Perceptual data fusion for autonomous systems and applications

Abstract

The invention discloses perceptual data fusion for autonomous systems and applications. Techniques are disclosed for fusing first information generated using one or more learning models with second information generated using one or more non-learning processes to generate one or more updated versions of third information including the first information and / or the second information. In some examples, the first information may indicate one or more locations associated with one or more first objects in the environment, and the second information may indicate one or more attributes associated with one or more second objects in the environment. In some cases, the learning model may generate the first information based at least on first sensor data generated using one or more first sensors of the machine, and the non-learning process may generate the second information based at least on second sensor data generated using one or more second sensors of the machine.

Description

Background Art

[0001] Machines (e.g., autonomous vehicles or machines, semi-autonomous vehicles or machines, etc.) may use various types of sensor modalities (including but not limited to image sensors, radar (RADAR) sensors, laser radar (LiDAR) sensors, and ultrasonic sensors) to obtain information associated with their surroundings. Because each sensor type may provide unique advantages and / or limitations, combining the advantages of these sensor modalities while mitigating their respective limitations is critical to achieving robust obstacle detection within the close proximity of the machine. Similarly, the perception systems of these machines may also employ various processing pipelines (such as learning models, algorithmic processing, probabilistic models, etc.) to accurately interpret the surrounding environment and make intelligent decisions for safe navigation.

[0002] However, like different types of sensor modalities, different processing pipelines used by machine perception systems may also have their own advantages and / or disadvantages. For example, learned models (e.g., machine learning models) can provide advantages for recognizing unclassified objects and accurately estimating object shapes and sizes, which makes learned models particularly valuable in complex environments where objects may have a wide variety of appearances and / or configurations. On the other hand, classical (e.g., non-learning) methods, such as algorithmic processing techniques, may excel in well-understood scenarios and maintain performance in edge cases where training data is limited, making them a more reliable choice for ensuring consistent performance. Summary of the Invention

[0003] Embodiments of the present disclosure relate to perceptual data fusion for autonomous or semi-autonomous systems and applications. For example, the systems and methods described herein can fuse first information generated using one or more learned models (e.g., one or more deep neural networks) with second information generated using one or more non-learning processes (e.g., one or more algorithmic processes) to generate third information comprising one or more updated (e.g., refined, improved, etc.) versions of the first information and / or the second information. That is, these systems and methods can combine the advantages of learned models and non-learning processes by refining and / or improving the first information based on at least the second information, and vice versa.

[0004] Compared to conventional systems, such as those described above, in some embodiments, the present system is able to flexibly combine one or more learned methods with one or more classical sensor processing techniques to robustly detect obstacles within the close proximity of a machine using a combination of different sensor modalities (e.g., image, RADAR, LiDAR, ultrasound, etc.). Thus, as described in more detail herein, by executing these processes, the present system is able to effectively combine the advantages of deep neural networks (such as their ability to identify unclassified objects and accurately estimate their shape and size) with the advantages of classical processing and / or algorithmic approaches (including their ability to reliably ensure consistent performance in well-understood scenarios while maintaining acceptable performance in edge cases). This provides an improvement over conventional systems, which require the outputs of these different techniques to be evaluated independently of each other, which can lead to inaccurate object detection and / or ultimately impact the machine's ability to make informed decisions. Furthermore, by fusing the information generated by the learned methods and classical methods, the present system can more accurately perceive the machine's near-field surroundings, including more precisely locating objects and correctly identifying their corresponding attributes. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The present system and method for perceptual data fusion for autonomous or semi-autonomous systems and applications will be described in detail below with reference to the accompanying drawings, in which:

[0006] Figure 1 is a data flow diagram illustrating an example process for fusing different information generated using learned and classical sensor processing techniques according to some embodiments of the present disclosure;

[0007] Figure 2 is an illustration of an example environment in which one or more machines according to some embodiments of the present disclosure may operate;

[0008] Figure 3 is an illustration of an example occupancy representation associated with an example environment, which may be generated using one or more learned techniques, according to some embodiments of the present disclosure;

[0009] Figure 4 is an illustration of example object data indicating one or more properties associated with one or more objects in an example environment, which may be generated using one or more processing techniques, according to some embodiments of the present disclosure;

[0010] Figure 5 is an illustration of an example occupancy history representation associated with an example environment according to some embodiments of the present disclosure;

[0011] Figure 6 is an illustration of an occupancy representation including example updates according to some embodiments of the present disclosure;

[0012] Figure 7 is a data flow diagram illustrating an example process for training a fusion component according to some embodiments of the present disclosure;

[0013] Figure 8 is a flow chart illustrating an example method for fusing information generated using learned and classical sensor processing techniques according to some embodiments of the present disclosure;

[0014] Figure 9 is a flow chart illustrating another example method for fusing information obtained using different processing techniques according to some embodiments of the present disclosure;

[0015] Figure 10A is an illustration of an example autonomous vehicle according to some embodiments of the present disclosure;

[0016] Figure 10B According to some embodiments of the present disclosure Figure 10A Examples of camera positions and fields of view for autonomous vehicles;

[0017] Figure 10C According to some embodiments of the present disclosure Figure 10A a block diagram of an example system architecture for an example autonomous vehicle;

[0018] Figure 10D According to some embodiments of the present disclosure, a method for Figure 10A System diagram of an example of communication between autonomous vehicles;

[0019] Figure 11 is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure; and

[0020] Figure 12 is a block diagram of an example data center suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0021] Systems and methods are disclosed relating to sensory data fusion for autonomous or semi-autonomous systems and applications. Although the present disclosure may be directed to an example autonomous or semi-autonomous vehicle or machine 1000 (also referred to herein as "vehicle 1000," "ego vehicle 1000," "ego machine 1000," or "machine 1000"), its examples may be directed to an example autonomous or semi-autonomous vehicle or machine 1000. Figures 10A-10DThe present disclosure may be described with respect to sensory data fusion in autonomous or semi-autonomous systems and applications, but this is not intended to be limiting, and the systems and methods described herein may be used with, 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, robots or robotic platforms with and without drivers, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, airships, boats, shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, engineering vehicles, underwater vehicles, drones, and / or other vehicle types. Furthermore, while the present disclosure may be described with respect to sensory data fusion in autonomous or semi-autonomous systems and 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, and / or any other technology space that may use object or feature detection and / or map creation.

[0022] For example, the system may generate first data indicating one or more first locations associated with one or more first objects in the environment. In some examples, the system may generate the first data using one or more learned models. In some cases, the learned models may include machine learning models, deep neural networks (DNNs), convolutional neural networks (CNNs), and / or any other type of model that can be trained to perform object detection, object classification, object tracking, etc. In some examples, the system may generate the first data based at least on first sensor data obtained using one or more first sensors of the machine. In some examples, the first sensors may include one or more image sensors, one or more RADAR sensors, one or more LiDAR sensors, one or more ultrasonic sensors, etc. Thus, the first sensor data may include one or more of image data, RADAR data, LiDAR data, ultrasonic data, etc. In some examples, the first data may be instantaneous (e.g., a single shot corresponding to a single instance of data or time, etc.) data that indicates a first location associated with a first object in the environment surrounding the machine at that instance in time.

[0023] In some examples, the first data may include occupancy data. For example, the first data may include an occupancy representation (e.g., a dense occupancy map, a dense occupancy grid, a top-down dense occupancy representation, etc.) associated with the environment from a top-down perspective (e.g., overhead, a bird's eye view (BEV), etc.). The occupancy representation may include one or more points and / or pixels that represent one or more samples obtained using the first sensor at a time instance. That is, the points and / or pixels may indicate various information associated with the environment, and each point / pixel of the occupancy representation may indicate information about a specific location in the environment. For example, as described below with respect to Figure 3 Described in more detail, one or more first points of the occupancy representation may correspond to a first location associated with a first object, one or more second points in the occupancy representation may correspond to one or more unoccupied locations in the environment at that time instance, and one or more third points in the occupancy representation may correspond to one or more occluded portions of the environment at that time instance. Additionally, in some cases, one or more values ​​of a point may correspond to or otherwise indicate at least one of a height or confidence associated with one or more samples. That is, the value (e.g., color, shading, opacity, etc.) of a point or pixel of the occupancy representation may indicate the height of at least a portion of an object at that location and / or the confidence or certainty associated with the point (e.g., whether the point actually corresponds to the portion of the object, the confidence of the height estimate of the portion of the object, etc.). Thus, multiple points or pixels in the occupancy representation and their corresponding values ​​may collectively indicate the shape of an object, the size of an object, an occupied portion of the environment, an unoccupied portion of the environment, an occluded portion of the environment, etc.

[0024] In some examples, the system may generate second data indicating one or more first attributes associated with one or more second objects in the environment. In some cases, the one or more second objects may be the same as the one or more first objects, and / or the one or more second objects may be different from the one or more first objects. The system may generate the second data based at least on second sensor data obtained using one or more second sensors of the machine. In some examples, the second data may include a list of objects in the environment. The object list may include first attributes associated with the second objects. For example, for one or more second objects included in the object list, the first attributes corresponding to these objects may also be listed. In some examples, the attributes that may be included in the object list may include one or more of the object's location (e.g., coordinates), the object's bounding shape (e.g., a bounding box, a bounding rectangle, a bounding polygon, etc.), the object's trajectory (e.g., velocity, acceleration, and / or direction), the object's pose or direction, the object's classification (e.g., vehicle, pedestrian, cyclist, animal, etc.), the object's shape, the object's size, whether the object is static or dynamic, etc.

[0025] In some examples, the system can use one or more classical (e.g., non-learning) methods (such as algorithmic sensor processing, probabilistic processing, threshold processing, feature extraction, filtering, etc.) to generate the second data, which methods can be unimodal and / or fused. In some examples, the second object, the second sensor data, and / or the second sensor can be the same as or different from the first object, the first sensor data, and / or the first sensor described above. For example, in some examples, one or more second sensors can include one or more image sensors, one or more RADAR sensors, one or more LiDAR sensors, one or more ultrasonic sensors, etc. In some examples, the second data can be time data, and the first attribute can be tracked or otherwise determined within a time period at least partially before the time instance. In this way, the second sensor data can include multiple instances and / or snapshots of sensor data obtained at one or more time instances throughout the time period. In some examples, the time period can be before the time instance, and in some cases, can include the time instance.

[0026] In some examples, the system can generate third data based at least on the first data and / or the second data. As described herein, in some examples, the system can generate third data by fusing the first data and the second data using one or more learned models, one or more classical (e.g., non-learning) methods, and / or any other fusion technology. The third data may include or otherwise represent one or more updated versions of the first data and / or the second data. For example, the third data and / or the updated version may include one or more refined and / or improved versions of the first data and / or the second data. In some cases, the system may use at least a portion of the first data and / or the second data to determine the updated version. That is, for example, the system may determine the updated version of the first data based at least on a portion (e.g., some or all) of the second data. Similarly, the system may determine the updated version of the second data based at least on a portion (e.g., some or all) of the first data. In this way, since the system uses the learned model to determine the first data and the classical method to determine the second data, the system can effectively combine the advantages of the learned model and the classical method with the generation of the third data.

[0027] In order to generate the third data, in some examples, the system can determine one or more first objects corresponding to one or more second objects. For example, the first data can indicate a first position associated with the first object, and the second data can indicate a first attribute associated with the second object. As described above and herein, the first attribute can include a position, an enclosing shape, a trajectory, a posture, and / or other information associated with the second object. Based at least on the first position of the first data and the various features of the first attribute, the system can determine, for example, a first position associated with the first object, which corresponds to the various features (position, enclosing shape, trajectory, posture, etc.) in the first attribute. In this way and in other ways, the system can determine that a first object in one or more first objects corresponds to a second object in one or more second objects (for example, by matching position, shape, size, posture, etc. between the first data and the second data).

[0028] In some examples, the third data may include an updated version of the first data. The updated version of the first data may indicate one or more second locations associated with the first object. In some examples, based at least on the first attribute of the second data, the second location may be more accurate and / or more refined than the first location. That is, in some cases, the system may determine the second location from a selective portion of the first data and / or the second data in such a manner that the second location more accurately / precisely corresponds to the actual location of the first object.

