USER AUTHENTICATION FOR ACCESS TO THE VEHICLE AND CABIN EXPERIENCE WITH INFRARED IMAGING

Abstract

In various examples, infrared image data can be used to capture one or more subcutaneous features (e.g., a palm venous topology) of a person (e.g., a person requesting access to a vehicle, a vehicle occupant) and to authenticate the user based on the one or more captured subcutaneous features. For example, infrared image data can be generated that represents one or more captured subcutaneous features (e.g., a topology of veins and / or other blood vessels in an area of ​​the palm, hand, neck, forearm, face, fingertip, eye, etc., of the authenticating user). Hand and / or palm capture can be applied to capture an area representing the user's hand or palm, and this area (or a subset thereof) can be segmented to create a representation of a captured venous topology.The recorded vein topology can be compared with one or more reference vein topologies stored in a database to determine whether the recorded vein topology matches one of the reference vein topologies.

Description

BACKGROUND

[0001] Vehicle designs typically rely on some form of authentication to ensure that only authorized users gain access to a vehicle. Traditional physical keys are one of the simplest authentication methods. The owner or authorized operator of the vehicle uses a unique key that fits into the vehicle's lock to gain access. Remote keyless entry systems (e.g., key fobs or smart keys) can grant access to the vehicle by emitting a radio frequency signal that communicates with the electronics in the vehicle to unlock the doors when the user is near the vehicle. Some keyless entry and start systems (e.g.,Push-to-start systems use proximity sensors to detect the presence of an authorized key fob and unlock the doors when the user presses a button on the door handle while the key fob is near the vehicle. Push-to-start systems detect when an authorized key fob is inside the vehicle and start the engine when the user presses a start button while the key fob is inside the vehicle. Some modern vehicles are equipped with biometric authentication systems such as fingerprint scanners or facial recognition systems. These systems can identify an authorized user based on their unique fingerprint or captured facial features and grant access to the vehicle accordingly.

[0002] One of the main disadvantages of existing biometric authentication systems is their reliance on prominent physical characteristics for user authentication. For example, conventional facial recognition systems in existing vehicles can be manipulated by positioning an image or a three-dimensional (3D) cutout of the vehicle owner in front of the system's camera, causing the system to falsely authenticate an intruder as the owner. As another example, conventional facial recognition systems may exhibit lower accuracy and fail in situations where a prominent physical characteristic is obscured (e.g., when the user is wearing a mask or other facial covering). Accordingly, existing biometric authentication methods are vulnerable to hacking and errors, which can compromise the security of vehicles and / or passengers.There is therefore a need for improved biometric authentication methods that can reduce the risk of unauthorized access and / or improve the convenience of entering and operating a vehicle. BRIEF DESCRIPTION OF THE INVENTION

[0003] The invention is defined by the claims. To illustrate the invention, aspects and embodiments are described herein, which may or may not fall within the scope of the claims.

[0004] Embodiments of the present disclosure relate to personal (e.g., subcutaneous) authentication for vehicle access and the in-cabin experience. Systems and methods are disclosed that use infrared image data to detect one or more subcutaneous features (e.g., a representation of vein topology, such as a palm vein map) of an individual (e.g., a person requesting access to a vehicle, a vehicle occupant) and authenticate the individual based on the one or more detected subcutaneous features.

[0005] Unlike conventional systems such as those described above, an infrared camera can use a near-infrared light source (e.g., with a wavelength of 850 nanometers, 940 nanometers, etc.) to generate infrared image data representing one or more detected subcutaneous features (e.g., a topology of veins and / or other blood vessels in an area of ​​the authenticating user's palm, hand, neck, forearm, face, fingertip, eye, etc.). Thus, the infrared image data can be segmented to generate a representation of the one or more detected subcutaneous features (e.g., a binary map representing segmented veins), and the one or more detected subcutaneous features can be matched to one or more corresponding subcutaneous reference features stored in a database to authenticate the user.

[0006] In various examples, infrared image data may be used to detect one or more subcutaneous features (e.g., a palm vein topology) of a person (e.g., a person requesting access to a vehicle, a vehicle occupant) and authenticate the user based on the one or more detected subcutaneous features. For example, infrared image data may be generated representing one or more detected subcutaneous features (e.g., a topology of veins and / or other blood vessels in an area of ​​the palm, hand, neck, forearm, face, fingertip, eye, etc., of the authenticating user). Hand and / or palm detection may be applied to detect a region representing the user's hand or palm, and this region (or a subset thereof) may be segmented to generate a representation of a detected vein topology.The acquired vein topology can be compared with one or more reference vein topologies stored in a database to determine whether the acquired vein topology matches one of the reference vein topologies.

[0007] In one embodiment involving access to an interior space (e.g., a vehicle cabin), an image capture loop may be triggered upon detecting a request to unlock one or more doors (e.g., via a key fob or smart key, via pressing a button, e.g., on a door handle), upon detecting the presence of an authorized key fob or smart key via a proximity sensor, and / or by other means. For example, with palm vein authentication, the authenticating user may place their hand in front of an infrared camera positioned and / or oriented to view the exterior of the vehicle, and the infrared camera may be used to generate an infrared image of the user's hand.Hand and / or palm detection can be applied to the infrared image to capture a region representing the user's hand or palm, and this region (or a subset thereof) can be segmented to generate a representation of a captured vein topology.

[0008] In one embodiment involving the authentication of an occupant of an interior space, an image capture loop may be triggered when one or more doors to the interior space are unlocked, upon detecting the presence of an authorized key fob or smart key in the interior space via a proximity sensor, upon detecting (e.g., via an OMS sensor) a body posture indicating that one or more keypoints representing the user's limbs, hand, wrist, and / or palm are within a threshold distance of the IR camera, and / or by other means. Using palm vein authentication as an example, the authenticating user may place their hand in front of an infrared camera (e.g.,a camera of the driver monitoring system positioned in the steering column), the infrared camera can be used to generate an infrared image of the user's hand, hand detection and / or palm detection can be applied to the infrared image to capture an area representing the user's hand and / or palm, and this area (or a subset thereof) can be segmented to generate a representation of a captured vein topology.

