with information about vehicles approaching the vehicle 1100 (e.g., vehicles in front of, to the sides of, and / or behind the vehicle 1100). This functionality can be part of the cooperative adaptive cruise control functionality of the vehicle 1100.

[0163] The network interface 1124 may include a SoC that provides modulation and demodulation functionality and enables the controller 1136 to communicate over a wireless network. The network interface 1124 may include an RF front-end for up-conversion from baseband to RF and down-conversion from RF to baseband. The frequency conversion may be performed by well-known processes and / or may be performed using a super-heterodyne process. In some examples, the RF front-end functionality may be provided by a separate chip. The network interface may include wireless functionality for communicating over LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0164] The vehicle 1100 may also include data storage 1128, which may include off-chip storage (e.g., outside the SoC 1104). The data storage 1128 may include one or more storage elements, including RAM, SRAM, DRAM, VRAM, flash memory, a hard disk, and / or other components and / or devices that can store at least one bit of data.

[0165] The vehicle 1100 may also include a GNSS sensor 1158. The GNSS sensor 1158 (e.g., GPS, assisted GPS sensor, differential GPS (DGPS) sensor, etc.) is used to assist with mapping, perception, occupancy grid generation, and / or path planning functions. Any number of GNSS sensors 1158 may be used, including, for example and without limitation, GPS using a USB connector with an Ethernet to serial (RS-232) bridge.

[0166] The vehicle 1100 may also include a RADAR sensor 1160. The RADAR sensor 1160 can be used by the vehicle 1100 for remote vehicle detection even in darkness and / or in adverse weather conditions. The RADAR functional safety level can be ASIL B. The RADAR sensor 1160 can use CAN and / or bus 1102 (e.g., to transmit data generated by the RADAR sensor 1160) for control and access to object tracking data, and in some examples access Ethernet to access raw data. A variety of RADAR sensor types can be used. For example and without limitation, the RADAR sensor 1160 can be suitable for front, rear, and side RADAR use. In some examples, a pulsed Doppler RADAR sensor is used.

[0167] The RADAR sensor 1160 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, and so on. In some examples, 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 a range of 250m) achieved through two or more independent scans. The RADAR sensor 1160 can help distinguish between static objects and moving objects and can be used by the ADAS system for emergency braking assistance and forward collision warning. The long-range RADAR sensor may include a single-station multimode RADAR with multiple (e.g., six or more) fixed RADAR antennas and high-speed CAN and FlexRay interfaces. In the example with six antennas, the central four antennas can create a focused beam pattern that is designed to record the surroundings of the vehicle 1100 at a higher rate with minimal traffic interference from adjacent lanes. The other two antennas can expand the field of view, making it possible to quickly detect vehicles entering or leaving the lane of the vehicle 1100.

[0168] As an example, a medium-range RADAR system may include a range of up to 1160m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 1150 degrees (rear). A short-range RADAR system may include, but is not limited to, a RADAR sensor designed to be mounted on both ends of the rear bumper. When mounted on both ends of the rear bumper, such a RADAR sensor system can create two beams that continuously monitor the blind spots behind and beside the vehicle.

[0169] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0170] Vehicle 1100 may also include ultrasonic sensors 1162. Ultrasonic sensors 1162, which may be located on the front, rear, and / or sides of vehicle 1100, may be used for parking assistance and / or for creating and updating an occupancy grid. A variety of ultrasonic sensors 1162 may be used, and different ultrasonic sensors 1162 may have different detection ranges (e.g., 2.5 m, 4 m). Ultrasonic sensors 1162 may operate at functional safety level ASIL B.

[0171] Vehicle 1100 may include a LIDAR sensor 1164. LIDAR sensor 1164 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. LIDAR sensor 1164 may be ASIL B functional safety level. In some examples, vehicle 1100 may include multiple LIDAR sensors 1164 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0172] In some examples, LIDAR sensor 1164 may be capable of providing a list of objects and their distances for a 360-degree field of view. Commercially available LIDAR sensors 1164 may have, for example, an advertised range of approximately 1100 meters, an accuracy of 2-3 cm, and support for 1100 Mbps Ethernet connections. In some examples, one or more non-obtrusive LIDAR sensors 1164 may be used. In such examples, LIDAR sensor 1164 may be implemented as a small device that can be embedded in the front, back, sides, and / or corners of vehicle 1100. In such examples, LIDAR sensor 1164 may provide a field of view of up to 120 degrees horizontally and 35 degrees vertically, with a range of 200 meters, even for low-reflectivity objects. Front-mounted LIDAR sensor 1164 may be configured for a horizontal field of view between 45 and 135 degrees.

[0173] In some examples, LIDAR technologies such as 3D flash LIDAR may also be used. 3D flash LIDAR uses flashes of laser as an emission source to illuminate the vehicle's surroundings up to about 200 m. The flash LIDAR unit includes a receiver that records the laser pulse transmission time and reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LIDAR can allow a highly accurate and distortion-free image of the surrounding environment to be generated with each laser flash. In some examples, four flash LIDAR sensors can be deployed, one on each side of the vehicle 1100. Available 3D flash LIDAR systems include solid-state 3D staring array LIDAR cameras (e.g., non-scanning LIDAR devices) with no moving parts other than a fan. The flash LIDAR device can use 5 nanosecond Class I (eye-safe) laser pulses per frame and can capture the reflected laser light in the form of a 3D range point cloud and co-registered intensity data. By using flash LIDAR, and because flash LIDAR is a solid-state device with no moving parts, the LIDAR sensor 1164 may be less susceptible to motion blur, vibration, and / or shock.

[0174] The vehicle may also include an IMU sensor 1166. In some examples, the IMU sensor 1166 may be located at the center of the rear axle of the vehicle 1100. The IMU sensor 1166 may include, for example and without limitation, 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 1166 may include an accelerometer and a gyroscope, while in a nine-axis application, the IMU sensor 1166 may include an accelerometer, a gyroscope, and a magnetometer.