[0029] In some examples, the updated version of the first data may include an updated version of the occupancy representation. The updated version of the occupancy representation may include one or more second points and / or pixels based at least on the second data. For example, the updated version of the occupancy representation may be refined and / or improved based on a first attribute included in the second data, and the updated version of the occupancy representation may include second points and / or pixels that may be changed or added relative to the original version of the occupancy representation in the first data. For example, based on the position of the enclosing shape of the object in the second data, one or more points / pixels in the updated version of the occupancy representation may be modified to indicate a more precise position, size, shape, etc. of the object. Additionally, in some examples, the first data and / or the updated version of the occupancy representation may indicate whether an occluded portion of the environment is occupied by at least one of the first objects or at least one of the second objects. In other words, the occupancy representation may be updated to indicate information related to the occluded portion of the environment that the model learned on the fly may not be able to determine.

[0030] In some examples, the third data may additionally or alternatively include an updated version of the second data. The updated version of the second data may indicate one or more second attributes associated with the second object. In some examples, based at least on the first position included in the first data, the second attribute may be more accurate and / or more refined than the first attribute. That is, in some cases, the system may determine the second attribute based on at least a selective portion of the first data and / or the second data in a manner such that the second attribute more accurately / precisely corresponds to the actual attribute or quality of the second object. For example, in the updated version of the second data, the first enclosing shape associated with the object included in the first attribute may be updated to a smaller or larger size, a different shape, etc., which more closely corresponds to the actual size or shape of the object. Additionally or alternatively, the first posture / orientation associated with the object in the first attribute may be updated in the updated version of the second data to more accurately correspond to the actual posture / orientation of the object. In some cases, the updated version of the first data may include an updated version of the object list.

[0031] In some examples, as part of determining the third data, the system may generate fourth data indicating one or more prior locations associated with a first object in the environment. The fourth data may represent a historical representation of occupancy associated with the environment over an entire time period. For example, the fourth data may include at least one or more points (e.g., pixels) representing one or more prior samples obtained using the first sensor over a time period, and may be refined / improved based on at least the first attribute. In some examples, each of the one or more points included in the fourth data may indicate historical speed information associated with the first object, historical height information associated with the first object, historical confidence information associated with the first object, and the like. In some examples, the system may determine the third data based at least on the fourth data.

[0032] In some examples, the one or more systems (and / or additional systems) can train one or more machine learning models to fuse the first data and the second data. That is, the system can train the machine learning model to combine the outputs determined using one or more learned methods and one or more classical sensor processing techniques to robustly detect obstacles within close range of the machine. For example, as described in more detail herein, the system can train the machine learning model using at least training data generated by different types of processing pipelines (e.g., learned models, classical methods, etc.) and true value data representing actual values ​​of parameters associated with objects represented by the training sensor data. In addition, during training, the system can generate and input data similar to the data input into the machine learning model when the vehicle or machine uses it.

[0033] In some examples, the system may cause the machine to perform one or more operations based at least on the third data. That is, in some cases, the system may cause the machine to perform one or more operations based at least on at least one of the updated version of the first data and / or the updated version of the second data. In some examples, the one or more operations may include providing the third data to a planning component of the machine to update the machine's trajectory. Additionally or alternatively, the system may update the machine's behavior based at least on the third data, such as causing the machine to operate at a slower speed than normal, causing the machine to increase the distance between itself and other objects in the environment, and so on.

[0034] 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, airships, boats, shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, engineering vehicles, submarines, drones, and / or other vehicle types. In addition, the systems and methods described herein may be used for a variety of purposes, such as, but not limited to, machine control, machine motion, machine driving, synthetic data generation, model training, perception, augmented reality, virtual reality, mixed reality, robotics, safety and surveillance, simulation and digital twins, autonomous or semi-autonomous machine applications, deep learning, environmental simulation, object or actor simulation and / or digital twins, data center processing, conversational 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.

[0035] The disclosed embodiments may be included in a variety of different systems, such as automotive systems (e.g., control systems for autonomous or semi-autonomous machines, perception systems for autonomous or semi-autonomous machines), systems implemented using robots, aviation systems, medical systems, boating systems, smart area monitoring systems, systems for performing deep learning operations, systems for performing simulation operations, systems for performing digital twin operations, systems implemented using edge devices, systems implementing a large language model (LLM), systems implementing one or more visual language models (VLM), 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.

[0036] refer to Figure 1 , Figure 1is a data flow diagram illustrating an example process 100 for fusing different information generated using learned and classical sensor processing techniques according to some embodiments of the present disclosure. It should be understood that this arrangement and other arrangements described herein are set forth by way of example only. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of 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 conjunction with other components and in any suitable combination and location. The various functions performed by the entities described herein may be implemented by hardware, firmware, and / or software. For example, the various functions may be implemented by a processor executing instructions stored in a memory. In some embodiments, the systems, methods, and processes described herein may be implemented using Figures 10A-10D Example autonomous vehicle 1000, Figure 11 The example computing device 1100 of , and / or Figure 12 The example data center 1200 may be performed using components, features, and / or functions similar to the components, features, and / or functions of the example data center 1200.

[0037] Process 100 includes one or more machine learning models 102 and one or more processing components 104 that obtain sensor data 106, perception data 108, and / or fused perception data 110. Machine learning model 102 and processing component 104 can then generate occupancy data 112 and object data 114, respectively, based at least on one or more of sensor data 106, perception data 108, and / or fused perception data 110.

[0038] In an example, sensor data 106 may be generated using one or more sensors 116 of a machine. Sensors 116 may include, but are not limited to, one or more of an image sensor, a RADAR sensor, a LiDAR sensor, an ultrasonic sensor, an environmental sensor, and / or any other type of sensor. Thus, sensor data 106 may include, but are not limited to, one or more of image data, RADAR data, LiDAR data, ultrasonic data, environmental data, and / or any other type of sensor data. In some examples, sensors 116 may include the first and second sensors described herein. Furthermore, sensor data 106 may include the first and second sensor data described herein.

[0039] In some examples, one or more perception components 118 can generate perception data 108 based at least on sensor data 106. Perception component 118 can include one or more object detectors and / or one or more object trackers of different sensor modalities. For example, perception component 118 can include a first object detector and / or a first object tracker that detects and / or tracks objects from, for example, image data, and a second object detector and / or a second object tracker that detects and / or tracks objects from, for example, RADAR data. In some examples, perception data 108 can, in some cases, include bounding shapes associated with detected and / or tracked objects in the environment. Thus, in some examples, perception data 108 can include first perception data associated with a first sensor modality (e.g., based on image data), second perception data associated with a second sensor modality (e.g., based on RADAR data), third perception data associated with a third sensor modality (e.g., based on ultrasonic data), and so on.

[0040] In an example, one or more object fusion components 120 can generate fused perception data 110 obtained by the machine learning model 102 and / or the processing component 104. The object fusion component 120 can generate the fused perception data 110 based on at least the perception data 108. For example, as described above and herein, the perception data 108 can include one or more instances of perception data of different sensor modalities, and the object fusion component 120 can fuse the different instances of perception data together as the fused perception data 110. That is, in some examples, the object fusion component 120 can improve the perception data 108 by selectively combining the strengths of each of the different sensor modalities into a single instance of the fused perception data 110.

[0041] In an example, sensor data 106, perception data 108, and / or fused perception data 110 may represent multiple objects in an environment. For example, Figure 2 is an illustration of an example environment 204 in which one or more machines, such as machine 202 (eg, a vehicle), may operate according to some embodiments of the present disclosure. Figure 2In the example of , while navigating around environment 204, machine 202 may use sensors (e.g., one or more sensors 116, etc.) to generate sensor data (e.g., sensor data 106, etc.) that represents multiple objects 206(1)-206(4) (also referred to as "object 206" in the singular or "multiple objects 206" in the plural) near machine 202 in environment 204. In addition, machine 202 may use one or more components (e.g., perception component 118, object fusion component 120, etc.) to process sensor data (e.g., sensor data 106, etc.) and generate perception information associated with object 206 (e.g., perception data 108, fused perception data 110, etc.). Although Figure 2 The example shows that object 206(1) includes another machine, objects 206(2)-206(3) include pedestrians, and object 206(4) includes traffic cones, but in other examples, objects 206 may include any other type of object (e.g., a cyclist, a structure, a road sign, an animal, vegetation, etc.). In addition, although Figure 2 The example of FIG202 describes sensor data as representing four objects 206 , but in other examples, machine 202 can generate sensor data representing any number of objects (e.g., one object, ten objects, fifty objects, one hundred objects, one thousand objects, etc.). Furthermore, while described as detecting, tracking, or determining parameters of objects, the objects can be extended to static objects, dynamic objects, undulations / perturbations in a surface (such as a driving surface), cavities or holes in a surface, and / or features in the environment (road markings, etc.).

[0042] Return Reference Figure 1 In an example, process 100 may include a machine learning model 102 that generates occupancy data 112 based on at least one or more of sensor data 106, perception data 108, and / or fused perception data 110. In some examples, machine learning model 102 may include one or more deep neural networks (DNNs) for processing sensor data 106, perception data 108, and / or fused perception data 110 to generate occupancy data 112. In some examples, occupancy data 112 may correspond to the first data described above, which indicates one or more first locations associated with one or more first objects in the environment. In some examples, occupancy data 112 may be instantaneous (e.g., a single shot corresponding to a single data instance and / or a single time instance, etc.) data that indicates a first location associated with a first object in the environment surrounding the machine—e.g., at a time instance.

[0043] In some examples, occupancy data 112 may include an occupancy representation associated with an environment (e.g., a dense occupancy map, a dense occupancy grid, etc.). Figure 3 is an illustration of an example occupancy representation 302 associated with an example environment 204, which may be generated using one or more learning techniques (e.g., machine learning model 102), in accordance with some embodiments of the present disclosure. In some examples, the occupancy representation 302 may be associated with a top-down (e.g., overhead, bird's-eye view, etc.) perspective. The occupancy representation 302 may include one or more points and / or pixels representing one or more samples obtained using sensor 116 at a time instance. That is, the points and / or pixels may indicate various information associated with the environment, and each point / pixel of the occupancy representation 302 may indicate information for a particular location in the environment. For example, one or more first points 304(1)-304(3) (also referred to as "first point 304" in the singular or "first points 304" in the plural) of the occupancy representation 302 may correspond to objects 206(1)-206(3), respectively. Additionally, one or more second points 306(1)-306(2) of the occupancy representation 302 (also referred to as "second point 306" in the singular or "second points 306" in the plural) can correspond to one or more occluded portions of the environment at the time instance. Furthermore, one or more third points of the occupancy representation 302 can correspond to one or more unoccupied portions of the environment at the time instance, which can be represented by Figure 3 302, such as one or more unoccupied portions 308. In other words, for ease of illustration and understanding, third points are not included in occupancy representation 302; however, blank (e.g., white) space within occupancy representation 302 may include one or more third points, and the third points may have a different appearance (e.g., shading, color, etc.) than first point 304 and / or second point 306.

[0044] Still refer to Figure 3 In some cases, one or more values ​​of one or more points can correspond to or otherwise indicate at least one of a height or a confidence level associated with a sample. That is, the value (e.g., color, shading, opacity, etc.) of one of the first point 304 and / or the second point 306 in the occupancy representation 302 can indicate the height of at least a portion of an object at that location and / or the confidence level or certainty associated with that point (e.g., whether the point actually corresponds to that portion of the object, the confidence level of the height estimate for that portion of the object, etc.). For example, a first point 304(1) corresponding to the object 206(1) can be represented using one or more first colors that indicate different height measurements associated with the object 206(1) at those corresponding points / locations. Additionally, a first point 304(2) corresponding to the object 206(2) can be represented using one or more second colors that indicate different height measurements associated with the object 206(2) at those corresponding points / locations.

[0045] Return Reference Figure 1 In an example of a process 100, a processing component 104 may be configured to process sensor data 106, perception data 108, and / or fused perception data 110 using one or more algorithmic sensor processing and / or other classical (e.g., non-learning) methods to generate object data 114. In some examples, the object data 114 may indicate one or more attributes corresponding to one or more objects represented in the sensor data 106, perception data 108, and / or fused perception data 110. In some examples, the object data 114 may include a list of objects and / or their respective attributes. These attributes associated with the objects may include, but are not limited to, the location of the object, the bounding shape associated with the object, the trajectory of the object, the pose / orientation associated with the object, and / or the classification of the object. In contrast to the occupancy data 112, which may be instantaneous or correspond to a single time or data instance, the object data 114 may be temporal and tracked over a period of time. However, in some embodiments, the occupancy data may also include a temporal element in order to apply smoothing to the results over time, such as by combining multiple current and past outputs using, for example, a weighting scheme that favors more recent data instances.