[0009] The sensed vein topology captured by an authenticating user may be compared to one or more reference vein topologies stored in a database to determine whether the sensed vein topology matches one of the reference vein topologies. For example, a sensed (e.g., binary) palm vein topology may be matched to a reference (e.g., binary) palm vein topology, and the sensed and reference palm vein topologies may be applied to one or more machine learning models (e.g., a discriminator network) to generate a representation of whether or not corresponding regions of the sensed and reference palm vein topologies are predicted to match. A measure of the likelihood of a match and / or similarity (e.g., the number of matching regions) may be identified and compared to a specified threshold (e.g.,B. may depend on the resolution of the palm vein topologies and / or the regions to be compared). If the degree of match probability and / or similarity (e.g., the number of matching regions) reaches or exceeds a specified threshold, the authenticating user can be authenticated. Otherwise, the captured palm vein topology can be tested against another reference palm vein topology in the database until, for example, the user is authenticated or no more reference palm vein topologies are available to test.

[0010] As such, and depending on the application, in response to the user's authentication, one or more corresponding actions may be performed (e.g., unlocking one or more doors; updating one or more vehicle, cabin, console, and / or other settings to reflect stored values ​​associated with a corresponding user profile, e.g., seat position, mirror adjustment, steering wheel position, heating and / or cooling settings, audio and / or volume presets, drive mode, console display settings, driver assistance settings, vehicle lighting, etc.; and / or other actions).

[0011] The disclosure extends to all novel aspects or features described and / or illustrated herein.

[0012] Further features of the disclosure are characterized by the independent and dependent claims.

[0013] Any feature of one aspect of the disclosure may be applied to other aspects of the disclosure in any suitable combination. In particular, method aspects may be applied to device or system aspects, and vice versa.

[0014] Furthermore, features implemented in hardware may be implemented in software, and vice versa. Any reference to software and hardware features herein should be construed accordingly.

[0015] Any system or device feature described herein may also be provided as a method feature, and vice versa. System and / or device aspects described functionally (including means plus functional features) may alternatively be expressed in terms of their corresponding structure, such as a suitably programmed processor and associated memory.

[0016] It is also to be understood that certain combinations of the various features described and defined in each aspect of the disclosure may be implemented and / or provided and / or used independently of one another.

[0017] The disclosure also provides computer programs and computer program products comprising software code configured to, when executed on a data processing device, perform any of the methods described herein and / or embody any of the device and system features described herein, including any or all component steps of a method.

[0018] The disclosure also includes a computer or computer system (including networked or distributed systems) having an operating system that supports a computer program for performing the methods described herein and / or for embodying the device or system features described herein.

[0019] The disclosure also provides a computer-readable medium on which one or more of the aforementioned computer programs are stored.

[0020] The disclosure also provides a signal carrying one or more of the aforementioned computer programs.

[0021] The disclosure extends to methods and / or devices and / or systems as described herein with reference to the accompanying drawings.

[0022] Aspects and embodiments of the disclosure will now be described, by way of example only, with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The present systems and methods for subcutaneous authentication are described in detail below with reference to the accompanying drawing figures, in which: Fig. 1 is a data flow diagram illustrating an example palm vein authentication pipeline according to some embodiments of the present disclosure; Fig. 2A illustrates an exemplary placement of a camera of the driver monitoring system within a steering column, and Fig. 2B is an exemplary infrared image of the palm of an authenticating user from the perspective of a camera of the driver monitoring system within a steering column, according to some embodiments of the present disclosure; Fig. 3 illustrates one or more exemplary regions of interest in the exemplary infrared image of Fig. 2B, according to some embodiments of the present disclosure; Fig. 4 illustrates an example representation of a segmented vein topology according to some embodiments of the present disclosure; Fig. 5A and Fig. 5B illustrate example representations of correspondences between a detected vein topology and a reference vein topology, according to some embodiments of the present disclosure; Fig. 6 is a flowchart illustrating a method for performing one or more actions based at least on matching a representation of one or more detected subcutaneous features with a corresponding representation of one or more subcutaneous reference features, according to some embodiments of the present disclosure; Fig. 7 is a flowchart illustrating a method for comparing detected and reference vein topologies according to some embodiments of the present disclosure; Fig. 8A is an illustration of an exemplary autonomous vehicle, according to some embodiments of the present disclosure; Fig. Figure 8B is an example of camera locations and fields of view for the example autonomous vehicle of Fig. 8A, according to some embodiments of the present disclosure; Fig. 8C is a block diagram of an example system architecture for the example autonomous vehicle of Fig. 8A, according to some embodiments of the present disclosure; Fig. 8D is a system diagram for communication between the one or more cloud-based servers and the example autonomous vehicle of Fig. 8A, according to some embodiments of the present disclosure; Fig. 9 is a block diagram of an exemplary computing device suitable for use in implementing some embodiments of the present disclosure; and Fig. 10 is a block diagram of an exemplary data center suitable for use in implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0024] Systems and methods are disclosed relating to user authentication for vehicle access and the in-cabin experience using infrared images. For example, systems and methods are disclosed that use infrared image data to detect one or more subcutaneous features (e.g., a representation of a vein topology, such as a palm vein map) of a person (e.g., a person requesting access to a vehicle, a vehicle occupant) and to authenticate the person based on the one or more detected subcutaneous features. The present techniques can be used to authenticate a person and / or perform a corresponding action thereon (e.g.,unlock one or more doors; update one or more vehicle, cabin, console and / or other settings to reflect stored values ​​associated with or otherwise customized for a corresponding user profile, such as seat position, mirror adjustment, steering wheel position, heating and / or cooling settings, audio and / or volume presets, drive mode, console display settings, driver assistance settings, vehicle lighting, etc.; and / or other actions) in systems with surround-view visualization, occupant monitoring (e.g., driver and / or passenger monitoring), and / or other types of systems.