[0175] In some embodiments, the IMU sensor 1166 can be implemented as a miniature, high-performance GPS-assisted inertial navigation system (GPS / INS) that combines micro-electromechanical systems (MEMS) inertial sensors, a high-sensitivity GPS receiver, and advanced Kalman filtering algorithms to provide estimates of position, velocity, and attitude. Thus, in some examples, the IMU sensor 1166 can enable the vehicle 1100 to estimate heading without requiring input from a magnetic sensor by directly observing and correlating velocity changes from the GPS to the IMU sensor 1166. In some examples, the IMU sensor 1166 and the GNSS sensor 1158 can be combined into a single integrated unit.

[0176] The vehicle may include microphones 1196 positioned in and / or around the vehicle 1100. The microphones 1196 may be used for, among other things, emergency vehicle detection and identification.

[0177] The vehicle may also include any number of camera types, including stereo cameras 1168, wide angle cameras 1170, infrared cameras 1172, surround cameras 1174, long and / or medium range cameras 1198, and / or other camera types. These cameras may be used to capture image data around the entire periphery of the vehicle 1100. The type of camera used depends on the embodiment and the requirements of the vehicle 1100, and any combination of camera types may be used to provide the necessary coverage around the vehicle 1100. Additionally, 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 not limitation, the cameras may support Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the cameras described herein may include a GMSL and / or Gigabit Ethernet network. Figure 11A and Figure 11B Described in more detail.

[0178] Vehicle 1100 may also include a vibration sensor 1142. Vibration sensor 1142 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 1142 are used, the difference between the vibrations can be used to determine friction or slippage of the road surface (e.g., when there is a vibration difference between a powered drive shaft and a freely rotating shaft).

[0179] The vehicle 1100 may include an ADAS system 1138. In some examples, the ADAS system 1138 may include a SoC. The ADAS system 1138 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.

[0180] The ACC system can utilize RADAR sensors 1160, LIDAR sensors 1164, and / or cameras. The ACC system can include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to the vehicle immediately in front of vehicle 1100 and automatically adjusts the vehicle speed to maintain a safe distance from the vehicle in front. Lateral ACC maintains distance and, when necessary, recommends that vehicle 1100 change lanes. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0181] CACC uses information from other vehicles, which can be received from other vehicles indirectly via a wireless link or through a network connection (e.g., through the Internet) via the network interface 1124 and / or the wireless antenna 1126. A direct link can be provided by a vehicle-to-vehicle (V2V) communication link, while an indirect link can be an infrastructure-to-vehicle (I2V) communication link. Typically, the V2V communication concept provides information about the vehicle immediately ahead (e.g., the vehicle immediately ahead of the vehicle 1100 and in the same lane as it), while the I2V communication concept provides information about traffic further ahead. The CACC system can include either or both of the I2V and V2V information sources. Given information about the vehicle ahead of the vehicle 1100, CACC can be more reliable, and it has the potential to improve the smoothness of traffic flow and reduce road congestion.

[0182] The FCW system is designed to alert the driver to hazards so that the driver can take corrective action. The FCW system uses a front-facing camera and / or RADAR sensor 1160 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 component. The FCW system can provide warnings in the form of, for example, audible, visual warnings, vibrations, and / or rapid brake pulses.

[0183] The AEB 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 specified time or distance parameters. The AEB system can use a front-facing camera and / or RADAR sensor 1160 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 the collision, and if the driver does not take corrective action, the AEB system can automatically apply the brakes in an effort to prevent or at least mitigate the effects of the predicted collision. The AEB system can include technologies such as dynamic brake support and / or collision approach braking.

[0184] The LDW system provides visual, audible, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle 1100 crosses a lane marking. When the driver indicates an intention to leave the lane by activating a turn signal, the LDW system is deactivated. The LDW system may utilize a front-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration components.

[0185] The LKA system is a variation of the LDW system. If the vehicle 1100 begins to leave its lane, the LKA system provides steering input or braking to correct the vehicle 1100.

[0186] The BSW system detects and warns the driver of vehicles in the car's blind spot. The BSW system can provide visual, audible, and / or tactile alerts to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses a turn signal. The BSW system can use a rear-facing camera and / or RADAR sensor 1160 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration component.

[0187] The RCTW system can provide visual, audible, and / or tactile notifications when an object is detected outside the range of the rear-mounted camera while the vehicle 1100 is in reverse. Some RCTW systems include automatic emergency braking (AEB) to ensure that the vehicle brakes are applied to avoid a collision. The RCTW system can use one or more rear-mounted RADAR sensors 1160 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration component.

[0188] Conventional ADAS systems can be prone to false positive results, which can be annoying and distracting to the driver, but are typically not catastrophic because the ADAS system alerts the driver and allows the driver to decide whether a safety condition actually exists and take action accordingly. However, in the autonomous vehicle 1100, in the event of conflicting results, the vehicle 1100 itself must decide whether to heed the results from the primary computer or the auxiliary computer (e.g., the first controller 1136 or the second controller 1136). For example, in some embodiments, the ADAS system 1138 can be a backup and / or auxiliary computer for providing perception information to the backup computer rationality module. The backup computer rationality monitor can run redundant and diverse software on hardware components to detect failures in perception and dynamic driving tasks. The output from the ADAS system 1138 can be provided to the supervisory MCU. If the outputs from the primary and auxiliary computers conflict, the supervisory MCU must determine how to reconcile the conflict to ensure safe operation.

[0189] In some examples, the primary computer can be configured to provide a confidence score to the supervisory MCU, indicating the primary computer's confidence in the selected result. If the confidence score exceeds a threshold, the supervisory MCU can follow the primary computer's direction, regardless of whether the secondary computer provides conflicting or inconsistent results. In the event that the confidence score does not meet the threshold and the primary and secondary computers indicate different results (e.g., a conflict), the supervisory MCU can arbitrate between these computers to determine the appropriate result.

[0190] The supervisory MCU can be configured to run a neural network that is trained and configured to determine conditions under which the secondary computer provides a false alarm based, at least in part, on outputs from the primary computer and the secondary computer. Thus, the neural network in the supervisory MCU can learn when the output of the secondary computer can be trusted and when it cannot. For example, when the secondary computer is a RADAR-based FCW system, the neural network in the supervisory MCU can learn when the FCW system is identifying a metal object that is not actually a danger, such as a drain grate or manhole cover, which triggers an alarm. Similarly, when the secondary computer is a camera-based LDW system, the neural network in the supervisory MCU can learn to disregard 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 can include at least one of a DLA or a GPU suitable for running the neural network with associated memory. In a preferred embodiment, the supervisory MCU can include and / or be included as a component of the SoC 1104.