[0046] In some examples, object data 114 indicating properties associated with an object can be represented visually and / or as a list. As an example of object data 114 being visually presented, Figure 4 is an illustration of example object data 114 indicating properties associated with an object 206 in an example environment 204 that may be generated using one or more processing techniques (e.g., algorithmic processing) employed by processing component 104, according to some embodiments of the present disclosure. Figure 4 In the example, the properties shown may correspond to Figure 2206(1)-206(4). The object data 114 may include or otherwise indicate one or more enclosing shapes 402(1)-402(4) (also referred to singularly as "enclosing shape 402" or plurally as "enclosing shapes 402"), one or more trajectories 404(1)-404(3) (also referred to singularly as "trajectory 404" or plurally as "trajectories 404"), and / or one or more positions 406(1)-406(4) (also referred to singularly as "positions 406" or plurally as "positions 406"). The enclosing shape 402(1), the trajectory 404(1), and the position 406(1) may correspond to the object 206(1). Similarly, enclosing shapes 402(2) and 402(3), trajectories 404(2) and 404(3), and positions 406(2) and 406(3), respectively, may correspond to objects 206(2) and 206(3), and enclosing shape 402(4) and position 406(4) may correspond to object 206(4). In some examples, object data 114 may further indicate a classification of object 206, such as by the color, shading, shape, etc., of enclosing shape 402(1), where a label indicates the classification, etc.

[0047] Additionally or alternatively, the object data 114 can be non-visually represented as a list of objects. For example, the list can include one or more identifiers corresponding to the objects 206, and each identifier for each object can be associated with one or more listed attributes. For an object identifier in the object list, the listed attributes can include, but are not limited to, the object's location (e.g., coordinates), the object's enclosing shape (e.g., coordinates or other measurements indicating the size or perimeter of the enclosing shape), the object's trajectory (e.g., speed, acceleration, and / or direction), the object's pose or orientation (e.g., the object's heading in degrees), the object's classification (e.g., vehicle, pedestrian, cyclist, animal, etc.), an indication of whether the object is static or dynamic, etc.

[0048] Return Reference Figure 1In the example process 100 shown in FIG, a fusion component 122 can obtain occupancy data 112 from a machine learning model 102 and object data 114 from a processing component 104 and generate updated occupancy data 124 and / or updated object data 126. In some examples, the fusion component 122 can generate updated occupancy data 124 based at least on at least a portion of the occupancy data 112 and the object data 114. Similarly, the fusion component 122 can generate updated object data 126 based at least on at least a portion of the object data 114 and the occupancy data 112. Furthermore, in some examples, the fusion component 122 can obtain sensor data 106 and / or perception data 108 directly from the sensor 116 and / or perception component 118 and also determine updated occupancy data 124 and / or updated object data 126 based at least on at least a portion of that data. In an example, the updated occupancy data 124 can represent a refined and / or improved version of the occupancy data 112. Similarly, the updated object data 126 can represent a refined and / or improved version of the object data 114.

[0049] As described above, in some cases, the fusion component 122 can use at least a portion of the object data 114 to determine updated occupancy data 124. That is, for example, the fusion component 122 can determine updated occupancy data 124 based on at least a portion (e.g., some or all) of the object data 114. For example, the fusion component 122 can use one or more bounding shapes from the object data 114 to refine points included in the occupancy data 112 and / or reduce occluded areas. Similarly, the fusion component 122 can use at least a portion of the occupancy data 112 to determine updated object data 126. For example, the fusion component 122 can use one or more points included in the occupancy data 112 to refine one or more bounding shapes, poses, classifications, etc. in the object data 114. In this way, the fusion component 122 can effectively combine the strengths of the machine learning model 102 and the processing component 104, while reducing and / or minimizing their weaknesses, to generate one or more improved data structures that can convey robust information associated with the surrounding environment so that the machine can make safer and / or more informed decisions.

[0050] To generate updated occupancy data 124 and / or updated object data 126, in some examples, the fusion component 122 can determine one or more corresponding objects between the occupancy data 112 and the object data 114. For example, the fusion component 122 can identify a first object in the occupancy data 112 that corresponds to a second object in the object data 114, identify a third object in the occupancy data 112 that corresponds to a fourth object in the object data 114, and so on. For example, referring to Figure 2-Figure 4, the occupancy representation 302 of the occupancy data 112 may include a first point 304(1) indicating a location associated with the object 206(1), and the object data 114 may indicate an enclosing shape 402(1) and / or a location 406(1) as attributes associated with the object 206(1), as well as other useful information not shown that may be used in the fusion process. In some examples, the fusion component 122 may process the occupancy data 112 and the object data 114 to determine an alignment between features included in the occupancy data 112 (e.g., the occupancy representation 302) and the object data 114. The fusion component 122 may then determine, based at least on the processing, that the first point 304(1), the enclosing shape 402(1), and / or the location 406(1) all correspond to the object 206(1). Based on the correspondence, the fusion component 122 may associate the features with the object 206(1). In examples, the fusion component 122 may repeat the process one or more times to determine which features / objects in the occupancy data correspond to features / objects in the object data 114, and / or vice versa.

[0051] Furthermore, in some examples, as part of generating updated occupancy data 124 and / or updated object data 126, fusion component 122 can generate occupancy history data indicating one or more previous locations associated with objects in the environment. For example, Figure 5 is an illustration of an example occupancy history representation 502 associated with an example environment 204 according to some embodiments of the present disclosure. The occupancy history representation 502 may indicate various locations of one or more previous points 504(1)-504(3) (also referred to singularly or collectively as "previous points 504") corresponding to objects 206 in the environment 204 throughout a time period. For example, previous point 504(1) may correspond to object 206(1), previous point 504(2) may correspond to object 206(2), and previous point 504(3) may correspond to object 206(3). In some examples, the previous points 504 may be determined from previous instances of updated occupancy data 124 and / or updated object data 126. Additionally or alternatively, the previous points 504 may represent one or more prior samples obtained using the sensors 116 during the time period and refined / refined based on at least the object data 114. In some examples, previous points 504 may indicate historical speed information associated with object 206 , historical altitude information associated with object 206 , historical confidence information, and the like.

[0052] Return Reference Figure 1For example, the updated occupancy data 124 may include one or more additional points and / or pixels relative to the occupancy data 112 based at least on the object data 114. Additionally or alternatively, the updated occupancy data 124 may include one or more improved, altered, and / or refined points and / or pixels relative to the occupancy data 112. That is, one or more values ​​of the original points included in the occupancy representation 302 may be changed from a first value to a second value based at least on the object data 114. For example, based on the position of the enclosing shape of the object in the object data 114, one or more points / pixels in the updated occupancy data / representation may be modified to indicate a more accurate position, size, shape, etc. of the object.

[0053] For example, Figure 6 is an illustration of an example updated occupancy representation 602 according to some embodiments of the present disclosure. In some examples, the updated occupancy representation 602 that can be included in the updated occupancy data 124 can include one or more updated points 604(1)-604(4) corresponding to the object 206 in the environment 204. Additionally, in some examples, the updated occupancy representation 602 can include bounding shapes 402(1)-402(4) and trajectories 404(1)-404(3) associated with the object 206. In some examples, these features from the object data 114 can be superimposed on the updated occupancy representation 602. In some examples, because the features from the object data 114 are temporal and tracked over time, these features (e.g., bounding shapes 402, trajectories 404, and other features) can help the fusion component 122 smooth the updated points 604 and generate a realistic representation.

[0054] In addition, the updated occupancy representation 602 may include one or more features that were not included in the original occupancy representation 302. For example, the machine learning model 102 may inadvertently miss detecting various objects in the environment, such as the object 206(4) including a traffic cone. As a result, the occupancy data 112 and / or the occupancy representation 302 may omit one or more points corresponding to these features. However, because the fusion component 122 is able to extract various features from the sensor data 106, the perception data 108, and / or the object data 114, such as the enclosing shape 402(4) corresponding to the object 206(4), the fusion component 122 may update the occupancy representation with one or more of these features. In addition, in some cases, the fusion component 122 may update the updated occupancy representation 602 to include the updated point 604(4) corresponding to the object 206(4). In some examples, the updated point 604(4) may be included in the sensor data 106 and / or the perception data 108.

[0055] Still refer to Figure 6In some examples, the updated occupancy representation 602 can update the occluded portions of the environment (e.g., those indicated by the second point 306 of the original occupancy representation 302) to indicate more information about these areas. For example, the fusion component 122 can generate the updated occupancy representation 602 to indicate whether the occluded portions are occupied by one or more objects. For example, if the object data 114 indicates a bounding box within one of the occluded portions, or if the sensor data 106 and / or the perception data 108 indicate one or more such features, the updated occupancy representation 602 can indicate the presence and / or location of these objects / features.

[0056] Now return to reference Figure 1 For example, in some examples, updated object data 126 may indicate one or more updated (e.g., refined, improved, etc.) properties associated with object 206. In some examples, updated object data 126 may include an updated object list. In some cases, such an updated object list may include more or fewer objects and / or properties than included in object data 114 (e.g., more objects if occupancy data 112 indicates additional objects, more properties if occupancy data 112 provides information indicating previously unknown properties, etc.). Additionally or alternatively, the updated object list in updated object data 126 may include one or more updated properties associated with the object (e.g., an updated bounding shape that more accurately represents the object, an updated trajectory based on a newly detected pose of the object from occupancy data 112, etc.). In some examples, based at least on the information included in occupancy data 112, the properties of updated object data 126 may be more accurate and / or more refined than the properties of object data 114.

[0057] In some examples, process 100 includes one or more downstream components 128 that obtain one or more of updated occupancy data 124 and / or updated object data 126. In some examples, downstream components 128 may include one or more systems of a machine, such as a tracking system configured to create, update, and / or terminate tracks associated with objects around the vehicle. While these are just a few examples of what additional downstream components 128 may include, in other examples, other downstream components 128 may include any other type of component, system, algorithm, etc. that utilizes updated occupancy data 124 and / or updated object data 126 to perform operations, tasks, actions, etc.

[0058] For example, the updated occupancy data 124 and / or updated object data 126 can be used by an autonomous or semi-autonomous driving software stack (which can be represented by downstream components 128) to perform one or more operations by a machine (and / or other ego vehicle type). For example, the driving stack can include a world model manager that can be used to generate, update, and / or define a world model. The world model manager can use information generated by and received from the perception components of the driving stack. The perception components can include obstacle sensors, path sensors, waiting sensors, map sensors, and / or other perception components. For example, the world model can be defined at least in part based on the affordances of obstacles, paths, and waiting conditions that can be perceived in real time or near real time by the obstacle sensors, path sensors, waiting sensors, and / or map sensors. The world model manager can continuously update the world model based on newly generated and / or received input (e.g., data) from the obstacle sensors, path sensors, waiting sensors, map sensors, and / or other components of the vehicle. For example, the world model manager and / or perception component may use the updated occupancy data 124 and / or the updated object data 126 to perform one or more operations.

[0059] The world model can be used to help inform the planning, control, obstacle avoidance, and / or actuation components of the driving stack. An obstacle sensor can perform obstacle perception, which can be based on locations where the vehicle is permitted to drive or can drive, and speeds at which the vehicle can drive without colliding with obstacles (e.g., objects such as structures, entities, vehicles, etc.) sensed by the vehicle (e.g., and represented in updated occupancy data 124 and / or updated object data 126).

[0060] The path perception device can perform path perception, such as by sensing a nominal path available in a particular situation. In some examples, the path perception device can further consider lane changes for path perception. The lane map can represent one or more paths available to the vehicle and can be as simple as a single path on a highway entrance ramp. In some examples, the lane map can include a path to a desired lane and / or can indicate available changes on a highway (or other road type), or can include nearby lanes, lane changes, forks, turns, cloverleaf interchanges, merges, and / or other information.

[0061] A waiting sensor can be responsible for determining constraints on the vehicle based on rules, conventions, and / or practical considerations. For example, rules, conventions, and / or practical considerations may relate to traffic lights, multiple stops, yields, merges, toll booths, gates, police or other emergency personnel, road crews, stopped buses or other vehicles, one-way bridge arbitration, ferry entrances, and the like. In some examples, the waiting sensor can be responsible for determining longitudinal constraints on the vehicle, which require the vehicle to wait or slow down until a certain condition is met. In some examples, the waiting condition is caused by potential obstacles, such as intersection traffic at an intersection, which may not be perceived through direct sensing by the obstacle sensor (e.g., using sensor data from a sensor because the obstacle may be obscured by the sensor's field of view). Therefore, the waiting sensor can provide situational awareness by addressing the danger of obstacles that are not always immediately perceptible through rules and conventions that can be sensed and / or learned. Therefore, the waiting sensor can be utilized to identify potential obstacles and implement one or more controls (e.g., slowing down, stopping, etc.) that may not be implemented by relying solely on the obstacle sensor.