[0025] Although the present disclosure is described with respect to an exemplary autonomous or semi-autonomous vehicle or machine 800 (alternatively referred to herein as “vehicle 800” or “ego machine 800,” an example of which is described with respect to Fig. 8A-8D), this is not to be construed as a limitation. For example, the systems and methods described herein may be used, without limitation, by non-autonomous vehicles or machines, semi-autonomous vehicles or machines (e.g., in one or more Advanced Driver Assistance Systems (ADAS)), autonomous vehicles or machines, piloted and unpiloted robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, flying vessels, boats, shuttles, emergency vehicles, motorcycles, electric or motorized bicycles, aircraft, construction vehicles, trains, underwater vehicles, remotely operated vehicles such as drones, and / or other vehicle types. Although the present disclosure is directed to the subcutaneous (e.g., vein topology, such asWhile the term "palm vein authentication" can be described as "fingerprint authentication" for vehicle access and the in-cabin experience, this is not to be construed as 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 area where authentication may be used.

[0026] At a high level, one or more infrared (IR) sensors (e.g., one or more IR cameras, combined RGB / infrared (RGB / IR) cameras) may be positioned to monitor an authenticating user. In an embodiment involving access to a vehicle, one or more IR sensors may be positioned to monitor a person requesting access to the vehicle. In an embodiment for an occupant monitoring system (OMS), one or more IR sensors may be positioned to monitor one or more occupants in a cabin, cockpit, or other interior space. An OMS may, for example, include a driver monitoring system (DMS), a system that monitors non-driver occupants, or a system that monitors driver occupants and / or non-driver occupants.The one or more IR sensors may use a near-infrared (NIR) light source (e.g., having a wavelength of 850 nanometers, 940 nanometers, etc.) to generate IR image data representing one or more detected subcutaneous features (e.g., a topology of veins and / or other blood vessels in an area of ​​the authenticating user's palm, hand, neck, forearm, face, fingertip, eye, etc.). Thus, the IR image data may be segmented to generate a representation of the one or more detected subcutaneous features (e.g., a binary map representing segmented veins), and the one or more detected subcutaneous features may be matched to one or more corresponding subcutaneous reference features stored in a database to authenticate the user.

[0027] Using palm vein authentication as an example, in some embodiments, an image capture loop may be triggered to generate one (or more) IR image(s) and detect whether a suitable palm vein topology is visible in the IR image. In an embodiment involving access to an interior space (e.g., a vehicle cabin), an image capture loop may be triggered upon detecting a request to unlock the door(s) (e.g., via a key fob or smart key, via pressing a button, e.g., on a door handle), upon detecting the presence of an authorized key fob or smart key via a proximity sensor, and / or by other means.In one embodiment involving authentication of an occupant of an interior space, the image capture loop may be triggered when the one or more doors to the interior space are unlocked, after detecting the presence of an authorized key fob or smart key in the interior space via a proximity sensor, upon detecting (e.g., via an OMS sensor) a body posture indicating that one or more keypoints representing the occupant's limbs, hand, wrist, and / or palm are within a threshold distance of the IR camera, and / or in another manner. Once triggered, an image capture loop may (e.g., repeatedly) generate an IR image, detect the presence of a hand or palm in the IR image, normalize the posture of a detected hand or palm (e.g.,by aligning a detected posture with the IR camera and distorting the IR image accordingly) and / or detecting the presence of veins in one or more areas of the detected hand or palm. In some embodiments, the image acquisition loop may check the detection accuracy of the hand, palm, and / or vein detection against a corresponding confidence threshold. If a hand, palm, and / or vein topology is not detected with at least a minimum confidence, the image acquisition loop may retrigger the loop and acquire a new IR image. In some embodiments, the image acquisition loop may provide feedback on a display visible to the authenticating user with posture instructions on how to position the hand to increase the probability of a positive detection (e.g.,The instructions can be based on acquisition accuracy, the size and / or position of the region of the IR image in which a captured hand or palm was detected, a captured pose (e.g., indicating a position of the user's limb, hand, wrist, and / or palm relative to the IR camera), and / or other features from a previous loop iteration. Thus, the image acquisition loop can iterate at any frame rate, for example, until a vein topology is captured with at least a minimum confidence.

[0028] In some embodiments, a detected region predicted to contain a hand and / or a palm (e.g., a predicted and / or normalized bounding box or other boundary shape) may be narrowed, fine-tuned, and / or otherwise processed to identify one or more regions of interest (e.g., one or more subset regions of a detected region), and palm vein detection may be performed for the predicted region and / or the one or more identified regions of interest. For example, most palms are likely to be located in a subset of the hand (e.g., the bottom 50% of the hand's bounding box, a subset region whose size is based on the correlation between palm size and the ratio of the index finger to the little finger, etc.).) so that one or more corresponding subset regions of a predicted bounding box or other boundary shape can be identified. In some embodiments, palm vein detection can be performed separately for different regions of interest (e.g., by sweeping one or more kernels over the rows of a predicted bounding box to test different subset regions, by performing a first pass for a larger region of interest and a second pass for smaller regions of interest, such as a subset region identified in the first pass, etc.). Thus, one or more subset regions of a predicted bounding box or other boundary shape can be identified, and palm vein detection can be performed for the one or more subset regions.