[0191] In other examples, the ADAS system 1138 may include an auxiliary computer that uses traditional computer vision rules to perform ADAS functions. In this way, the auxiliary computer can use classic computer vision rules (if-then), and the presence of a neural network in the supervisory MCU can improve reliability, safety, and performance. For example, diverse implementations and intentional non-identity make the entire system more fault-tolerant, especially with respect to failures caused by software (or software-hardware interface) functions. For example, if there is a software vulnerability or bug in the software running on the main computer and the non-identical software code running on the auxiliary computer provides the same overall result, the supervisory MCU can be more confident that the overall result is correct and that the vulnerability in the software or hardware on the main computer did not cause a substantial error.

[0192] In some examples, the output of the ADAS system 1138 can be fed into the primary computer's perception block and / or the primary computer's dynamic driving task block. For example, if the ADAS system 1138 indicates a forward collision warning due to an object immediately ahead, the perception block can use this information when identifying the object. In other examples, the secondary computer can have its own neural network that is trained and thus reduces the risk of false positives as described herein.

[0193] The vehicle 1100 may also include an infotainment SoC 1130 (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 separate components. The infotainment SoC 1130 may include a combination of hardware and software that can be used to provide audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming, etc.), phone (e.g., hands-free calling), network connectivity (e.g., LTE, WiFi, etc.), and / or information services (e.g., a navigation system, rear parking assistance, radio data system, vehicle-related information such as fuel level, total distance covered, brake fuel level, oil level, door open / closed, air filter information, etc.) to the vehicle 1100. For example, the infotainment SoC 1130 may include a radio, a disc player, a navigation system, a video player, USB and Bluetooth connectivity, an onboard computer, in-car entertainment, WiFi, steering wheel audio controls, hands-free voice controls, a head-up display (HUD), an HMI display 1134, a telematics device, a control panel (e.g., for controlling and / or interacting with various components, features, and / or systems), and / or other components. The infotainment SoC 1130 may further be used to provide information (e.g., visual and / or auditory) to a user of the vehicle, such as information from an ADAS system 1138, 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.

[0194] The infotainment SoC 1130 may include GPU functionality. The infotainment SoC 1130 may communicate with other devices, systems, and / or components of the vehicle 1100 via a bus 1102 (e.g., a CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 1130 may be coupled to a supervisory MCU so that in the event of a failure of a primary controller 1136 (e.g., a primary and / or backup computer of the vehicle 1100), the infotainment system's GPU may perform some self-driving functions. In such an example, the infotainment SoC 1130 may place the vehicle 1100 in a driver-safe parking mode as described herein.

[0195] The vehicle 1100 may also include an instrument cluster 1132 (e.g., a digital instrument panel, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 1132 may include a controller and / or a supercomputer (e.g., a separate controller or a supercomputer). The instrument cluster 1132 may include a set of instruments, such as a speedometer, fuel level, oil pressure, a tachometer, an odometer, a turn indicator, a gear position indicator, a seat belt warning light, a parking brake warning light, an engine check light, airbag (SRS) system information, lighting controls, safety system controls, navigation information, etc. In some examples, information can be displayed and / or shared between the infotainment SoC 1130 and the instrument cluster 1132. In other words, the instrument cluster 1132 may be included as part of the infotainment SoC 1130, or vice versa.

[0196] Figure 11D For cloud-based servers and Figure 11A 11. System diagram of communication between an autonomous vehicle 1100 and an example system 1176. System 1176 may include a server 1178, a network 1190, and a vehicle including vehicle 1100. Server 1178 may include multiple GPUs 1184(A)-1284(H) (collectively referred to herein as GPUs 1184), PCIe switches 1182(A)-1182(H) (collectively referred to herein as PCIe switches 1182), and / or CPUs 1180(A)-1180(B) (collectively referred to herein as CPUs 1180). GPUs 1184, CPUs 1180, and PCIe switches may be interconnected with a high-speed interconnect such as, for example and without limitation, NVLink interface 1188 developed by NVIDIA and / or PCIe connection 1186. In some examples, GPUs 1184 are connected via NVLink and / or NVSwitch SoCs, and GPUs 1184 and PCIe switches 1182 are connected via PCIe interconnects. Although eight GPUs 1184, two CPUs 1180, and two PCIe switches are shown, this is not intended to be limiting. Depending on the embodiment, each of the servers 1178 may include any number of GPUs 1184, CPUs 1180, and / or PCIe switches. For example, each of the servers 1178 may include eight, sixteen, thirty-two, and / or more GPUs 1184.

[0197] Server 1178 can receive image data from a vehicle via network 1190, the image data representing images showing unexpected or changed road conditions, such as recently begun road construction. Server 1178 can transmit neural network 1192, updated neural network 1192, and / or map information 1194, including information about traffic and road conditions, via network 1190 and to the vehicle. Updates to map information 1194 can include updates to HD map 1122, such as information about construction sites, potholes, curves, flooding, or other obstacles. In some examples, neural network 1192, updated neural network 1192, and / or map information 1194 can be generated from new training and / or data received from any number of vehicles in the environment and / or based on experience from training performed at a data center (e.g., using server 1178 and / or other servers).

[0198] Server 1178 can be used to train a machine learning model (e.g., a neural network) based on training data. The training data can be generated by the vehicle and / or can be generated in simulation (e.g., using a game engine). In some examples, the training data is labeled (e.g., in cases 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., in cases 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 the following categories: 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 variants or combinations thereof. Once the machine learning model is trained, the machine learning model can be used by the vehicle (e.g., transmitted to the vehicle via network 1190), and / or the machine learning model can be used by server 1178 to remotely monitor the vehicle.

[0199] In some examples, server 1178 can receive data from the vehicle and apply the data to the latest real-time neural network for real-time intelligent reasoning. Server 1178 can include a deep learning supercomputer and / or a dedicated AI computer powered by GPU 1184, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 1178 can include the deep learning infrastructure of a data center using only CPU power.