[0062] A map sensor may include mechanisms that can identify behaviors and, in some examples, determine specific instances of which conventions apply in a particular region.

[0063] The planning component may include a route planner, a lane planner, a behavior planner, and a behavior selector, among other components, features, and / or functionality. The route planner may use information from a map sensor, a map manager, and / or a positioning manager, among other information, to generate a planned path that may consist of GNSS waypoints (e.g., GPS waypoints). Waypoints may represent specific distances in the future for the vehicle, such as a number of city blocks, a number of kilometers / miles, a number of meters / feet, etc., which may be used as targets for the lane planner.

[0064] The lane planner can use a lane map (e.g., a lane map from a path sensor, which can be generated at least in part using updated occupancy data 124 and / or updated object data 126), object poses within the lane map (e.g., according to a localization manager), and / or a target point and direction at a certain distance in the future from the route planner as input. The target point and direction can be mapped to the best matching drivable point and direction in the lane map (e.g., based on GNSS and / or compass direction). A graph search algorithm can then be executed on the lane map starting from the current edge in the lane map to find the shortest path to the target point.

[0065] The behavior planner can determine the feasibility of basic behaviors of the vehicle, such as staying in the lane or changing lanes to the left or right, so that the feasible behaviors can be matched with the most desired behavior output by the lane planner. For example, if it is determined that the desired behavior is unsafe and / or unavailable, a default behavior can be selected (for example, the default behavior can be to stay in the lane when the desired behavior or lane change is unsafe).

[0066] The control component may follow as closely as possible the trajectory or path (lateral and longitudinal) approach that has been received from the behavior selector of the planning component and that is within the capabilities of the vehicle.

[0067] The obstacle avoidance component can help the vehicle avoid collisions with objects (e.g., moving and stationary objects). In some examples, the obstacle avoidance component can be used independently of the components, features, and / or functions of the vehicle that are required to comply with traffic regulations and drive courteously. In these examples, the obstacle avoidance component can ignore traffic laws, road rules, and courteous driving norms to ensure that no collision occurs between the vehicle and any objects. Therefore, the obstacle avoidance layer can be a layer separate from the road layer rules, and the obstacle avoidance layer can ensure that the vehicle performs safe actions only from the perspective of obstacle avoidance. On the other hand, the road layer rules can ensure that the vehicle complies with traffic laws and conventions, and adheres to legal and regular rights of way.

[0068] In some examples, process 100 may continue to reuse new sensor data 106 generated using sensor 116 to continue determining updated information associated with the environment surrounding the machine. For example, if sensor data 106 is associated with a frame rate, process 100 may repeat based on the frame rate. For a first example, if the frame rate associated with sensor data 106 (and / or other sensor data, sensory data, etc.) is 30 frames per second, process 100 may repeat 30 times per second (e.g., every frame). For a second example, if the frame rate associated with first sensor data 106 (and / or other sensor data, sensory data, etc.) is again 30 frames per second, process 100 may repeat 15 times per second (e.g., every other frame). In some examples, a time instance, as used herein, may correspond to one or more frames at a frame rate. For example, a time instance may correspond to one frame of sensor data, two frames of sensor data, three frames of sensor data, and so on. Furthermore, a time period, as used herein, may correspond to one or more time instances. For example, the time period may correspond to one time instance, two time instances, three time instances, and so on.

[0069] Now refer to Figure 7 , Figure 7is a data flow diagram illustrating an example process 700 for training the fusion component 122 according to some embodiments of the present disclosure. As described above, the fusion component 122 can use one or more classical methods (e.g., algorithmic processes) and / or one or more learning models (e.g., machine learning, deep neural networks, etc.) to generate updated occupancy data 124 and / or updated object data 126. Thus, the process 700 can be used to train one or more learning models of the fusion component 122.

[0070] As shown, the fusion component 122 can be trained using various input data 702 (e.g., training input data), which can include one or more of the sensor data 106, the perception data 108, the fused perception data 110, the occupancy data 112, and / or the object data 114. In some examples, the input data 702 can include one or more actual (e.g., previously generated and / or stored) versions of the sensor data 106, the perception data 108, the fused perception data 110, the occupancy data 112, and / or the object data 114. Additionally or alternatively, the input data 702 can be based on actual versions of the sensor data 106, the perception data 108, the fused perception data 110, the occupancy data 112, and / or the object data 114. For example, the input data 702 can include one or more modified versions of the sensor data 106, the perception data 108, the fused perception data 110, the occupancy data 112, and / or the object data 114.

[0071] The fusion component 122 can be trained using input data 702 and corresponding ground truth data 704. The ground truth data 704 can include annotations, labels, masks, and the like. For example, in some embodiments, the ground truth data 704 can indicate actual values ​​of parameters 706 associated with one or more objects within the environment. For example, for an object, the parameters 706 can include, but are not limited to, x-coordinate position, y-coordinate position, z-coordinate position, height, width, length, x-direction velocity, y-direction velocity, orientation, classification, point position, enclosing shape position, enclosing shape size, object classification, and / or any other parameters. In some examples, the ground truth data 704 can be generated in a drawing program (e.g., an annotation program), a computer-aided design (CAD) program, a labeling program, another type of program suitable for generating ground truth data 704, and / or can be drawn manually. In any examples, ground truth data 704 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., a tagger or annotation expert defines the locations of the labels), and / or a combination thereof (e.g., a human identifies the vertices of a polyline and a machine generates the polygons using a polygon rasterizer).

[0072] The training engine 708 may use one or more loss functions that measure the loss (e.g., error) of the fused output data 710 generated by the fusion component 122 compared to the ground truth data 704. The fused output data 710 may include updated occupancy data 124 and / or updated object data 126. In some examples, any type of loss function may 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 may have different loss functions. For example, the x-coordinate position may include a first loss, the y-coordinate position may include a second loss, the z-coordinate position may include a third loss, and so on. In these examples, the loss functions may be combined to form a total loss, and the total loss may be used to train the fusion component 122 (e.g., to update its parameters). In any example, a backward pass calculation may be performed to recursively calculate the gradients of the loss functions with respect to the training parameters. In some examples, the weights and biases of the fusion component 122 may be used to calculate these gradients.

[0073] Now refer to Figure 8 and Figure 9, each block of methods 800 and 900 described herein includes a computational process that can be performed using any combination of hardware, firmware, and / or software. For example, various functions can be implemented by a processor executing instructions stored in a memory. These methods can also be embodied as computer-usable instructions stored on a computer storage medium. These methods 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 800 and 900 are provided by way of example with respect to Figure 1 However, these methods may additionally or alternatively be performed by any one system or any combination of systems, including but not limited to the systems described herein.

[0074] First reference Figure 8 , Figure 8 8 is a flow chart illustrating an example method 800 for fusing disparate information generated using learned and classical sensor processing techniques according to some embodiments of the present disclosure. At block B802, method 800 may include generating, using one or more neural networks and based at least on first sensor data generated using one or more first sensors of one or more first sensor modalities, first data indicating one or more first locations associated with one or more first objects in an environment. For example, machine learning model 102 may generate, using a neural network and based at least on first sensor data (e.g., sensor data 106) generated using a first sensor of the machine (e.g., sensor 116), first data (e.g., occupancy data 112) indicating a first location associated with a first object in the environment.

[0075] At block B804, method 800 may include generating second data indicating one or more first properties associated with one or more second objects in the environment based at least on second sensor data generated using one or more second sensors of one or more second sensor modalities. For example, processing component 104 may generate second data (e.g., object data 114) indicating first properties associated with second objects in the environment based at least on second sensor data (e.g., sensor data 106) generated using a second sensor of the machine (e.g., sensor 116). In some examples, the second sensor modality may be the same as or different from the first sensor modality, and / or the first sensor may be the same as or different from the second sensor.

[0076] At block B806A, the method 800 may include generating third data including updated versions of the first data indicating one or more second locations associated with the one or more first objects. For example, the fusion component 122 may generate third data (e.g., updated occupancy data 124) including updated versions of the first data (e.g., occupancy data 112) indicating the second locations associated with the first objects.

[0077] Alternatively or additionally, at block B806B, method 800 may include generating third data including an updated version of second data indicating one or more second attributes associated with one or more second objects. For example, fusion component 122 may generate third data (e.g., updated object data 126) including an updated version of second data (e.g., object data 114) indicating second attributes associated with the second objects.

[0078] At block B808, method 800 may include performing one or more operations related to machine control based at least on the third data. For example, downstream component 128 may include one or more components of the machine, such as a planning component for planning a trajectory for the machine, a positioning component for determining the position of the machine relative to its operating environment, a prediction component for predicting the behavior of objects in the environment, and the like. Downstream component 128 may cause the machine to perform one or more operations based at least on the third data. For example, downstream component 128 may cause the machine to follow a specific trajectory through the environment.

[0079] Now refer to Figure 9 , Figure 9 is a flow chart illustrating another example method 900 for fusing information obtained using different processing techniques according to some embodiments of the present disclosure. At block B 902, method 900 may include obtaining, at a time instance, first information associated with one or more first objects in an environment, the first information being generated based on at least first sensor data obtained using one or more first sensors. For example, fusion component 122 may obtain first information (e.g., occupancy data 112) associated with a first object in the environment at the time instance. Additionally, the first information (e.g., occupancy data 112) may be generated based on at least first sensor data (e.g., sensor data 106) obtained from a first sensor (e.g., sensor 116).

[0080] At block B904, method 900 may include obtaining second information associated with one or more second objects in the environment, the second information being determined based at least on second sensor data obtained using one or more second sensors during a time period that at least partially precedes the time instance. For example, fusion component 122 may obtain second information associated with a second object in the environment (e.g., object data 114). Furthermore, the second information (e.g., object data 114) may be determined based at least on second sensor data (e.g., sensor data 106) obtained from a second sensor (e.g., sensor 116) during the time period.

[0081] At block B906, method 900 may include generating an updated version of the first information based at least on the second information, the updated version of the first information indicating at least one or more updated locations associated with the one or more first objects. For example, the fusion component 122 may generate an updated version (e.g., updated occupancy data 124) of the first information (e.g., occupancy data 112) based at least on the second information (e.g., object data 114). Additionally, the updated version (e.g., updated occupancy data 124) of the first information (e.g., occupancy data 112) may indicate at least the updated locations associated with the first objects.

[0082] At block B908, method 900 may include generating an updated version of the second information based at least on the first information, the updated version of the second information indicating at least one or more updated properties associated with the one or more second objects. For example, fusion component 122 may generate an updated version (e.g., updated object data 126) of the second information (e.g., object data 114) based at least on the first information (e.g., occupancy data 112). Furthermore, the updated version (e.g., updated object data 126) of the second information (e.g., object data 114) may indicate at least the updated properties associated with the second objects.

[0083] Example autonomous vehicle

[0084] Figure 10Ais an illustration of an example autonomous vehicle 1000 according to some embodiments of the present disclosure. Autonomous vehicle 1000 (alternatively referred to herein as "vehicle 1000") may include, but is not limited to, a passenger vehicle such as a car, a truck, a bus, a first responder vehicle, a shuttle, an electric or motorized bicycle, a motorcycle, a fire truck, a police vehicle, an ambulance, a boat, a construction vehicle, an underwater vessel, a robotic vehicle, a drone, an airplane, a vehicle coupled to a trailer (e.g., a semi-tractor trailer for hauling cargo), and / or another type of vehicle (e.g., a vehicle that is unmanned and / or accommodates 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 such standards). The vehicle 1000 may be capable of implementing one or more functions consistent with levels 3 through 5 of the autonomous driving levels. The vehicle 1000 may be capable of implementing one or more functions consistent with levels 1 through 5 of the autonomous driving levels. For example, depending on the embodiment, the vehicle 1000 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). The term “autonomy” as used herein may include any and / or all types of autonomy of the vehicle 1000 or other machine, such as fully autonomous, highly autonomous, conditionally autonomous, partially autonomous, with assisted autonomy, semi-autonomous, primarily autonomous, or other designations.

[0085] Vehicle 1000 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 1000 may include a propulsion system 1050, such as an internal combustion engine, a hybrid power plant, an all-electric engine, and / or another type of propulsion system. Propulsion system 1050 may be connected to a drivetrain of vehicle 1000, which may include a transmission, to achieve propulsion of vehicle 1000. Propulsion system 1050 may be controlled in response to receiving a signal from throttle / accelerator 1052.