[0029] In some embodiments, an IR image (e.g., resulting from multiplying the original IR image with a binary mask corresponding to an identified region of interest, rotated to align an identified region of interest with an input grid, such as a rectangular grid) may be segmented to extract a representation of a detected vein topology (e.g., a palm vein topology) that may include binary values ​​indicating whether each pixel is expected to represent a detected vein. The detected vein topology, captured by an authenticating user, may be compared to one or more reference vein topologies stored in a database to determine whether the detected vein topology matches one of the reference vein topologies. For example, a detected (e.g., binary) palm vein topology may be compared to a reference (e.g.,binary) palm vein topology can be matched, and the sensed and reference palm vein topologies can be applied to one or more machine learning models (e.g., a discriminator network) to generate a representation of whether corresponding regions of the sensed and reference palm vein topologies are predicted to match. A measure of the probability of a match and / or similarity (e.g., the number of matching regions) can be identified and compared to a specified threshold (which may, for example, depend on the resolution of the palm vein topologies and / or the regions being compared). If the measure of the probability of a match and / or similarity (e.g., the number of matching regions) meets or exceeds a specified threshold, the authenticating user can be authenticated.Otherwise, the captured palm vein topology can be tested, for example, against another reference palm vein topology in the database until the user is authenticated or there are no more reference palm vein topologies to test.

[0030] Thus, the techniques described herein can be used to authenticate an individual and / or perform a corresponding action (e.g., unlock one or more doors; update one or more vehicle, cabin, console, and / or other settings to reflect stored values ​​associated with a corresponding user profile; and / or other actions). The present techniques improve upon conventional biometric authentication techniques in terms of security and privacy. For example, subcutaneous authentication is more robust than conventional facial capture systems because subcutaneous features (e.g., a captured palm vein topology) are only visible under infrared illumination and are therefore more difficult to forge than facial capture systems that rely on visible light.Furthermore, embodiments that authenticate based on infrared illumination of unobscured subcutaneous features (e.g., captured from the palm, hand, neck, forearm, unobscured area of ​​the face, fingertip, eye, etc.) avoid the errors that occur when conventional solutions attempt to authenticate based on an obscured physical feature (e.g., when the user is wearing a mask or other face covering). Furthermore, embodiments that rely on a user positioning themselves in front of a camera to trigger an authentication challenge provide additional flexibility by allowing the user to choose whether or not to authenticate, unlike conventional facial capture techniques where there is no way to hide from the camera and trigger an authentication challenge.Thus, the techniques described here represent improved biometric authentication techniques that can reduce the risk of unauthorized access and / or improve the convenience of entering and operating a vehicle.

[0031] Referring to Fig. 1 shows Fig. 1 illustrates an example palm vein authentication pipeline 100 according to some embodiments of the present disclosure. It should be understood that these and other arrangements described herein are set forth as examples only. Other arrangements and elements (e.g., engines, interfaces, functions, arrangements, groupings of functions, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities that may be implemented as individual or distributed components, or in conjunction with other components, and in any suitable combination and location. Various functions performed by entities described herein may be performed by hardware, firmware, and / or software.For example, various functions may be performed by a processor executing instructions stored in memory. In some embodiments, the systems, methods, and processes described herein may be implemented using similar components, features, and / or functionality as those of the exemplary autonomous vehicle 800 of FIG. Fig. 8A-8D, the exemplary computing device 900 of Fig. 9 and / or the exemplary data center 1000 from Fig. 10.

[0032] In the Fig. 1, the palm vein authentication pipeline 100 includes an IR camera 105, a loop control component 110, an image acquisition component 120, a region of interest generator 130, a palm vein detector 140, and an authentication component 150. At a high level, the loop control component 110 can trigger and control an image acquisition loop in which the image acquisition component 120 can use the IR camera 105 to generate IR image data representing a part (e.g., a palm, a hand, a neck, a forearm, a face, a fingertip, an eye, etc.).) of an authenticating user and identify and / or locate a detected hand or palm represented in the IR image data, the region of interest generator 130 can identify a region of the detected hand or palm, and the palm vein detector 140 can segment a vein topology (or one or more other subcutaneous features visible in the infrared illumination) from the identified region of the detected hand or palm. The loop control component 110 can iterate the loop at any rate until a detected hand or palm region and / or segmented vein topology is detected with a detection accuracy that meets or exceeds a confidence threshold.Thus, the authentication component 150 may compare this captured vein topology to one or more reference vein topologies 160 representing authorized users to determine whether the authenticating user is an authorized user. If so, the authentication component 150 may determine that the authenticating user is authenticated and perform an appropriate action, such as granting access (e.g., unlocking one or more vehicle locks 170, applying one or more vehicle settings 180 associated with the authenticated user's profile, etc.).

[0033] The IR camera 105 (which may, for example, be a dedicated IR camera, an RGB / IR camera, and / or another device that includes one or more IR sensors) may include or otherwise be coordinated with a near-infrared (NIR) light source (e.g., having a wavelength of 850 nanometers, 940 nanometers, etc.) to illuminate a target area within the field of view of the IR camera 105. Thus, the IR camera 105 may include one or more IR sensors that can be used to observe IR light (e.g., emitted and / or reflected IR light) and generate IR image data representing the scene and / or the one or more objects within the field of view of the IR camera 105. In general, the IR image data (e.g.,An IR image (e.g., an IR image) may contain pixels that store values ​​corresponding to sensed temperatures, and the IR image data may take the form of a grayscale image or a false-color representation, which uses different shades of gray or different colors to represent different temperatures.