[0200] The deep learning infrastructure of server 1178 may be capable of rapid real-time inference and may use this capability to assess and verify the health of the processors, software, and / or associated hardware in vehicle 1100. For example, the deep learning infrastructure may receive periodic updates from vehicle 1100, such as an image sequence and / or objects that vehicle 1100 has located in the image sequence (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural networks to identify objects and compare them to the objects identified by vehicle 1100, and if the results do not match and the infrastructure concludes that the AI in vehicle 1100 has malfunctioned, server 1178 may transmit a signal to vehicle 1100 instructing the vehicle's 1100 fail-safe computer to take control, notify passengers, and complete a safe parking maneuver.

[0201] For inference, server 1178 may include a GPU 1184 and one or more programmable inference accelerators (e.g., NVIDIA's TensorRT). The combination of GPU-powered servers and inference acceleration can enable real-time responses. In other examples, such as where performance is less important, CPU, FPGA, and other processor-powered servers can be used for inference.

[0202] Example computing device

[0203] Figure 12 1 is a block diagram of an example computing device 1200 suitable for implementing some embodiments of the present disclosure. Computing device 1200 may include an interconnect system 1202 that directly or indirectly couples the following devices: memory 1204, one or more central processing units (CPUs) 1206, one or more graphics processing units (GPUs) 1208, a communication interface 1210, input / output (I / O) ports 1212, I / O components 1214, a power supply 1216, one or more presentation components 1218 (e.g., display(s)), and one or more logic units 1220. In at least one embodiment, computing device(s) 1200 may include one or more virtual machines (VMs), and / or any of its components may include virtual components (e.g., virtual hardware components). For non-limiting examples, one or more of GPUs 1208 may include one or more vGPUs, one or more of CPUs 1206 may include one or more vCPUs, and / or one or more of logic units 1220 may include one or more virtual logic units. As such, computing device(s) 1200 may include discrete components (e.g., a full GPU dedicated to computing device 1200), virtual components (e.g., a portion of a GPU dedicated to computing device 1200), or a combination thereof.

[0204] although Figure 12 The various blocks of are shown as being connected via interconnect system 1202 using wires, but this is not intended to be limiting and is provided for clarity only. For example, in some embodiments, presentation component 1218 (such as a display device) may be considered to be I / O component 1214 (e.g., if the display is a touch screen). As another example, CPU 1206 and / or GPU 1208 may include memory (e.g., memory 1204 may represent a storage device in addition to the memory of GPU 1208, CPU 1206, and / or other components). In other words, Figure 12 The term computing device is illustrative only. 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 are considered within the Figure 12 within the range of computing devices.

[0205] Interconnect system 1202 can represent one or more links or buses, such as an address bus, a data bus, a control bus, or a combination thereof. Interconnect system 1202 can include one or more bus or link types, such as an Industry Standard Architecture (ISA) bus, an Extended Industry Standard Architecture (EISA) bus, a Video Electronics Standards Association (VESA) bus, a Peripheral Component Interconnect (PCI) bus, a Peripheral Component Interconnect Express (PCIe) bus, and / or another type of bus or link. In some embodiments, there is a direct connection between components. As an example, CPU 1206 can be directly connected to memory 1204. Further, CPU 1206 can be directly connected to GPU 1208. In the case where there is a direct or point-to-point connection between components, interconnect system 1202 can include a PCIe link to perform the connection. In these examples, the PCI bus does not need to be included in computing device 1200.

[0206] Memory 1204 may include any of a variety of computer-readable media. Computer-readable media can be any available media that can be accessed by computing device 1200. Computer-readable media can include volatile and non-volatile media, as well as removable and non-removable media. By way of example and not limitation, computer-readable media can include computer storage media and communication media.

[0207] Computer storage media may include volatile and nonvolatile media and / or removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, and / or other data types. For example, memory 1204 may store computer-readable instructions (e.g., representing (one or more) programs and / or (one or more) program elements, such as an operating system). Computer storage media may include, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage devices or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by computing device 1200. As used herein, computer storage media does not include signals themselves.

[0208] Computer storage media may embody computer-readable instructions, data structures, program modules, and / or other data types in a modulated data signal such as a carrier wave or other transport mechanism, and includes any information delivery media. The term "modulated data signal" may refer to a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, computer storage media may include wired media (such as a wired network or direct-wired connection) and wireless media (such as acoustic, RF, infrared, and other wireless media). Combinations of any of the above are also intended to be included within the scope of computer-readable media.

[0209] The CPU 1206 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 1200 to perform one or more of the methods and / or processes described herein. The CPUs 1206 may each include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of handling numerous software threads simultaneously. The CPU 1206 may include any type of processor and may include different types of processors depending on the type of computing device 1200 being implemented (e.g., a processor with fewer cores for a mobile device and a processor with more cores for a server). For example, depending on the type of computing device 1200, 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). The computing device 1200 may also include one or more CPUs 1206 in addition to one or more microprocessors or supplemental coprocessors (such as a math coprocessor).

[0210] In addition to or in lieu of CPU(s) 1206, GPU(s) 1208 may be configured to execute at least some of the computer-readable instructions to control one or more components of computing device 1200 to perform one or more of the methods and / or processes described herein. One or more of GPUs 1208 may be integrated GPUs (e.g., with one or more of CPUs 1206) and / or one or more of GPUs 1208 may be discrete GPUs. In embodiments, one or more of GPUs 1208 may be coprocessors for one or more of CPUs 1206. GPU 1208 may be used by computing device 1200 to render graphics (e.g., 3D graphics) or perform general-purpose computations. For example, GPU 1208 may be used for general-purpose computing on a GPU (GPGPU). GPU 1208 may include hundreds or thousands of cores capable of handling hundreds or thousands of software threads simultaneously. The GPU 1208 may generate pixel data for an output image in response to rendering commands (e.g., rendering commands received from the CPU 1206 via a host interface). The GPU 1208 may include graphics memory (e.g., display memory) for storing pixel data or any other suitable data (e.g., GPGPU data). The display memory may be included as part of the memory 1204. The GPU 1208 may include two or more GPUs operating in parallel (e.g., via a link). The link may directly connect the GPUs (e.g., using NVLINK) or may connect the GPUs through a switch (e.g., using NVSwitch). When combined, each GPU 1208 may generate pixel data or GPGPU data for a different portion of the output or for different outputs (e.g., a first GPU for a first image and a second GPU for a second image). Each GPU may include its own memory or may share memory with other GPUs.