[0086] A steering system 1054, which may include a steering wheel, may be used to steer vehicle 1000 (e.g., along a desired path or route) when propulsion system 1050 is operating (e.g., while the vehicle is in motion). Steering system 1054 may receive signals from steering actuator 1056. For fully automated (Level 5) functionality, a steering wheel may be optional.

[0087] Brake sensor system 1046 may be used to operate the vehicle brakes in response to receiving signals from brake actuator 1048 and / or brake sensors.

[0088] May include one or more system on chip (SoC) 1004 ( Figure 10C ) and / or one or more controllers 1036 of one or more GPUs can provide flags (e.g., representing commands) to one or more components and / or systems of the vehicle 1000. For example, the one or more controllers can send flags to operate the vehicle brakes via one or more brake actuators 1048, to operate the steering system 1054 via one or more steering actuators 1056, and to operate the propulsion system 1050 via one or more throttles / accelerators 1052. The one or more controllers 1036 can include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process the sensor flags and output operational commands (e.g., flags representing commands) to enable autonomous driving and / or assist a human driver in driving the vehicle 1000. The one or more controllers 1036 may include a first controller 1036 for autonomous driving functions, a second controller 1036 for functional safety functions, a third controller 1036 for artificial intelligence functions (e.g., computer vision), a fourth controller 1036 for infotainment functions, a fifth controller 1036 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 1036 may handle two or more of the above functions, two or more controllers 1036 may handle a single function, and / or any combination thereof.

[0089] The one or more controllers 1036 may provide indicia for controlling one or more components and / or systems of the vehicle 1000 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 1058 (e.g., a global positioning system sensor), a RADAR sensor 1060, an ultrasonic sensor 1062, a LIDAR sensor 1064, an inertial measurement unit (IMU) sensor 1066 (e.g., an accelerometer, a gyroscope, a magnetic compass, a magnetometer, etc.), a microphone 1096, a stereo camera 1068, a wide-angle camera 1070 (e.g., a fisheye camera), an infrared camera 1072, a surround camera 1074 (e.g., a 360-degree camera), a long-range and / or mid-range camera 1098, a speed sensor 1044 (e.g., for measuring the velocity of the vehicle 1000), a vibration sensor 1042, a steering sensor 1040, a brake sensor (e.g., as part of a brake sensor system 1046), and / or other sensor types.

[0090] One or more of the controllers 1036 may receive input (e.g., represented by input data) from the instrument set 1032 of the vehicle 1000 and provide output (e.g., represented by output data, display data, etc.) via a human machine interface (HMI) display 1034, an audible indicator, a speaker, and / or via other components of the vehicle 1000. These outputs may include information such as vehicle speed, velocity, time, map data (e.g., Figure 10C The HMI display 1034 may display information such as a high-definition ("HD") map 1022 of the vehicle 1000, position data (e.g., the position of the vehicle 1000 on the map), directions, positions of other vehicles (e.g., an occupancy grid), information about objects and states of objects as sensed by the controller 1036, and the like. For example, the HMI display 1034 may display information about the presence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information about driving maneuvers that the vehicle has made, is making, or will make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0091] The vehicle 1000 also includes a network interface 1024 that can communicate over one or more networks using one or more wireless antennas 1026 and / or a modem. For example, the network interface 1024 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 1026 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.

[0092] Figure 10B For use according to some embodiments of the present disclosure Figure 10A An example of camera positions and fields of view for autonomous vehicle 1000 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 1000.

[0093] 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 800. 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, etc., 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, clear pixel cameras such as cameras with RCCC, RCCB, and / or RBGC color filter arrays may be used in efforts to improve light sensitivity.

[0094] 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).

[0095] 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 may 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.

[0096] A camera with a field of view that includes a portion of the environment in front of the vehicle 1000 (e.g., a front-facing camera) 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 1036 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.

[0097] 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 1070, which may be used to sense objects entering the field of view from the periphery (e.g., pedestrians, intersection traffic, or bicycles). Although Figure 10B The figure shows only one wide-angle camera, but there can be any number (including zero) of wide-angle cameras 1070 on the vehicle 1000. In addition, any number of long-range cameras 1098 (e.g., a pair of long-view stereo cameras) can be used for depth-based object detection, especially for objects for which neural networks have not yet been trained. Long-range cameras 1098 can also be used for object detection and classification and basic object tracking.

[0098] Any number of stereo cameras 1068 may also be included in the front configuration. In at least one embodiment, one or more stereo cameras 1068 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 1068 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 functions. Other types of stereo cameras 1068 may be used in addition to or alternatively to those described herein.

[0099] Cameras with a field of view that includes portions of the environment to the sides of the vehicle 1000 (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 1074 (e.g., Figure 10B Four surround cameras 1074 (shown in FIG) can be placed on the vehicle 1000. The surround cameras 1074 can include a wide-angle camera 1070, 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 1074 (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.

[0100] A camera having a field of view that includes a portion of the environment behind the vehicle 1000 (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 1098, stereo cameras 1068, infrared cameras 1072, etc.).

[0101] Figure 10C For use according to some embodiments of the present disclosure Figure 10A1000 . 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.

[0102] Figure 10C Each of the components, features, and systems of vehicle 1000 is illustrated as being connected via bus 1002. Bus 1002 may include a controller area network (CAN) data interface (alternatively, referred to herein as a "CAN bus"). CAN may be a network internal to vehicle 1000 that assists in controlling various features and functions of vehicle 1000, such as driving, braking, 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., 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.

[0103] Although bus 1002 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 1002 is represented by a single line, this is not intended to be limiting. For example, there may be any number of buses 1002, 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 1002 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 1002 may be used for collision avoidance functionality, and a second bus 1002 may be used for driving control. In any example, each bus 1002 may communicate with any component of vehicle 1000, and two or more buses 1002 may communicate with the same component. In some examples, each SoC 1004 , each controller 1036 , and / or each computer within the vehicle may have access to the same input data (e.g., input from sensors of the vehicle 1000 ) and may be connected to a common bus such as a CAN bus.

[0104] The vehicle 1000 may include one or more controllers 1036, such as those described herein. Figure 10A Controller 1036 may be used for a variety of functions. Controller 1036 may be coupled to any other various components and systems of vehicle 1000 and may be used for control of vehicle 1000, artificial intelligence of vehicle 1000, infotainment for vehicle 1000, and / or the like.

[0105] The vehicle 1000 may include one or more system-on-chips (SoCs) 1004. The SoCs 1004 may include a CPU 1006, a GPU 1008, a processor 1010, a cache 1012, an accelerator 1014, a data store 1016, and / or other components and features not shown. The SoCs 1004 may be used to control the vehicle 1000 in a variety of platforms and systems. For example, the one or more SoCs 1004 may be combined with an HD map 1022 in a system (e.g., a system of the vehicle 1000), which may be downloaded from one or more servers (e.g., a server) via a network interface 1024. Figure 10D One or more servers 1078) obtain map refreshes and / or updates.

[0106] The CPU 1006 may include a CPU cluster or CPU complex (alternatively, referred to herein as a "CCPLEX"). The CPU 1006 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU 1006 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 1006 may include four dual-core clusters, each of which has a dedicated L2 cache (e.g., a 2MB L2 cache). The CPU 1006 (e.g., CCPLEX) may be configured to support simultaneous cluster operations such that any combination of CPU 1006 clusters can be active at any given time.

[0107] CPU 1006 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 1006 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.

[0108] The GPU 1008 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The GPU 1008 may be programmable and efficient for parallel workloads. In some examples, the GPU 1008 may use an enhanced tensor instruction set. The GPU 1008 may include one or more streaming microprocessors, each of which may include an L1 cache (e.g., an L1 cache with at least 96KB of storage capacity), and two or more of these streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB of storage capacity). In some embodiments, the GPU 1008 may include at least eight streaming microprocessors. The GPU 1008 may use a computing application programming interface (API). In addition, the GPU 1008 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0109] In the case of automotive and embedded use, GPU 1008 can be power optimized to achieve optimal performance. For example, GPU 1008 can be manufactured on fin field effect transistors (FinFETs). However, this is not intended to be limiting, and GPU 1008 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.

[0110] GPU 1008 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.

[0111] The GPU 1008 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 1008 to directly access the CPU 1006 page tables. In such an example, when the GPU 1008 memory management unit (MMU) experiences a miss, an address translation request may be transmitted to the CPU 1006. In response, the CPU 1006 may look up the virtual-to-physical mapping for the address in its page table and transmit the translation back to the GPU 1008. In this way, unified memory technology may allow a single unified virtual address space to be used for memory of both the CPU 1006 and the GPU 1008, thereby simplifying GPU 1008 programming and porting applications to the GPU 1008.

[0112] Additionally, GPU 1008 may include access counters that can track how often GPU 1008 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.

[0113] SoC 1004 may include any number of caches 1012, including those described herein. For example, cache 1012 may include an L3 cache available to both CPU 1006 and GPU 1008 (e.g., connected to both CPU 1006 and GPU 1008). Cache 1012 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.

[0114] The SoC 1004 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 1000, such as processing a DNN. Furthermore, the SoC 1004 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 1004 may include one or more FPUs integrated as execution units within the CPU 1006 and / or GPU 1008.

[0115] SoC 1004 may include one or more accelerators 1014 (e.g., hardware accelerators, software accelerators, or a combination thereof). For example, SoC 1004 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 secondary GPU 1008 and offload some tasks of GPU 1008 (e.g., freeing up more cycles of GPU 1008 for performing other tasks). As an example, accelerator 1014 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).

[0116] Accelerator 1014 (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.

[0117] 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.

[0118] The DLA can perform any function of the GPU 1008, and by using an inference accelerator, for example, the designer can target any function to either the DLA or the GPU 1008. For example, the designer can focus the processing of CNNs and floating-point operations on the DLA and leave other functions to the GPU 1008 and / or other accelerators 1014.

[0119] The accelerator 1014 (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.

[0120] The RISC core can interact with an image sensor (e.g., an image sensor of any camera described herein), an image signature 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.

[0121] The DMA can enable components of the PVA to access system memory independently of the CPU 1006. 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.

[0122] A vector processor can be a programmable processor that can be designed to efficiently and flexibly execute programming for computer vision algorithms and provide logo 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 logo processor, such as, for example, a single instruction multiple data (SIMD), a very long instruction word (VLIW) digital logo processor. The combination of SIMD and VLIW can enhance throughput and speed.

[0123] 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.

[0124] The accelerator 1014 (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 1014. 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).

[0125] The on-chip computer vision network can include an interface that ensures that both the PVA and DLA provide ready and valid flags before transmitting any control flags / addresses / data. Such an interface can provide separate phases and separate channels for transmitting control flags / addresses / data, as well as burst-based communication for continuous data transmission. This type of interface can comply with ISO 26262 or IEC 615010 standards, but other standards and protocols can also be used.

[0126] In some examples, SoC 1004 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 visual simulations for RADAR signature interpretation, for sound propagation synthesis and / or analysis, for SONAR system simulation, for general wave propagation simulation, 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.

[0127] The accelerator 1014 (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.

[0128] 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.

[0129] 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.

[0130] 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 the confidence level 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 as 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 1066 output related to the orientation and distance of the vehicle 1000, 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LIDAR sensor 1064 or RADAR sensor 1060), etc.

[0131] SoC 1004 may include one or more data stores 1016 (e.g., memory). Data store 1016 may be on-chip memory of SoC 1004 that may store neural networks to be executed on the GPU and / or DLA. In some examples, for redundancy and safety, data store 1016 may be large enough to store multiple instances of the neural network. Data store 1012 may include an L2 or L3 cache 1012. References to data store 1016 may include references to memory associated with the PVA, DLA, and / or other accelerators 1014 as described herein.

[0132] SoC 1004 may include one or more processors 1010 (e.g., embedded processors). Processor 1010 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 1004 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 1004 thermal and temperature sensor management, and / or SoC 1004 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and SoC 1004 may use the ring oscillator to detect the temperature of CPU 1006, GPU 1008, and / or accelerator 1014. If the temperature is determined to exceed a threshold, the boot and power management processor may enter a temperature fault routine and place SoC 1004 in a lower power state and / or place vehicle 1000 in a driver safety parking mode (e.g., to safely park vehicle 1000).

[0133] The processor 1010 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 logo processor with dedicated RAM.

[0134] Processor 1010 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.

[0135] Processor 1010 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.

[0136] Processor 1010 may also include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0137] The processor 1010 may also include a high dynamic range logo processor, which may include an image logo processor, which is a hardware engine that is part of the camera processing pipeline.

[0138] The processor 1010 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 1070, the surround camera 1074, 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.