[0034] The IR camera 105 can be positioned and / or oriented in various ways to observe an authenticating user depending on the application (e.g., the security application, the body area to be examined for one or more subcutaneous features, etc.) and / or the embodiment. For example, in some embodiments that observe a topology of veins and / or other blood vessels in a particular area of ​​the authenticating user's body, the IR camera 105 can be positioned to be accessible to the authenticating user and / or to facilitate positioning of the salient part of an authenticating user's body within the field of view and at a suitable distance from the IR camera 105 (e.g., by the user placing their palm or forearm in front of the camera; placing their face, neck, or eye in front of the camera; etc.). In an example where access (e.g.,to a physical space such as a room, building, or compartment; to software or electronics such as a computer or storage device; to an (e.g., access-controlled) account; to an online service; to a database or record; etc.) is secured via subcutaneous authentication, the IR camera 105 may be positioned on a wall or other component of the physical space to be secured, in a computer or storage device used for the access request (e.g., an IR webcam, IR smartphone camera, etc.), and / or in another manner. In one embodiment where a vehicle is entered, the IR camera 105 may be an ambient camera (e.g., a side view camera of the one or more ambient cameras 874 of the vehicle 800 of FIG. Fig. 8A-8D, which is equipped with infrared sensor functionality) and / or otherwise positioned to observe a person requesting access to the vehicle. In an embodiment that includes authorization to start the vehicle engine or the detection of a specific identity and loading one or more stored vehicle settings associated with that identity, the IR camera 105 may be part of an OMS or DMS. Vehicle manufacturers tend to vary the number of OMS cameras from model to model and depending on the trim level. Base models typically have an IR camera facing the driver (e.g., in a steering column, vehicle pillar, or infotainment console). Higher trim levels may include any number of additional cameras (e.g.,one in the steering column facing the driver, one in the rearview mirror facing the driver or the cabin, one in a vehicle pillar facing a specific row of seats, etc.). Fig. 2A illustrates an example implementation where the IR camera 105 corresponds to a strain gauge camera 210 mounted in a steering column. This is intended as an example only, and other types of IR cameras and / or IR sensors may be positioned and / or oriented within the scope of the present disclosure.

[0035] In some embodiments, the loop control component 110 may initiate and / or control an image capture loop that uses the IR camera 105 to repeatedly generate IR image data (e.g., an IR image) and / or otherwise access IR image data generated by the IR camera 105. In an embodiment involving access to an interior space (e.g., a vehicle cabin), the loop control component 110 may initiate the image capture loop when the loop control component 110 or another component detects a request to unlock the one or more doors (e.g., via a key fob or smart key, via pressing a button, e.g., on a door handle), when the loop control component 110 or another component detects the presence of an authorized key fob or smart key via a proximity sensor, and / or otherwise.In an embodiment involving the authentication of an occupant of an interior (e.g., in connection with authorization to start the vehicle engine, detecting a particular identity and loading one or more stored vehicle settings associated with that identity, etc.), the image capture loop may be triggered when the loop control component 110 or another component detects that the one or more doors to the interior are unlocked, when the loop control component 110 or another component detects the presence of an authorized key fob or smart key within the interior via a proximity sensor, when the loop control component 110 or another component (e.g.,via an OMS sensor, which may be a different camera than the IR camera 105) detects a posture indicating that one or more key points representing the user's limbs, hand, wrist, and / or palm are within a threshold distance of the IR camera 105, and / or otherwise. In one embodiment for detecting posture, an OMS camera (e.g., in the rearview mirror) may be used to detect the vehicle operator's posture, and the loop control component 110 may evaluate the detected posture to determine when the detected posture indicates that the vehicle operator is positioning their hand or palm within a threshold distance of a DMS camera in the steering column.

[0036] Thus, the authenticating user may position themselves and / or the salient part of their body within the field of view of the IR camera 105. In some embodiments (e.g., when the loop control component 110 initiates an image capture loop), the image capture component 120 may use the IR camera 105 to generate and / or otherwise access IR image data depicting a portion (e.g., a palm, a hand, a neck, a forearm, a face, a fingertip, an eye, etc.) of the authenticating user. Fig. For example, Figure 2B is an exemplary infrared image of the palm of an authenticating user that may be generated in accordance with some embodiments of the present disclosure. In some embodiments, the image capture component 120 of Fig. 1 may identify and / or locate a detected hand or palm represented in the IR image data to enable determination of whether a suitable image of the salient part of the authenticating user's body was generated, and the loop control component 110 may test a detection accuracy of the detected hand or palm. If the loop control component 110 determines that the detection accuracy is below a minimum confidence, the loop control component 110 may capture a new IR image (e.g., at a certain frame rate).

[0037] In the Fig. 1, the one or more subcutaneous features used for authentication may include a palm vein topology, so the image capture component 120 may include a palm detection network 122 and a normalization component 124. At a high level, the palm detection network 122 (or a hand detection network) may include one or more machine learning models (e.g., one or more deep neural networks or "DNNs") that detect and / or locate a hand or palm in the IR image data (e.g., a bounding box or other boundary shape and / or one or more keypoints that indicate the pose of a hand or palm in the IR image data), and the normalization component 124 may normalize the pose of the detected hand or palm (e.g., by aligning a detected pose with the IR camera 105 and warping the IR image accordingly).

[0038] In some embodiments, the palm detection network 122 (and / or any of the other machine learning models and / or detection networks described herein, such as the palm vein detection network 144 and / or the palm vein comparison network 154) may be implemented using a DNN, such as a convolutional neural network (CNN). Although certain embodiments are described in which the palm detection network 122, the palm vein detection network 144, and / or the palm vein comparison network 154 are implemented using neural networks, this is not intended to be limiting.For example, and without limitation, the palm detection network 122, the palm vein detection network 144, and / or the palm vein comparison network 154 may include any of a variety of different networks or machine learning models, such as one or more machine learning models using linear regression, logistic regression, decision trees, support vector machines (SVM), naive Bayes, k-nearest neighbor (Knn), k-means clustering, random forest, dimensionality reduction algorithms, gradient boosting algorithms, neural networks (e.g., auto-encoders, convolutional, transformers, recurrent, perceptrons, long / short term memory (LSTM), Hopfield, Boltzmann, deep belief, de-convolutional, generative adversarial, liquid state machine, etc.), and / or other types of machine learning models.