[0211] In addition to or in lieu of the CPU 1206 and / or GPU 1208, the logic unit 1220 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 1200 to perform one or more of the methods and / or processes described herein. In embodiments, the CPU(s) 1206, the GPU(s) 1208, and / or the logic unit(s) 1220 may execute any combination of methods, processes, and / or portions thereof, either discretely or jointly. One or more of the logic units 1220 may be part of and / or integrated into one or more of the CPU 1206 and / or GPU 1208, and / or one or more of the logic units 1220 may be discrete components or otherwise external to the CPU 1206 and / or GPU 1208. In embodiments, one or more of logic units 1220 may be a co-processor to one or more of CPUs 1206 and / or one or more of GPUs 1208 .

[0212] Examples of logic unit 1220 include one or more processing cores and / or components thereof, such as a data processing unit (DPU), a tensor core (TC), a tensor processing unit (TPU), a pixel vision core (PVC), a vision processing unit (VPU), a graphics processing cluster (GPC), a texture processing cluster (TPC), a streaming multiprocessor (SM), a tree transverse unit (TTU), an artificial intelligence accelerator (AIA), a deep learning accelerator (DLA), an arithmetic logic unit (ALU), an application-specific integrated circuit (ASIC), a floating point unit (FPU), an input / output (I / O) element, a peripheral component interconnect (PCI) or a peripheral component interconnect express (PCIe) element, etc.

[0213] The communication interface 1210 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 1200 to communicate with other computing devices via an electronic communication network (including wired and / or wireless communications). The communication interface 1210 may include components and functionality that implement communication over any of a number of different networks, such as a wireless network (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), a wired network (e.g., via Ethernet or Wi-Fi), a low-power wide-area network (e.g., LoRaWAN, SigFox, etc.), and / or the Internet. In one or more embodiments, the one or more logic units 1220 and / or the communication interface 1210 may include one or more data processing units (DPUs) for transmitting data received over the network and / or over the interconnect system 1202 directly to one or more GPUs 1208 (e.g., their memory).

[0214] I / O ports 1212 can enable computing device 1200 to be logically coupled to other devices including I / O components 1214, presentation components (one or more) 1218, and / or other components, some of which may be built into (e.g., integrated into) computing device 1200. Illustrative I / O components 1214 include a microphone, a mouse, a keyboard, a joystick, a game pad, a game controller, a satellite dish, a scanner, a printer, a wireless device, and the like. I / O components 1214 can provide a natural user interface (NUI) that processes air gestures, voice, or other physiological input generated by a user. In some cases, the input can be transmitted to an appropriate network element for further processing. The NUI can implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, gesture recognition on and near the screen, air gestures, head and eye tracking, and touch recognition (as described in more detail below) associated with the display of computing device 1200. Computing device 1200 may include a depth camera for gesture detection and recognition, such as a stereo camera system, an infrared camera system, an RGB camera system, touch screen technology, and combinations thereof. Additionally, computing device 1200 may include an accelerometer or gyroscope that enables detection of motion (e.g., as part of an inertial measurement unit (IMU)). In some examples, computing device 1200 may use the output of the accelerometer or gyroscope to render immersive augmented or virtual reality.

[0215] The power supply 1216 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 1216 may provide power to the computing device 1200 to enable the components of the computing device 1200 to operate.

[0216] The presentation component 1218 may include a display (e.g., a monitor, a touch screen, a television screen, a head-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The presentation component 1218 may receive data from other components (e.g., the GPU 1208, the CPU 1206, the DPU, etc.) and output the data (e.g., as images, video, sound, etc.).

[0217] Sample Data Center

[0218] Figure 13 An example data center 1300 that can be used in at least one embodiment of the present disclosure is shown. The data center 1300 can include a data center infrastructure layer 1310, a framework layer 1320, a software layer 1330, and / or an application layer 1340.

[0219] like Figure 13 As shown, the data center infrastructure layer 1310 may include a resource coordinator 1312, grouped computing resources 1314, and node computing resources ("node CRs") 1316(1)-1316(N), where "N" represents any complete positive integer. In at least one embodiment, the node CRs 1316(1)-1316(N) may include, but is not limited to, any number of central processing units (CPUs) or other processors (including DPUs, accelerators, field programmable gate arrays (FPGAs), graphics processors or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memories), 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 the node CRs 1316(1)-1316(N) may correspond to a server having one or more of the above-mentioned computing resources. Furthermore, in some embodiments, node CRs 1316(1)-13161(N) may include one or more virtual components, such as vGPUs, vCPUs, etc., and / or one or more of node CRs 1316(1)-1316(N) may correspond to a virtual machine (VM).

[0220] In at least one embodiment, the grouped computing resources 1314 may include separate groups of node CRs 1316 housed in one or more racks (not shown), or multiple racks housed in data centers at different geographical locations (also not shown). Separate groups of node CRs 1316 within the grouped computing resources 1314 may include grouped computing, network, memory, or storage resources that can be configured or allocated to support one or more workloads. In at least one embodiment, several node CRs 1316 including CPUs, GPUs, DPUs, and / or other processors may be grouped in 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.

[0221] Resource coordinator 1312 may configure or otherwise control one or more node CRs 1316(1)-1316(N) and / or grouped computing resources 1314. In at least one embodiment, resource coordinator 1312 may comprise a software design infrastructure (SDI) management entity for data center 1300. Resource coordinator 1312 may comprise hardware, software, or some combination thereof.