[0139] 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.

[0140] 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 1008 does not need to continuously render new surfaces. Even when the GPU 1008 is powered on and active for 3D rendering, the video image compositor can be used to offload the GPU 1008 to improve performance and responsiveness.

[0141] SoC 1004 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. SoC 1004 may also include an input / output controller that may be controlled by software and may be used to receive I / O flags that are not assigned to a specific role.

[0142] The SoC 1004 may also include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management, and / or other devices. The SoC 1004 may be used to process data from cameras (connected via Gigabit multimedia serial links and Ethernet), sensors (e.g., LIDAR sensor 1064, RADAR sensor 1060, etc., which may be connected via Ethernet), data from the bus 1002 (e.g., vehicle 1000 speed, steering wheel position, etc.), and data from the GNSS sensor 1058 (connected via Ethernet or a CAN bus). The SoC 1004 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 1006 from routine data management tasks.

[0143] SoC 1004 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 1004 can be faster, more reliable, and even more energy- and space-efficient than conventional systems. For example, when combined with CPU 1006, GPU 1008, and data storage 1016, accelerator 1014 can provide a fast and efficient platform for Level 3-5 autonomous vehicles.

[0144] This technology thus provides capabilities and functionality not achievable 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.

[0145] In contrast to conventional systems, by providing a CPU complex, a GPU complex, and a cluster of hardware accelerators, the technology described herein allows for 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 1020) 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.

[0146] 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 1008.

[0147] 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 1000. 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 1004 provides security against theft and / or carjacking.

[0148] In another example, a CNN for emergency vehicle detection and identification can use data from microphone 1096 to detect and identify emergency vehicle sirens. In contrast to conventional systems that use general classifiers to detect sirens and manually extract features, SoC 1004 uses CNNs to classify environmental and urban sounds as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative 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 1058. Thus, for example, when operating in the European Union, the CNN will seek to detect EU 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 1062, 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.

[0149] The vehicle may include a CPU 1018 (e.g., a discrete CPU or dCPU) that may be coupled to the SoC 1004 via a high-speed interconnect (e.g., PCIe). The CPU 1018 may include, for example, an X106 processor. The CPU 1018 may be used to perform any of a variety of functions, including, for example, arbitrating potentially inconsistent results between ADAS sensors and the SoC 1004, and / or monitoring the status and health of the controller 1036 and / or the infotainment SoC 1030.

[0150] The vehicle 1000 may include a GPU 1020 (e.g., a discrete GPU or dGPU) that may be coupled to the SoC 1004 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU 1020 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 1000 (e.g., sensor data).

[0151] The vehicle 1000 may also include a network interface 1024, which may include one or more wireless antennas 1026 (e.g., one or more wireless antennas for different communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 1024 can be used to enable wireless connections to the cloud (e.g., to a server 1078 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 1000 with information about vehicles approaching the vehicle 1000 (e.g., vehicles in front of, to the side of, and / or behind the vehicle 1000). This functionality can be part of the cooperative adaptive cruise control functionality of the vehicle 1000.

[0152] The network interface 1024 may include a SoC that provides modulation and demodulation functionality and enables the controller 1036 to communicate over a wireless network. The network interface 1024 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.

[0153] The vehicle 1000 may also include data storage 1028, which may include off-chip storage (e.g., outside the SoC 1004). The data storage 1028 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.

[0154] The vehicle 1000 may also include a GNSS sensor 1058. The GNSS sensor 1058 (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 1058 may be used, including, for example and without limitation, GPS using a USB connector with an Ethernet to serial (RS-232) bridge.

[0155] The vehicle 1000 may also include a RADAR sensor 1060. The RADAR sensor 1060 can be used by the vehicle 1000 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 1060 can use CAN and / or bus 1002 (e.g., to transmit data generated by the RADAR sensor 1060) 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 1060 can be suitable for front, rear, and side RADAR use. In some examples, a pulse Doppler RADAR sensor is used.

[0156] The RADAR sensor 1060 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 250m) achieved by two or more independent scans. The RADAR sensor 1060 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 1000 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 1000.

[0157] As an example, a medium-range RADAR system may include a range of up to 1060m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 1050 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.

[0158] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0159] Vehicle 1000 may also include ultrasonic sensors 1062. Ultrasonic sensors 1062, which may be located on the front, rear, and / or sides of vehicle 1000, may be used for parking assistance and / or for creating and updating an occupancy grid. A variety of ultrasonic sensors 1062 may be used, and different ultrasonic sensors 1062 may have different detection ranges (e.g., 2.5 m, 4 m). Ultrasonic sensors 1062 may operate at functional safety level ASIL B.

[0160] Vehicle 1000 may include a LIDAR sensor 1064. LIDAR sensor 1064 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. LIDAR sensor 1064 may be ASIL B functional safety level. In some examples, vehicle 1000 may include multiple LIDAR sensors 1064 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0161] In some examples, the LIDAR sensor 1064 may be capable of providing a list of objects and their distances for a 360-degree field of view. Commercially available LIDAR sensors 1064 may have, for example, an advertised range of approximately 1000 meters, an accuracy of 2-3 cm, and support for 1000 Mbps Ethernet connections. In some examples, one or more non-obtrusive LIDAR sensors 1064 may be used. In such examples, the LIDAR sensor 1064 may be implemented as a small device that can be embedded in the front, back, sides, and / or corners of the vehicle 1000. In such examples, the LIDAR sensor 1064 may provide a field of view of up to 120 degrees horizontally and 35 degrees vertically, even for low-reflectivity objects, with a range of 200 meters. The front-mounted LIDAR sensor 1064 may be configured for a horizontal field of view between 45 and 135 degrees.

[0162] In some examples, LIDAR technology such as 3D flash LIDAR can also be used. 3D flash LIDAR uses flashes of laser light as an emission source to illuminate the vehicle's surroundings up to about 200 meters away. The flash LIDAR unit includes a receiver that records the laser pulse transmission time and the 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 1000. 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 fans. 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 1064 may be less susceptible to motion blur, vibration, and / or shock.

[0163] The vehicle may also include an IMU sensor 1066. In some examples, the IMU sensor 1066 may be located at the center of the rear axle of the vehicle 1000. The IMU sensor 1066 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 1066 may include an accelerometer and a gyroscope, while in a nine-axis application, the IMU sensor 1066 may include an accelerometer, a gyroscope, and a magnetometer.

[0164] In some embodiments, the IMU sensor 1066 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 1066 can enable the vehicle 1000 to estimate heading without input from a magnetic sensor by directly observing and correlating velocity changes from the GPS to the IMU sensor 1066. In some examples, the IMU sensor 1066 and the GNSS sensor 1058 can be combined into a single integrated unit.

[0165] The vehicle may include microphones 1096 positioned in and / or around the vehicle 1000. The microphones 1096 may be used for, among other things, emergency vehicle detection and identification.

[0166] The vehicle may also include any number of camera types, including stereo cameras 1068, wide angle cameras 1070, infrared cameras 1072, surround cameras 1074, long and / or medium range cameras 1098, and / or other camera types. These cameras may be used to capture image data around the entire periphery of the vehicle 1000. The type of camera used depends on the embodiment and the requirements of the vehicle 1000, and any combination of camera types may be used to provide the necessary coverage around the vehicle 1000. 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 10GbE GPIO (Gigabit Multimedia Serial Link) and / or a 1GbE GPIO (Gigabit Ethernet). Figure 10A and Figure 10B Described in more detail.

[0167] Vehicle 1000 may also include a vibration sensor 1042. Vibration sensor 1042 can measure the vibration of a component of the vehicle, such as an axle. For example, changes in vibration can indicate changes in the road surface. In another example, when two or more vibration sensors 1042 are used, the difference between the vibrations can be used to determine friction or slip on the road surface (e.g., when there is a vibration difference between a powered axle and a free-wheeling axle).

[0168] The vehicle 1000 may include an ADAS system 1038. In some examples, the ADAS system 1038 may include a SoC. The ADAS system 1038 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.

[0169] The ACC system can utilize RADAR sensors 1060, LIDAR sensors 1064, 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 1000 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 1000 change lanes. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0170] 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 1024 and / or the wireless antenna 1026. 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 1000 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 1000, CACC can be more reliable, and it has the potential to improve the smoothness of traffic flow and reduce road congestion.

[0171] 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 1060 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components. The FCW system can provide warnings in the form of, for example, audible, visual warnings, vibrations, and / or rapid brake pulses.

[0172] An 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. An AEB system can use a front-facing camera and / or RADAR sensor 1060 coupled to a dedicated processor, DSP, FPGA, and / or ASIC. When an 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. An AEB system can include technologies such as dynamic brake support and / or collision approach braking.

[0173] The LDW system provides visual, audible, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle 1000 crosses a lane marking. When the driver indicates an intention to leave the lane by activating a turn sign, 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 vibrating component.

[0174] The LKA system is a variation of the LDW system. If the vehicle 1000 begins to leave its lane, the LKA system provides steering input or braking to correct the vehicle 1000.

[0175] 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 sign. The BSW system can use a rear-facing camera and / or RADAR sensor 1060 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibrating component.

[0176] 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 1000 is in reverse. Some RCTW systems include 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 1060 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.

[0177] 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 act accordingly. However, in the autonomous vehicle 1000, in the event of conflicting results, the vehicle 1000 itself must decide whether to heed the results from the primary computer or the auxiliary computer (e.g., the first controller 1036 or the second controller 1036). For example, in some embodiments, the ADAS system 1038 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 1038 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.

[0178] 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.

[0179] 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 grid 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 preferred embodiments, the supervisory MCU can include and / or be included as a component of SoC 1004.

[0180] In other examples, the ADAS system 1038 can 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.

[0181] In some examples, the output of the ADAS system 1038 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 1038 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.

[0182] The vehicle 1000 may also include an infotainment SoC 1030 (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 1030 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 1000. For example, the infotainment SoC 1030 may include a radio, a disc player, a navigation system, a video player, USB and Bluetooth connectivity, an onboard computer, in-vehicle entertainment, WiFi, steering wheel audio controls, hands-free voice controls, a head-up display (HUD), an HMI display 1034, 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 1030 may further be used to provide information (e.g., visual and / or auditory) to a user of the vehicle, such as information from the ADAS system 1038, 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.

[0183] The infotainment SoC 1030 may include GPU functionality. The infotainment SoC 1030 may communicate with other devices, systems, and / or components of the vehicle 1000 via a bus 1002 (e.g., a CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 1030 may be coupled to a supervisory MCU so that in the event of a failure of a primary controller 1036 (e.g., a primary and / or backup computer of the vehicle 1000), the infotainment system's GPU may perform some self-driving functions. In such an example, the infotainment SoC 1030 may place the vehicle 1000 in a driver-safe parking mode as described herein.

[0184] Vehicle 1000 may also include a dashboard 1032 (e.g., a digital dashboard, an electronic dashboard, a digital dashboard, etc.). The dashboard 1032 may include a controller and / or a supercomputer (e.g., a separate controller or a supercomputer). The dashboard 1032 may include a set of instruments, such as a speedometer, fuel level, oil pressure, a tachometer, an odometer, a turn indicator, a shift 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 may be displayed and / or shared between the infotainment SoC 1030 and the dashboard 1032. In other words, the dashboard 1032 may be included as part of the infotainment SoC 1030, or vice versa.

[0185] Figure 10D For cloud-based servers and Figure 10A 10. System diagram of communication between an example autonomous vehicle 1000. System 1076 may include a server 1078, a network 1090, and a vehicle including vehicle 1000. Server 1078 may include multiple GPUs 1084(A)-1084(H) (collectively referred to herein as GPUs 1084), PCIe switches 1082(A)-1082(H) (collectively referred to herein as PCIe switches 1082), and / or CPUs 1080(A)-1080(B) (collectively referred to herein as CPUs 1080). GPUs 1084, CPUs 1080, and PCIe switches may be interconnected with a high-speed interconnect such as, for example and without limitation, the NVLink interface 1088 developed by NVIDIA and / or PCIe connections 1086. In some examples, GPUs 1084 are connected via NVLink and / or NVSwitch SoCs, and GPUs 1084 and PCIe switches 1082 are connected via PCIe interconnects. Although eight GPUs 1084, two CPUs 1080, and two PCIe switches are shown, this is not intended to be limiting. Depending on the embodiment, each of the servers 1078 may include any number of GPUs 1084, CPUs 1080, and / or PCIe switches. For example, each of the servers 1078 may include eight, sixteen, thirty-two, and / or more GPUs 1084.