[0039] In some embodiments, the palm detection network 122 (and / or any of the other machine learning models and / or detection networks described herein, such as the palm vein detection network 144 and / or the palm vein comparison network 154) may include a common root (or stream of layers) with multiple input and / or output heads (or at least partially discrete streams of layers) to accept various inputs and / or predict various outputs. For example, without limitation, the palm detection network 122 may include a feature extractor (e.g., a DNN, an encoder / decoder, etc.) comprising convolutional layers, pooling layers, and / or other layer types, with the output of the feature extractor provided as input to a plurality of output heads (e.g.,Classification and / or regression heads) may be provided to detect and / or classify pixels likely to represent a detected hand or palm (or one or more keypoints or other parts thereof), to detect and / or regress coordinates of a bounding shape (or one or more keypoints or other parts) of a detected hand and / or palm, and / or the like. Thus, any known post-processing technique may be applied to decode the predicted classification data and / or regression data to generate a bounding box, one or more closed polylines, and / or other bounding shape that identifies the location, size, and / or orientation of a detected hand or palm and / or one or more keypoints representing a detected pose of the hand or palm.These and the other architectures described herein are intended as examples only, and other architectures may be implemented within the scope of the present disclosure. Although some embodiments are described with respect to detecting the hand or palm, some embodiments may additionally or alternatively detect and / or localize other body parts, e.g., based on the location of the subcutaneous feature(s) to be authenticated (e.g., neck, forearm, face, fingertip, eye, a prominent part thereof, etc.).

[0040] In some embodiments, the loop control component 110 may check a detection accuracy of the detected hand or palm. For example, the palm detection network 122 may predict classification data (e.g., a binary classification for the entire IR image, per-pixel classification data) representing a probability that the IR image and / or each individual pixel represents a hand and / or a palm. Thus, the loop control component 110 may identify a prediction confidence (e.g., for the binary classification of the entire IR image predicted by the palm detection network 122, an aggregate and / or statistical value representing the prediction confidence of a plurality of per-pixel classifications, such as an average prediction confidence for pixels located within a predicted bounding box, etc.).), and the loop control component 110 may use the identified prediction confidence as the detection accuracy and test the prediction confidence against a specified confidence threshold. If the loop control component 110 determines that the detection accuracy is below a minimum confidence level, the loop control component 110 may acquire a new IR image (e.g., at a specified frame rate). If the loop control component 110 determines that the detection accuracy meets or exceeds the minimum confidence level, the loop control component 110 may proceed to the next step in the process (e.g., in the palm vein authentication pipeline 100). Although the application of this threshold is described as occurring at a specific step in the process, this need not be the case, and the application of this and / or other thresholds described herein may occur at any step in the process.

[0041] In some embodiments, image acquisition component 120 includes a normalization component 124. For example, in some embodiments, an attempt may be made to capture one or more subcutaneous features in a particular body part, such as the hand or palm, and certain postures may facilitate positive capture. Positive captures may be facilitated, for example, by positioning an open palm facing the IR camera 105. Thus, in some embodiments, normalization component 124 may match a captured pose (e.g., represented as 2D and / or 3D coordinates of a plurality of keypoints captured by the palm capture network 122) to a particular pose (e.g., facing the IR camera 105), e.g., by fitting the captured keypoints into a 3D model of a hand and rotating the 3D model to obtain the particular pose.Thus, the normalization component 124 may warp the IR image data (and / or the corresponding predicted classification data, regression data, boundary shape, etc.) to the rotated keypoints to produce normalized IR image data representing the determined pose.

[0042] In some embodiments, the image capture component 120 includes a region of interest generator 130. The region of interest generator 130 may generate a representation of one or more bounded, fine-tuned, and / or otherwise identified regions of interest (e.g., one or more subset regions) of the captured hand or palm (or other region of the authenticating user's body) represented in the (e.g., normalized) IR image data. Taking a captured hand as an example, most of the palms are likely located in a subset region of the hand (e.g., the lower 50% of the bounding box of the hand, a subset region determined based on the correlation between the palm size and the ratio between the index finger and the little finger, etc.), so the region of interest generator 130 may generate one or more corresponding subset regions of a (e.g.,normalized) bounding box or other boundary shape predicted to contain a detected hand (e.g., by scaling a predicted bounding box in one or more dimensions by a specified percentage or calculated ratio). Fig. 3 illustrates an example region of interest 310 (e.g., the palm of an authenticating user) that is a subset of the example infrared image of FIG. 2B, according to some embodiments of the present disclosure.

[0043] Thus, region of interest generator 130 may provide a representation of one or more regions of interest of the detected hand or palm, and palm vein detector 140 may detect the presence of detected veins (e.g., a predicted bounding box or other boundary shape) within a detected hand or palm and / or one or more regions of interest therein and / or otherwise generate a representation thereof. In some embodiments, palm vein detector 140 may be executed separately for different regions of interest (e.g., by sweeping one or more kernels across the rows of a predicted bounding box to test different subset regions, performing a first pass for a larger region of interest and a second pass for smaller regions of interest, such as a subset region identified in the first pass, etc.).Thus, the region of interest generator 130 may identify and provide a representation of one or more subset regions (e.g., a predicted bounding box or other boundary shape) representing a detected hand or palm to the palm vein detector 140, and the palm vein detector 140 may process the detected hand or palm and / or the one or more subset regions to detect (e.g., segment) the veins represented in the detected hand or palm and / or the one or more subset regions.

[0044] In the Fig. 1, the palm vein detector 140 includes an input generator 142 and a palm vein detection network 144. Given a representation (e.g., a binary mask) representing a bounding box or other bounding shape representing a detected hand and / or palm and / or a subset thereof, the input generator 142 may generate and / or otherwise condition IR image data representing the detected hand, palm, and / or subset region in a format acceptable to the palm vein detection network 144. For example, the input generator 142 may multiply the IR image data (e.g., the IR image generated by the IR camera 105, the normalized IR image data generated by the normalization component 124) by a binary mask representing the region to be examined (e.g., the region of interest 310 of Fig. 3), rotate the (e.g., resulting) IR image data to align the area to be inspected with a rectangular grid, and / or apply the (e.g., resulting) IR image data to the palm vein detection network 144.