[0222] In at least one embodiment, Figure 13 As shown, the framework layer 1320 may include a job scheduler 1333, a configuration manager 1334, a resource manager 1336 and / or a distributed file system 1338. The framework layer 1320 may include a framework that supports software 1332 of the software layer 1330 and / or one or more applications 1342 of the application layer 1340. The software 1332 or the application 1342 may include network-based service software or applications, such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. The framework layer 1320 may be, but is not limited to, a free and open source software network application framework (such as Apache Spark) that can utilize the distributed file system 1338 for large-scale data processing (e.g., "big data"). TM (hereinafter referred to as "Spark"). In at least one embodiment, the job scheduler 1333 may include a Spark driver to facilitate scheduling workloads supported by the different layers of the data center 1300. The configuration manager 1334 may be capable of configuring the different layers, such as the software layer 1330 and the framework layer 1320 (which includes Spark and a distributed file system 1338 for supporting large-scale data processing). The resource manager 1336 may be capable of managing clustered or grouped computing resources that are mapped to the distributed file system 1338 and the job scheduler 1333 or allocated to support the distributed file system 1338 and the job scheduler 1333. In at least one embodiment, the clustered or grouped computing resources may include the grouped computing resources 1314 at the data center infrastructure layer 1310. The resource manager 1336 may coordinate with the resource coordinator 1312 to manage these mapped or allocated computing resources.

[0223] In at least one embodiment, the software 1332 included in the software layer 1330 may include software used by at least a portion of the node CRs 1316(1)-1316(N), the grouped computing resources 1314, and / or the distributed file system 1338 of the framework layer 1320. The 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.

[0224] In at least one embodiment, the applications 1342 included in the application layer 1340 may include one or more types of applications used by at least a portion of the node CRs 1316(1)-1316(N), the grouped computing resources 1314, and / or the distributed file system 1338 of the framework layer 1320. The one or more types of applications may include, but are not limited to, any number of genomic 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 conjunction with one or more embodiments.

[0225] In at least one embodiment, any of configuration manager 1334, resource manager 1336, and resource coordinator 1312 can implement any number and type of self-modification actions based on any amount and type of data acquired in any technically feasible manner. The self-modification actions can save a data center operator of data center 1300 from making potentially poor configuration decisions and potentially avoiding underutilized and / or poorly performing portions of the data center.

[0226] According to one or more embodiments described herein, data center 1300 may include tools, services, software, or other resources to train one or more machine learning models or use one or more machine learning models to predict or infer information. For example, the machine learning model(s) may be trained by computing weight parameters according to a neural network architecture using the software and / or computing resources described above with respect to data center 1300. 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 1300 using weight parameters computed using one or more training techniques, such as, but not limited to, those described herein.

[0227] In at least one embodiment, data center 1300 may use a CPU, an application-specific integrated circuit (ASIC), a GPU, an FPGA, and / or other hardware (or virtual computing resources corresponding thereto) to perform training and / or inference using the aforementioned resources. In addition, one or more software and / or hardware resources described above may be configured to allow a user to train or perform inference services on information, such as image recognition, speech recognition, or other artificial intelligence services.

[0228] Sample network environment

[0229] A network environment suitable for implementing embodiments of the present disclosure may include one or more client devices, servers, network attached storage (NAS), other backend devices, and / or other device types. The client devices, servers, and / or other device types (e.g., each device) may be configured to: Figure 12 The backend devices 1300 may be implemented on one or more instances of the computing device(s) 1200 - for example, each device may include similar components, features, and / or functionality of the computing device(s) 1200. In addition, in the case of implementing backend devices (e.g., servers, NAS, etc.), the backend devices may be included as part of the data center 1300, examples of which are discussed herein with respect to FIG. Figure 13 Describe in more detail.

[0230] The components of the network environment can communicate with each other via a network, which can be wired, wireless, or both. The network can include multiple networks or one of multiple networks. For example, the 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 the case 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.

[0231] Compatible network environments 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 functionality described herein for the server may be implemented on any number of client devices.

[0232] In at least one embodiment, the network environment may include one or more cloud-based network environments, distributed computing environments, combinations thereof, and the like. The 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 software supporting the software layer and / or a framework for one or more applications at the application layer. The software or application may include network-based service software or applications, respectively. In an embodiment, one or more client devices may use network-based service software or applications (e.g., by accessing the service software and / or application via one or more application programming interfaces (APIs)). The framework layer may be, but is not limited to, a free and open source software network application framework that can use a distributed file system for large-scale data processing (e.g., "big data").

[0233] A cloud-based network environment can provide cloud computing and / or cloud storage that performs any combination of the computing and / or data storage functions described herein (or one or more portions thereof). Any of these different functions can be distributed across multiple locations from a central or core server (e.g., one or more data centers that can be distributed across a state, region, country, global, etc.). If the connection to the user (e.g., 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).

[0234] The client device(s) may include the Figure 12 At least some of the components, features, and functionality of the described example computing device(s) 1200. By way of example and not limitation, the client device may be implemented as a personal computer (PC), a laptop computer, a mobile device, a smartphone, a tablet computer, a smartwatch, a wearable computer, a personal digital assistant (PDA), an MP3 player, a virtual reality headset, a global positioning system (GPS) or device, a video player, a camera, a surveillance device or system, a vehicle, a boat, a spacecraft, a virtual machine, a drone, a robot, a handheld communication device, a hospital device, a gaming device or system, an entertainment system, a vehicle computer system, an embedded system controller, a remote control, an appliance, a consumer electronic device, a workstation, an edge device, any combination of the depicted devices, or any other suitable device.

[0235] The present disclosure can be described in the general context of machine-usable instructions or computer code executed by a computer or other machine such as a personal digital assistant or other handheld device, including computer-executable instructions such as program modules. Generally, program modules including routines, programs, objects, components, data structures, etc. refer to code that performs a specific task or implements a specific abstract data type. The present disclosure can be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. The present disclosure can also be practiced in a distributed computing environment where tasks are performed by remote processing devices linked through a communication network.

[0236] As used herein, the phrase "and / or" with respect to two or more elements should be interpreted as referring to only one element or combination of elements. For example, "element A, element B, and / or element C" may 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. In addition, "at least one of element A or element B" may 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" may 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.