[0186] Server 1078 can receive image data from a vehicle via network 1090, the image data representing images showing unexpected or changed road conditions, such as recently begun road construction. Server 1078 can transmit neural network 1092, updated neural network 1092, and / or map information 1094, including information about traffic and road conditions, to the vehicle via network 1090. Updates to map information 1094 can include updates to HD map 1022, such as information about construction sites, potholes, curves, flooding, or other obstacles. In some examples, neural network 1092, updated neural network 1092, and / or map information 1094 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 1078 and / or other servers).

[0187] Server 1078 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 1090), and / or the machine learning model can be used by server 1078 to remotely monitor the vehicle.

[0188] In some examples, server 1078 can receive data from the vehicle and apply the data to the latest real-time neural network for real-time intelligent reasoning. Server 1078 can include a deep learning supercomputer and / or a dedicated AI computer powered by GPU 1084, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 1078 can include the deep learning infrastructure of a data center using only CPU power.

[0189] The deep learning infrastructure of server 1078 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 1000. For example, the deep learning infrastructure may receive periodic updates from vehicle 1000, such as an image sequence and / or objects located in the image sequence that vehicle 1000 has located (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural network to identify objects and compare them to the objects identified by vehicle 1000, and if the results do not match and the infrastructure concludes that the AI ​​in vehicle 1000 has malfunctioned, server 1078 may transmit a flag to vehicle 1000 instructing the fail-safe computer of vehicle 1000 to take control, notify passengers, and complete a safe parking maneuver.

[0190] For inference, the server 1078 may include a GPU 1084 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.

[0191] Example computing device

[0192] Figure 11 1 is a block diagram of an example computing device 1100 suitable for implementing some embodiments of the present disclosure. Computing device 1100 may include an interconnect system 1102 that directly or indirectly couples the following devices: memory 1104, one or more central processing units (CPUs) 1106, one or more graphics processing units (GPUs) 1108, a communication interface 1110, input / output (I / O) ports 1112, I / O components 1114, a power supply 1116, one or more presentation components 1118 (e.g., one or more displays), and one or more logic units 1120. In at least one embodiment, one or more computing devices 1100 may include one or more virtual machines (VMs), and / or any of their components may include virtual components (e.g., virtual hardware components). For non-limiting examples, one or more of GPUs 1108 may include one or more vGPUs, one or more of CPUs 1106 may include one or more vCPUs, and / or one or more of logic units 1120 may include one or more virtual logic units. As such, one or more computing devices 1100 may include discrete components (eg, a full GPU dedicated to computing device 1100 ), virtual components (eg, a portion of a GPU dedicated to computing device 1100 ), or a combination thereof.

[0193] although Figure 11 The various blocks of are shown as being connected via interconnect system 1102 using wires, but this is not intended to be limiting and is provided for clarity only. For example, in some embodiments, presentation component 1118 (such as a display device) may be considered to be I / O component 1114 (e.g., if the display is a touch screen). As another example, CPU 1106 and / or GPU 1108 may include memory (e.g., memory 1104 may represent a storage device in addition to the memory of GPU 1108, CPU 1106, and / or other components). In other words, Figure 11 The term computing device is illustrative only. No distinction is made between categories within this category 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 this category. Figure 11 within the range of computing devices.

[0194] Interconnect system 1102 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 1102 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 1106 can be directly connected to memory 1104. Further, CPU 1106 can be directly connected to GPU 1108. In the case where there is a direct or point-to-point connection between components, interconnect system 1102 can include a PCIe link to perform the connection. In these examples, the PCI bus does not need to be included in computing device 1100.

[0195] Memory 1104 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 1100. 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.

[0196] 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 1104 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 the computing device 1100. As used herein, computer storage media does not include the identifier itself.

[0197] Computer storage media may embody computer-readable instructions, data structures, program modules, and / or other data types in modulated data symbols, such as a carrier wave or other transport mechanism, and includes any information delivery media. The term "modulated data symbol" may refer to a symbol that has one or more characteristics set or changed in a manner that encodes information in the symbol. 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.

[0198] The CPU 1106 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 1100 to perform one or more of the methods and / or processes described herein. The CPUs 1106 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 1106 may include any type of processor and may include different types of processors depending on the type of computing device 1100 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 1100, 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 1100 may also include one or more CPUs 1106 in addition to one or more microprocessors or secondary coprocessors (such as a math coprocessor).

[0199] In addition to or in lieu of one or more CPUs 1106, one or more GPUs 1108 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 1100 to perform one or more of the methods and / or processes described herein. One or more of the GPUs 1108 may be integrated GPUs (e.g., with one or more of the CPUs 1106) and / or one or more of the GPUs 1108 may be discrete GPUs. In embodiments, one or more of the GPUs 1108 may serve as coprocessors for one or more of the CPUs 1106. The GPUs 1108 may be used by the computing device 1100 to render graphics (e.g., 3D graphics) or perform general-purpose computations. For example, the GPUs 1108 may be used for general-purpose computing on a GPU (GPGPU). The GPUs 1108 may include hundreds or thousands of cores capable of handling hundreds or thousands of software threads simultaneously. The GPUs 1108 may generate pixel data for an output image in response to rendering commands (e.g., rendering commands received from the CPUs 1106 via a host interface). The GPU 1108 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 1104. The GPU 1108 may include two or more GPUs operating in parallel (e.g., via a link). The link may connect the GPUs directly (e.g., using NVLINK) or may connect the GPUs through a switch (e.g., using NVSwitch). When combined, each GPU 1108 may generate pixel data or GPGPU data for a different portion 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.

[0200] In addition to or in lieu of the CPU 1106 and / or GPU 1108, the logic unit 1120 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 1100 to perform one or more of the methods and / or processes described herein. In embodiments, the one or more CPUs 1106, the one or more GPUs 1108, and / or the one or more logic units 1120 may perform any combination of methods, processes, and / or portions thereof, either discretely or jointly. One or more of the logic units 1120 may be part of and / or integrated into one or more of the CPUs 1106 and / or GPUs 1108, and / or one or more of the logic units 1120 may be discrete components or otherwise external to the CPU 1106 and / or GPU 1108. In embodiments, one or more of logic units 1120 may be a co-processor to one or more of CPUs 1106 and / or one or more of GPUs 1108 .

[0201] Examples of logic unit 1120 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.

[0202] The communication interface 1110 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 1100 to communicate with other computing devices via an electronic communication network (including wired and / or wireless communications). The communication interface 1110 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 1120 and / or the communication interface 1110 may include one or more data processing units (DPUs) to send data received over the network and / or over the interconnect system 1102 directly to one or more GPUs 1108 (e.g., memory).

[0203] The I / O ports 1112 can enable the computing device 1100 to be logically coupled to other devices including I / O components 1114, one or more presentation components 1118, and / or other components, some of which can be built into (e.g., integrated into) the computing device 1100. Illustrative I / O components 1114 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. The I / O components 1114 can provide a natural user interface (NUI) that processes air gestures, voice, or other physiological input generated by the 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 the computing device 1100. The computing device 1100 can 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 1100 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 1100 may use the output of the accelerometer or gyroscope to render immersive augmented or virtual reality.

[0204] The power supply 1116 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 1116 may provide power to the computing device 1100 to enable the components of the computing device 1100 to operate.

[0205] The presentation component 1118 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 1118 may receive data from other components (e.g., the GPU 1108, the CPU 1106, etc.) and output the data (e.g., as images, video, sound, etc.).

[0206] Sample Data Center

[0207] Figure 12 An example data center 1200 that can be used in at least one embodiment of the present disclosure is shown. The data center 1200 can include a data center infrastructure layer 1210, a framework layer 1220, a software layer 1230, and / or an application layer 1240.

[0208] like Figure 12 As shown, the data center infrastructure layer 1210 may include a resource coordinator 1210, grouped computing resources 1214, and node computing resources ("node CRs") 1216(1)-1216(N), where "N" represents any complete positive integer. In at least one embodiment, the node CRs 1216(1)-1216(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 of the node CRs 1216(1)-1216(N) may correspond to a server having one or more of the above-mentioned computing resources. Furthermore, in some embodiments, node CRs 1216(1)-12161(N) may include one or more virtual components, such as vGPUs, vCPUs, etc., and / or one or more of node CRs 1216(1)-1216(N) may correspond to virtual machines (VMs).

[0209] In at least one embodiment, the grouped computing resources 1214 may include individual groups of node CRs 1216 housed in one or more racks (not shown), or multiple racks housed in data centers at different geographical locations (also not shown). Individual groups of node CRs 1216 within the grouped computing resources 1214 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 1216 including CPUs, GPUs, 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.

[0210] Resource coordinator 1222 may configure or otherwise control one or more node CRs 1216(1)-1216(N) and / or grouped computing resources 1214. In at least one embodiment, resource coordinator 1222 may comprise a software design infrastructure ("SDI") management entity for data center 1200. Resource coordinator 1222 may comprise hardware, software, or some combination thereof.

[0211] In at least one embodiment, Figure 12 As shown, the framework layer 1220 may include a job scheduler 1232, a configuration manager 1234, a resource manager 1236, and / or a distributed file system 1238. The framework layer 1220 may include a framework that supports the software 1232 of the software layer 1230 and / or one or more applications 1242 of the application layer 1240. The software 1232 or the application 1242 may include network-based service software or applications, such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. The framework layer 1220 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 1238 for large-scale data processing (e.g., "big data"). TM(hereinafter referred to as "Spark"). In at least one embodiment, the job scheduler 1232 may include a Spark driver to facilitate scheduling workloads supported by the different layers of the data center 1200. The configuration manager 1234 may be capable of configuring the different layers, such as the software layer 1230 and the framework layer 1220 (which includes Spark and a distributed file system 1238 for supporting large-scale data processing). The resource manager 1236 may be capable of managing clustered or grouped computing resources that are mapped to the distributed file system 1238 and the job scheduler 1232 or allocated to support the distributed file system 1238 and the job scheduler 1232. In at least one embodiment, the clustered or grouped computing resources may include the grouped computing resources 1214 at the data center infrastructure layer 1210. The resource manager 1236 may coordinate with the resource coordinator 1210 to manage these mapped or allocated computing resources.

[0212] In at least one embodiment, the software 1232 included in the software layer 1230 may include software used by at least a portion of the node CRs 1216(1)-1216(N), the grouped computing resources 1214, and / or the distributed file system 1238 of the framework layer 1220. 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.

[0213] In at least one embodiment, the applications 1242 included in the application layer 1240 may include one or more types of applications used by at least a portion of the node CRs 1216(1)-1216(N), the grouped computing resources 1214, and / or the distributed file system 1238 of the framework layer 1220. 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.

[0214] In at least one embodiment, any of configuration manager 1234, resource manager 1236, and resource coordinator 1210 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 1200 from making potentially poor configuration decisions and potentially avoiding underutilized and / or poorly performing portions of the data center.

[0215] According to one or more embodiments described herein, data center 1200 may include tools, services, software, or other resources to train one or more machine learning models or to predict or infer information using one or more machine learning models. For example, one or more machine learning models 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 1200. 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 1200 using weight parameters computed using one or more training techniques, such as, but not limited to, those described herein.

[0216] In at least one embodiment, data center 1200 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.

[0217] Sample network environment

[0218] 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 11 The system may be implemented on one or more instances of one or more computing devices 1100 - for example, each device may include similar components, features and / or functions of one or more computing devices 1100. In addition, in the case of implementing a backend device (e.g., a server, NAS, etc.), the backend device may be included as part of the data center 1200, an example of which is described herein with respect to the backend device 1200. Figure 12 Describe in more detail.

[0219] 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 a plurality of networks or one of more 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.

[0220] 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.

[0221] 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").

[0222] 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).

[0223] One or more client devices may include Figure 11At least some of the components, features, and functionality of one or more example computing devices 1100 are described. By way of example and not limitation, a 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.

[0224] The present disclosure may be described in the general context of machine-usable instructions or computer code, including computer-executable instructions, such as program modules, executed by a computer or other machine such as a personal digital assistant or other handheld device. Generally, program modules, including routines, programs, objects, components, data structures, and the like, refer to code that performs a specific task or implements a specific abstract data type. The present disclosure may be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, and the like. The present disclosure may also be practiced in distributed computing environments in which tasks are performed by remote processing devices linked via a communications network.

[0225] 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.

[0226] 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.