[0045] For example, the palm vein detection network 144 may segment the (e.g., resulting) input IR image data to generate a representation of one or more detected veins represented in the input IR image data. For example, without limitation, the palm vein detection network 144 may include a feature extractor (e.g., a DNN, an encoder / decoder, etc.) that includes convolutional layers, pooling layers, and / or other layer types, where the output of the feature extractor may be provided to a classification head to detect and / or classify pixels likely to represent a vein (or other blood vessel or other subcutaneous feature visible under IR illumination). In some embodiments, the palm vein detection network 144 may perform class segmentation (e.g.,by applying a per-pixel prediction confidence threshold) to identify pixels likely to represent a vein. Thus, palm vein detection network 144 may predict, segment, and / or otherwise generate a (e.g., binary) representation of a detected (e.g., palm) vein topology for an authenticating user. Fig. 4 illustrates an example representation of a segmented palm vein topology 410 according to some embodiments of the present disclosure.

[0046] In some embodiments, the loop control component 110 may test a detection accuracy of the detected veins. For example, the palm vein detection network 144 may predict classification data (e.g., a binary classification for the entire input IR image, per-pixel classification data) representing a probability that the input IR image and / or each individual pixel represents a detected vein. The loop control component 110 may thus determine a prediction confidence (e.g., for the binary classification of the entire input IR image predicted by the vein detection network 144, an aggregate and / or statistical value representing the prediction confidence of a plurality of per-pixel classifications, such as an average prediction confidence for the pixels within the tested region, etc.).), and the loop control component 110 can use the identified prediction confidence as the detection accuracy and test the prediction confidence against a specified confidence threshold. If the loop control component 110 determines that the detection accuracy is below a minimum confidence level, the loop control component 110 can acquire a new IR image. If the loop control component 110 determines that the detection accuracy meets or exceeds the minimum confidence level, the loop control component 110 can proceed to the next step in the process (e.g., in the palm vein authentication pipeline 100).

[0047] In some embodiments, when the loop control component 110 triggers the image capture component 120 or otherwise actuates it with respect to triggering the image capture component 120 to capture another frame of IR image data, the loop control component 110 may trigger a presentation of feedback on a display visible to the authenticating user with posture guidance instructions on how to position their hand (or another part of their body) to increase the likelihood of a positive capture (e.g., move your hand closer or farther from the camera, move your hand left or right, move your hand up or down, rotate your hand to face the camera, etc.). The loop control component 110 may, based on the capture accuracy (e.g.,the palm detection network 122, the palm vein detection network 144), the size and / or position of the region of the IR image in which a detected hand or palm was detected (e.g., by the palm detection network 122), a detected posture (e.g., indicating a position of the user's limb, hand, wrist, and / or palm relative to the IR camera), and / or other features from a previous iteration of the loop determine which instructions to issue. In one embodiment including posture detection, an OMS camera (e.g., in the rearview mirror) may be used to detect the posture of the vehicle operator, and the loop control component 110 may evaluate the detected posture to detect the distance and / or position of the vehicle operator's hand relative to a DMS camera in the steering column.For example, the loop control component 110 can determine the direction in which the vehicle driver should move their hand and / or what changes in their hand posture orientation are recommended, and the loop control component 110 can control the display to display corresponding instructions. For example, the loop control component 110 can trigger any number of iterations of an image acquisition loop operating at any frame rate (e.g., 5 frames per second (fps), 10 fps, 20, 30, etc.) until a vein topology (or other subcutaneous feature) is detected with at least a minimum confidence level.

[0048] In some embodiments, the authentication component 150 may compare the acquired palm vein topology for the authenticating user to one or more reference palm vein topologies 160 representing authorized users to determine whether the authenticating user is an authorized user. The one or more reference palm vein topologies 160 may be registered using any suitable technique. For example, a vehicle owner and / or an authorized user may create or update a user profile (e.g., via interactions with a display such as an infotainment console). As part of the process, the vehicle owner and / or the authorized user may be prompted to register their palm vein topology (or other subcutaneous features); in this case, a representation of their palm vein topology may be acquired using the techniques described herein (e.g.,using the loop control component 110, the image acquisition component 120, the region of interest generator 130, and / or the palm vein detector 140). In some embodiments, one or more cabin, console, and / or other settings may be stored and / or otherwise associated with their user profile, such as seat position, mirror adjustment, steering wheel position, heating and / or cooling settings, audio and / or volume presets, drive mode, console display settings, driver assistance settings, vehicle lighting, and / or others.

[0049] As such, the authentication component 150 may access one or more of the reference palm vein topologies 160 registered in a database (e.g., in the vehicle or at a remote location accessible via a wireless communication network) and compare the detected vein topology with the reference palm vein topology 160. In the Fig. 1, the authentication component 150 includes an input generator 152, a palm vein comparison network 154, and a matching component 156.

[0050] In some embodiments, the input generator 152 may match (e.g., binary) representations of a detected palm vein topology with a reference palm vein topology (e.g., rotate one or both to align with an (e.g., rectangular) input grid, rotate one or both to have a comparable pose, center or otherwise position one or both in a particular region of a frame such as a central region, etc.).Thus, the input generator 152 may apply the detected and reference palm vein topologies to the palm vein comparison network 154 to generate a representation indicating whether corresponding regions of the detected and reference palm vein topologies are predicted to match, and the matching component 156 may decode the output of the palm vein comparison network 154 to determine whether the detected and reference palm vein topologies are predicted to match.