[0237] The subject matter of the present disclosure is described in detail herein to meet statutory requirements. However, the description itself is not intended to limit the scope of the present disclosure. On the contrary, the present inventors have contemplated that the claimed subject matter may also be embodied in other ways to include steps that are different from the steps described herein in conjunction with other current or future technologies, or combinations of similar steps. Moreover, although the terms "step" and / or "frame" may be used herein to imply different elements of the method employed, these terms should not be interpreted as implying any particular order among or between the various steps disclosed herein, unless the order of the steps is explicitly described.

[0238] Sample Clauses

[0239] A: A method comprising: generating one or more input tokens representing one or more features associated with an environment based at least on one or more representations corresponding to the environment; processing the one or more input tokens based at least on one or more language models to generate one or more output tokens representing one or more attributes associated with the one or more features; and determining information associated with the one or more features within the environment based at least on the one or more output tokens.

[0240] B: A method as described in paragraph A, wherein: the one or more features include one or more surface lines represented by the one or more representations; the one or more input tags represent the one or more surface lines represented by the one or more representations; and the one or more output tags represent at least information associated with the one or more surface lines within the one or more representations.

[0241] C: The method as described in paragraph A or paragraph B further includes determining at least one of the following based at least on the one or more output tags: one or more categories associated with one or more points corresponding to the one or more features; one or more types associated with the one or more features; one or more colors associated with the one or more features; or one or more shapes associated with the one or more features.

[0242] D: A method as described in any of paragraphs AC, wherein the one or more output tags include at least: one or more first tags associated with a first point, the first point corresponding to a feature among the one or more features; and one or more second tags associated with a second point corresponding to the feature.

[0243] E: A method as described in paragraph D, wherein determining the information associated with the one or more features within the environment includes: determining a first position associated with the first point within the environment based at least on the one or more first markers; determining a second position associated with the second point within the environment based at least on the one or more second markers; and determining a connection between the second point and the first point.

[0244] F: The method described in paragraph E further includes: determining that the first point includes a starting point associated with the feature based at least on the one or more first markers; and determining that the second point includes at least one of an intermediate point or an end point associated with the feature based at least on the one or more second markers, wherein determining the connection between the second point and the first point is based at least on the first point including the starting point and the second point including at least one of the intermediate point or the end point.

[0245] G: A method as described in any of paragraphs AF, wherein generating the one or more input tags includes: processing the one or more representations based at least on one or more models to generate at least one of feature data associated with the one or more representations or heat map data associated with the one or more representations; and generating the one or more input tags representing the one or more features associated with the environment based at least on the feature data or the at least one of the heat map data.

[0246] H: A method as described in any of paragraphs AG, wherein the one or more representations include one or more of: an intensity image corresponding to the environment; a color image corresponding to the environment; a height image corresponding to the environment; or a point cloud corresponding to the environment.

[0247] I: The method as described in any of paragraphs AH further includes one or more of the following: causing the map to be updated to at least indicate information associated with the one or more features; or causing the machine to travel based on at least the information associated with the one or more features.

[0248] J: A system comprising: one or more processors for: generating one or more representations corresponding to one or more features located within an environment; processing input data associated with the one or more representations based at least on one or more language models to generate output data representing one or more attributes associated with the one or more features; and performing one or more operations based at least on the output data.

[0249] K: A system as described in paragraph J, wherein: the input data represents one or more input tags associated with the one or more features; and the output data represents one or more output tags associated with the one or more attributes.

[0250] L: A method as described in paragraph J or paragraph K, wherein: the one or more features include one or more traffic features represented by the one or more representations; the input data represents the one or more traffic features represented by the one or more representations; and the output data represents at least one or more locations associated with the one or more features within the one or more representations.

[0251] M: A system as described in any of paragraphs JL, wherein the one or more processors are further used to determine at least one of the following based at least on the output data: one or more categories associated with one or more points corresponding to the one or more features; one or more types associated with the one or more features; one or more colors associated with the one or more features; or one or more shapes associated with the one or more features.

[0252] N: A system as described in any of paragraphs JM, wherein the output data includes at least: first output data associated with a first point, the first point corresponding to a feature among the one or more features; and second output data associated with a second point corresponding to the feature.

[0253] O: A system as described in paragraph N, wherein the one or more processors are further used to determine one or more locations associated with the one or more features within the environment by at least: determining a first location associated with the first point within the environment based on at least the first output data; determining a second location associated with the second point within the environment based on at least the second output data; and determining a connection between the second point and the first point.

[0254] P: A system as described in paragraph O, wherein the one or more processors are further used to: determine, based at least on the one or more first markers, that the first point includes a starting point associated with the feature; and determine, based at least on the one or more second markers, that the second point includes at least one of an intermediate point or an end point associated with the feature, wherein the connection between the second point and the first point is determined based at least on the first point including the starting point and the second point including at least one of the intermediate point or the end point.

[0255] Q: A system as described in any of paragraphs JP, wherein the one or more representations include one or more of: an intensity image corresponding to the environment; a color image corresponding to the environment; a height image corresponding to the environment; or a point cloud corresponding to the environment.

[0256] R: A system as described in any of paragraphs JQ, wherein the system is included in at least one of the following: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing one or more simulation operations; a system for performing one or more digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation of 3D assets; a system for performing one or more deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing one or more generative AI operations; a system for performing operations using one or more large language models (LLMs); a system for performing one or more conversational AI operations; a system for generating synthetic data; a system for presenting at least one of virtual reality content, augmented reality content, or mixed reality content; a system comprising one or more virtual machines (VMs); a system implemented at least in part in a data center; or a system implemented at least in part using cloud computing resources.

[0257] S: A processor comprising: one or more processors for performing one or more operations based at least on one or more attributes associated with one or more surface lines within an environment, wherein the one or more attributes are determined based at least on one or more first tags output using one or more language models, and the one or more language models process one or more second tags associated with the one or more surface lines identified using one or more images.