[0227] Example paragraph

[0228] A. A method comprising: generating, using one or more neural networks and based at least on first sensor data generated using one or more first sensors of one or more first sensor modalities, first data indicating one or more first locations associated with one or more first objects in an environment; generating, based at least on second sensor data generated using one or more second sensors of one or more second sensor modalities different from the one or more first sensor modalities, second data indicating one or more first attributes associated with one or more second objects in the environment; generating, based at least on at least a first portion of the first data and at least a second portion of the second data, third data comprising at least one of: an updated version of the first data indicating one or more second locations associated with the one or more first objects; or an updated version of the second data indicating one or more second attributes associated with the one or more second objects; and performing one or more operations regarding control of a machine based at least on the third data.

[0229] B. A method as described in paragraph A, wherein: the first data is instantaneous data indicating the one or more first locations associated with the one or more first objects in the environment around the machine at a time instance, and the second data is time data, and the one or more first attributes are tracked over a time period that at least partially precedes the time instance.

[0230] C. A method as described in any of paragraphs AB, wherein: the first data further indicates one or more obscured parts of the environment, and the updated version of the first data further indicates whether the one or more obscured parts of the environment are occupied by at least one first object among the one or more first objects or at least one second object among the one or more second objects.

[0231] D. A method as described in any of paragraphs AC, wherein: the attributes of the one or more first attributes include at least one of the following: a first enclosing shape associated with an object of the one or more second objects, a first position associated with the object, a first posture associated with the object, a first trajectory associated with the object, or a first classification associated with the object, and the attributes of the one or more second attributes include at least one of the following: a second enclosing shape associated with the object, a second position associated with the object, a second posture associated with the object, a second trajectory associated with the object, or a second classification associated with the object.

[0232] E. A method as described in any of paragraphs AD, wherein the first data is a dense occupancy representation of the environment from a top-down perspective, the dense occupancy representation including one or more points representing one or more samples obtained using the one or more first sensors at a time instance, wherein one or more first points of the one or more points correspond to the one or more first locations associated with the one or more first objects, and one or more second points of the one or more points correspond to one or more unoccupied locations in the environment at the time instance.

[0233] F. The method of any of paragraphs AE, wherein the one or more values ​​of the one or more points correspond to at least one of a height or a confidence associated with the one or more samples.

[0234] G. The method of any of paragraphs AF, further comprising: determining that a first object of the one or more first objects corresponds to a second object of the one or more second objects, wherein generating the third data is further based at least on the first object corresponding to the second object.

[0235] H. The method as described in any of paragraphs AG further includes: generating fourth data indicating one or more previous locations associated with the one or more first objects in the environment, the fourth data including one or more points representing one or more previous samples obtained using the one or more first sensors within a time period and refined at least based on the one or more first attributes; wherein generating the third data is further based at least on the fourth data.

[0236] I. The method of any of paragraphs AH, wherein a first point of the one or more points included in the fourth data indicates a speed associated with an object of the one or more first objects.

[0237] J. The method of any of paragraphs AI, further comprising causing the machine to perform one or more operations based on at least one of the updated version of the first data or the updated version of the second data.

[0238] K. A system comprising: one or more processors, the one or more processors being configured to: obtain first information indicating one or more locations associated with one or more first objects in an environment at a time instance, the first information being generated using one or more neural networks and based at least on first sensor data obtained using one or more first sensors of a machine; obtain second information indicating one or more attributes associated with one or more second objects in the environment, the second information being determined at least based on second sensor data obtained using one or more second sensors of the machine during a time period at least partially preceding the time instance; and generate at least one of: an updated version of the first information based at least on the second information, the updated version of the first information at least indicating one or more updated locations associated with the one or more first objects; or an updated version of the second information based at least on the first information, the updated version of the second information at least indicating one or more updated attributes associated with the one or more second objects.

[0239] L. The system of paragraph K, wherein the first information is associated with the time instance, and the updated version of the first information is time information associated with the time instance and at least a portion of the time period.

[0240] M. A system as described in any of paragraphs KL, wherein: the first information indicates one or more occluded portions of the environment at the time instance, and the updated version of the first information further indicates whether the one or more occluded portions of the environment are occupied by at least one first object of the one or more first objects or at least one second object of the one or more second objects.

[0241] N. A system as described in any of paragraphs KM, wherein: the attributes of the one or more updated attributes include at least one of the following: an enclosing shape associated with an object of the one or more second objects, a position associated with the object, a posture associated with the object, a trajectory associated with the object, or a classification associated with the object.

[0242] O. A system as described in any of paragraphs KN, wherein: the first information is a dense occupancy representation of the environment from a top-down perspective, the dense occupancy representation including one or more points representing one or more samples obtained using the one or more first sensors at the time instance, one or more first points in the one or more points corresponding to the one or more locations associated with the one or more first objects, one or more second points in the one or more points corresponding to one or more unoccupied locations in the environment at the time instance, and one or more values ​​in the one or more points corresponding to at least one of a height or a confidence associated with the one or more samples.

[0243] P. A system as described in any of paragraphs KO, wherein the one or more processors are further used to: generate third information indicating one or more prior locations associated with the one or more first objects in the environment, the third information including one or more points representing one or more prior samples obtained using the one or more first sensors during the time period and refined based at least on the second information, wherein generating at least one of the updated version of the first information or the updated version of the second information is further based at least on the third information.

[0244] Q. The system of any of paragraphs KP, wherein the one or more first sensors include one or more of: an image sensor; a radar sensor; an ultrasonic sensor; or a LiDAR sensor.

[0245] R. A system as described in any of paragraphs KQ, 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 simulation operations; a system for performing digital twin operations; a system for performing light transport simulations; a system for performing collaborative content creation of 3D assets; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system implementing one or more large language models (LLMs); a system for generating synthetic data; 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.

[0246] S. A processor comprising: one or more processing units, the one or more processing units being configured to: generate one or more updated versions associated with an occupancy representation corresponding to an environment or at least one of one or more attributes associated with one or more objects detected in the environment, wherein the occupancy representation is generated using one or more neural networks and based at least on first sensor data obtained using one or more first sensors of one or more first sensor modalities, and the one or more attributes are determined using one or more algorithmic processes and based at least on second sensor data obtained using one or more second sensors of one or more second sensor modalities different from the one or more first sensor modalities.

[0247] T. A processor as described in paragraph S, wherein the processor 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 simulation operations; a system for performing digital twin operations; a system for performing light transport simulations; a system for performing collaborative content creation of 3D assets; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system implementing one or more large language models (LLMs); a system for generating synthetic data; 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.

Claims

1. A method comprising: generating, using the one or more neural networks and based at least on first sensor data generated using one or more first sensors of the one or more first sensor modalities, first data indicative of one or more first locations associated with one or more first objects in the environment; generating second data indicative of one or more first properties associated with one or more second objects in the environment based at least on second sensor data generated using one or more second sensors of one or more second sensor modalities different from the one or more first sensor modalities; Generating third data based on at least a first portion of the first data and at least a second portion of the second data, the third data comprising at least one of the following: an updated version of the first data indicating one or more second locations associated with the one or more first objects; or an updated version of the second data indicating one or more second properties associated with the one or more second objects; as well as One or more operations regarding control of the machine are performed based at least on the third data.

2. The method of claim 1, wherein: The first data is instantaneous data indicating the one or more first locations associated with the one or more first objects in the environment around the machine at an instance in time, and The second data is time data, and the one or more first attributes are tracked over a time period that at least partially precedes the time instance.

3. The method of claim 1, wherein: The first data further indicates one or more occluded portions of the environment, and The updated version of the first data further indicates whether the one or more occluded portions of the environment are occupied by at least one first object of the one or more first objects or at least one second object of the one or more second objects.

4. The method of claim 1, wherein: The one or more first attributes include at least one of the following: a first enclosing shape associated with an object of the one or more second objects, a first location associated with the object, a first pose associated with the object, a first track associated with the object, or a first category associated with the object, and The one or more second attributes include at least one of the following: a second enclosing shape associated with the object, a second location associated with the object, a second pose associated with the object, a second track associated with the object, or A second classification is associated with the object.

5. The method of claim 1 , wherein the first data is a dense occupancy representation of the environment from a top-down perspective, the dense occupancy representation comprising one or more points representing one or more samples obtained using the one or more first sensors at a time instance, wherein one or more first points of the one or more points correspond to the one or more first locations associated with the one or more first objects, and one or more second points of the one or more points correspond to one or more unoccupied locations in the environment at the time instance. 6 . The method of claim 5 , wherein the one or more values ​​of the one or more points correspond to at least one of a height or a confidence level associated with the one or more samples.

7. The method of claim 1, further comprising: determining that a first object of the one or more first objects corresponds to a second object of the one or more second objects, Wherein generating the third data is further based on at least the first object corresponding to the second object.

8. The method of claim 1, further comprising: generating fourth data indicative of one or more prior locations associated with the one or more first objects in the environment, the fourth data comprising one or more points representing one or more prior samples obtained using the one or more first sensors over a time period and refined based on at least the one or more first attributes; Wherein generating the third data is further based on at least the fourth data. 9 . The method of claim 8 , wherein a first point of the one or more points included in the fourth data indicates a speed associated with an object of the one or more first objects.

10. The method of claim 1, further comprising: The machine is caused to perform one or more operations based on at least one of the updated version of the first data or the updated version of the second data.

11. A system comprising: One or more processors configured to: obtaining first information indicating one or more locations associated with one or more first objects in the environment at a time instance, the first information generated using one or more neural networks and based on at least first sensor data obtained using one or more first sensors of the machine; obtaining second information indicative of one or more attributes associated with one or more second objects in the environment, the second information determined based on at least second sensor data obtained using one or more second sensors of the machine during a time period that at least partially precedes the time instance; Generate at least one of the following: an updated version of the first information based at least on the second information, the updated version of the first information indicating at least one or more updated locations associated with the one or more first objects; or An updated version of the second information is based at least on the first information, the updated version of the second information indicating at least one or more updated properties associated with the one or more second objects.

12. The system of claim 11, wherein the first information is associated with the time instance, and the updated version of the first information is time information associated with the time instance and at least a portion of the time period.

13. The system of claim 11, wherein: The first information indicates one or more occluded portions of the environment at the time instance, and The updated version of the first information further indicates whether the one or more occluded portions of the environment are occupied by at least one first object of the one or more first objects or at least one second object of the one or more second objects.

14. The system of claim 11, wherein: The one or more updated attributes include at least one of the following: an enclosing shape associated with an object of the one or more second objects, the location associated with the object, the pose associated with the object, the track associated with the object, or The classification associated with the object.

15. The system of claim 11, wherein: The first information is a dense occupancy representation of the environment from a top-down perspective, the dense occupancy representation comprising one or more points representing one or more samples obtained using the one or more first sensors at the time instance, One or more first points of the one or more points correspond to the one or more locations associated with the one or more first objects, One or more second points of the one or more points correspond to one or more unoccupied locations in the environment at the time instance, and One or more values ​​in the one or more points correspond to at least one of a height or a confidence level associated with the one or more samples.

16. The system of claim 11 , wherein the one or more processors are further configured to generate third information indicating one or more prior locations associated with the one or more first objects in the environment, the third information comprising one or more points representing one or more prior samples obtained using the one or more first sensors during the time period and refined based at least on the second information, wherein generating at least one of the updated version of the first information or the updated version of the second information is further based at least on the third information.

17. The system of claim 11, wherein the one or more first sensors include one or more of: Image sensor; radar sensors; Ultrasonic sensor; or LiDAR sensor.

18. The system of claim 11, 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 simulation operations; Systems for performing digital twin operations; a system for performing light transport simulations; A system for performing collaborative content creation of 3D assets; Systems for performing deep learning operations; Systems implemented using edge devices; Systems implemented using robots; Systems for performing conversational AI operations; A system implementing one or more large language models (LLMs); Systems for generating synthetic data; 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 processing units configured to generate one or more updated versions associated with an occupancy representation corresponding to an environment or at least one of one or more attributes associated with one or more objects detected in the environment, wherein the occupancy representation is generated using one or more neural networks and based on at least first sensor data obtained using one or more first sensors of one or more first sensor modalities, and the one or more attributes are determined using one or more algorithmic processes and based on at least second sensor data obtained using one or more second sensors of one or more second sensor modalities different from the one or more first sensor modalities.

20. The processor of claim 19, wherein the processor is included in at least one of: control systems for autonomous or semi-autonomous machines; Perception systems for autonomous or semi-autonomous machines; a system for performing simulation operations; Systems for performing digital twin operations; a system for performing light transport simulations; A system for performing collaborative content creation of 3D assets; Systems for performing deep learning operations; Systems implemented using edge devices; Systems implemented using robots; Systems for performing conversational AI operations; A system implementing one or more large language models (LLMs); Systems for generating synthetic data; 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