[0258] T: A processor as described in paragraph S, wherein the system is included in at least one of the following: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing one or more simulation operations; a system for performing one or more digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation of 3D assets; a system for performing one or more deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing one or more generative AI operations; a system for performing operations using one or more large language models (LLMs); a system for performing one or more conversational AI operations; a system for generating synthetic data; a system for presenting at least one of virtual reality content, augmented reality content, or mixed reality content; a system comprising one or more virtual machines (VMs); a system implemented at least in part in a data center; or a system implemented at least in part using cloud computing resources.< / vtx> < / eol>

Claims

1. A method comprising: generating one or more input tokens representing one or more features associated with the environment based at least on the one or more representations corresponding to the environment; processing the one or more input tokens based at least on one or more language models to generate one or more output tokens representing one or more attributes associated with the one or more features; as well as Information associated with the one or more features within the environment is determined based at least on the one or more output indicia.

2. The method according to claim 1, wherein: The one or more features include one or more surface lines represented by the one or more representations; the one or more input indicia representing the one or more surface lines represented by the one or more representations; as well as The one or more output indicia represent at least information associated with the one or more surface lines within the one or more representations.

3. The method of claim 1 , further comprising determining, based at least on the one or more output tags, at least one of: one or more categories associated with one or more points corresponding to the one or more features; one or more types associated with the one or more features; one or more colors associated with the one or more features; or One or more shapes associated with the one or more features.

4. The method of claim 1 , wherein the one or more output tags include at least: one or more first markings associated with a first point corresponding to a feature of the one or more features; as well as One or more second markers associated with a second point corresponding to the feature.

5. The method of claim 4, wherein determining information associated with the one or more features within the environment comprises: determining a first location associated with the first point within the environment based at least on the one or more first markers; determining a second location associated with the second point within the environment based at least on the one or more second markers; as well as A connection between the second point and the first point is determined.

6. The method according to claim 5, further comprising: determining, based at least on the one or more first markers, that the first point comprises a starting point associated with the feature; as well as determining, based at least on the one or more second markers, that the second point comprises at least one of an intermediate point or an endpoint associated with the feature, The determining of the connection between the second point and the first point is based on at least the first point including the starting point and the second point including at least one of the intermediate point or the end point.

7. The method of claim 1 , wherein generating the one or more input tokens comprises: processing the one or more representations based at least on one or more models to generate at least one of feature data associated with the one or more representations or heat map data associated with the one or more representations; as well as The one or more input tags representing the one or more features associated with the environment are generated based at least on the at least one of the feature data or the heat map data.

8. The method of claim 1 , wherein the one or more representations include one or more of: an intensity image corresponding to the environment; a color image corresponding to the environment; a height image corresponding to the environment; or A point cloud corresponding to the environment.

9. The method of claim 1, further comprising one or more of the following: causing a map to be updated to indicate at least information associated with the one or more features; or The machine is caused to travel based on at least the information associated with the one or more characteristics.

10. A system comprising: One or more processors for: generating one or more representations corresponding to one or more features located within the environment; processing input data associated with the one or more representations based at least on one or more language models to generate output data representing one or more attributes associated with the one or more features; and One or more operations are performed based at least on the output data.

11. The system of claim 10, wherein: The input data represents one or more input tags associated with the one or more features; and The output data represents one or more output tags associated with the one or more attributes.

12. The method according to claim 10, wherein: The one or more characteristics include one or more traffic characteristics represented by the one or more representations; The input data represents the one or more traffic characteristics represented by the one or more representations; and The output data represents at least one or more locations associated with the one or more features within the one or more representations.

13. The system of claim 10, wherein the one or more processors are further configured to determine, based at least on the output data, at least one of: one or more categories associated with one or more points corresponding to the one or more features; one or more types associated with the one or more features; one or more colors associated with the one or more features; or One or more shapes associated with the one or more features.

14. The system according to claim 10, wherein the output data comprises at least: first output data associated with a first point, the first point corresponding to a feature of the one or more features; as well as Second output data is associated with a second point corresponding to the feature.

15. The system of claim 14, wherein the one or more processors are further configured to determine one or more locations associated with the one or more features within the environment by at least: determining a first location associated with the first point within the environment based at least on the first output data; determining a second location associated with the second point within the environment based at least on the second output data; and A connection between the second point and the first point is determined.

16. The system according to claim 15, wherein: The one or more processors are further configured to: determining, based at least on the one or more first markers, that the first point comprises a starting point associated with the feature; and determining, based at least on the one or more second markers, that the second point comprises at least one of an intermediate point or an endpoint associated with the feature, The connection between the second point and the first point is determined based on at least the first point including the start point and the second point including at least one of the intermediate point or the end point.

17. The system of claim 10, wherein the one or more representations include one or more of: an intensity image corresponding to the environment; a color image corresponding to the environment; a height image corresponding to the environment; or A point cloud corresponding to the environment.

18. The system of claim 10, wherein the system is included in at least one of the following: control systems for autonomous or semi-autonomous machines; Perception systems for autonomous or semi-autonomous machines; a system for performing one or more simulation operations; a system for performing one or more digital twin operations; a system for performing light transport simulations; A system for performing collaborative content creation of 3D assets; a system for performing one or more deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system for performing one or more generative AI operations; A system for performing operations using one or more large language models (LLMs); A system for performing one or more conversational AI operations; Systems for generating synthetic data; a system for presenting at least one of virtual reality content, augmented reality content, or mixed reality content; A system comprising one or more virtual machines VM; A system implemented at least in part in a data center; or A system implemented at least in part using cloud computing resources.

19. A processor comprising: One or more processors for performing one or more operations based at least on one or more attributes associated with one or more surface lines within an environment, wherein the one or more attributes are determined based at least on one or more first tags output using one or more language models, and the one or more language models process one or more second tags associated with the one or more surface lines identified using one or more images.

20. The system of claim 19, wherein the system is included in at least one of the following: control systems for autonomous or semi-autonomous machines; Perception systems for autonomous or semi-autonomous machines; a system for performing one or more simulation operations; a system for performing one or more digital twin operations; a system for performing light transport simulations; A system for performing collaborative content creation of 3D assets; a system for performing one or more deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system for performing one or more generative AI operations; A system for performing operations using one or more large language models (LLMs); A system for performing one or more conversational AI operations; Systems for generating synthetic data; a system for presenting at least one of virtual reality content, augmented reality content, or mixed reality content; A system comprising one or more virtual machines VM; A system implemented at least in part in a data center; or A system implemented at least in part using cloud computing resources.

Images