[0051] The palm vein comparison network 154 may, for example, include one or more machine learning models (e.g., a discriminator network) that compare corresponding regions of the detected and reference palm vein topologies (e.g., binary masks representing these topologies) and generate a representation of whether the corresponding regions match. The palm vein comparison network 154 may, for example, include, without limitation, one or more feature extractors (e.g., a DNN, an encoder / decoder, etc.), including convolutional layers, pooling layers, and / or other layer types, to extract and / or compare features of the detected and reference palm vein topologies. The palm vein comparison network 154 may predict a correspondence map that quantifies the similarity and / or probability that corresponding regions of the detected and reference palm vein topologies match.For example, each element in the correspondence map may quantify the similarity and / or probability of a match between a portion of one topology (e.g., a binary mask representing one topology) and a corresponding portion of the other topology (e.g., a binary mask representing the other topology) within a receptive field of a particular size (e.g., 5x5 pixels). As a non-limiting example, a representation of a palm vein topology may have a pixel resolution of 630x430, and the palm vein comparison network 154 may predict a correspondence map with a resolution of 126x86 based on a receptive field comparing corresponding 5x5 pixel regions of the sensed and reference palm vein topologies. Thus, each element in the predicted output (e.g.,a predicted correspondence map) quantify the probability that the corresponding regions of the acquired and reference palm vein topology match.

[0052] Accordingly, the matching component 156 may quantify the number of matching regions, e.g., by applying a confidence threshold (e.g., 85%) to each element of the predicted output (e.g., a predicted correspondence map). Thus, the matching component 156 may identify the number of matching regions between the acquired and reference palm vein topologies, and the matching component 156 may compare this number to a predetermined threshold, which may depend on the resolution of the palm vein topologies, the receptive field size, the body area being compared, and / or other factors.To stay with the above example with a palm vein topology with a pixel resolution of 630x430 and a correspondence map with a resolution of 126x86, a comparison between acquired and reference vein topologies that do not match may yield about 20 matching regions, whereas a comparison between acquired and reference vein topologies that do match may result in about 100 matches, so a threshold like 70 can be applied. Fig. 5A is a visualization of example correspondences 530 between a detected vein topology 510 and a reference vein topology 520 that do not match, and Fig. 5B is a visualization of example correspondences 550 between a detected vein topology 510 and a reference vein topology 540 that match.

[0053] If the matching component 156 determines whether the number of matching regions meets or exceeds a certain threshold, the matching component 156 may determine that the authenticating user is an authorized user and is therefore authenticated. Otherwise, the authentication component 150 may compare the acquired palm vein topology to another reference vein topology 160 in the database, for example, until the user is authenticated or there are no further reference palm vein topologies to compare. If the authentication component 150 (e.g., the matching component 156) determines that the user is authenticated, the authentication component 150 may perform an appropriate action, such as granting access (e.g.,triggering one or more actuators to unlock one or more corresponding vehicle locks 170, applying one or more stored settings (e.g., one or more vehicle settings 180) associated with the profile of the authenticated user, etc.).

[0054] Referring to Fig. 6 and Fig. 7, each block of the methods 600 and 700 described herein comprises a computational process that may be performed using any combination of hardware, firmware, and / or software. Various functions may be performed, for example, by a processor executing instructions stored in memory. The methods may also be embodied as computer-usable instructions stored on computer storage media. The methods may be provided by a standalone application, a service or hosted service (standalone or in combination with another hosted service), or a plug-in for another product, to name a few. Furthermore, the methods 600 and 700 are exemplified with respect to the palm vein authentication pipeline 100 of Fig. 1. However, these methods may additionally or alternatively be performed by any system or combination of systems, including, but not limited to, the systems described herein.

[0055] Fig. 6 is a flowchart illustrating a method 600 for performing one or more actions based at least on matching a representation of one or more detected subcutaneous features with a corresponding representation of one or more subcutaneous reference features, according to some embodiments of the present disclosure. The method 600 includes, at block B602, generating IR image data representing an area of ​​an authenticating user. With respect to Fig. 1, for example, the image capture component 120 may use the IR camera 105 to generate IR image data representing a portion (e.g., a palm, a hand, a neck, a forearm, a face, a fingertip, an eye, etc.) of an authenticating user.

[0056] The method 600 includes, in block B604, generating a representation of one or more detected subcutaneous features based at least on the segmentation of at least a portion of the IR image data. Fig. 1, for example, the image capture component 120 (e.g., the palm capture network 122) may identify and / or locate a captured hand or a captured palm represented in the IR image data, and / or the palm vein detector 140 may segment a vein topology (or one or more other subcutaneous features visible under infrared illumination) from the IR image data and / or a portion thereof. In some embodiments, the region of interest generator 130 may generate a representation of one or more subset regions of a captured hand or palm (e.g.,a predicted bounding box or other boundary shape representing a detected hand or palm) and provide it to the palm vein detector 140, and the palm vein detector 140 may process the detected hand or palm and / or the one or more subset regions to detect (e.g., segment) the veins represented in the detected hand or palm and / or the one or more subset regions.

[0057] The method 600 includes, in block B606, performing one or more actions based at least on comparing the representation of the one or more detected subcutaneous features with a corresponding representation of one or more subcutaneous reference features stored in a database. Fig. 1, for example, the authentication component 150 may compare a captured palm vein topology for an authenticating user to one or more reference palm vein topologies 160 representing authorized users to determine whether the authenticating user is an authorized user. If the authentication component 150 (e.g., the matching component 156) determines that the user is authenticated, the authentication component 150 may perform an appropriate action, such as granting access (e.g., triggering one or more actuators to unlock one or more corresponding vehicle locks 170, applying one or more stored settings (e.g., one or more vehicle settings 180) associated with the authenticated user's profile, etc.).

[0058] Fig. 7 is a flowchart illustrating a method 700 for comparing acquired and reference vein topologies according to some embodiments of the present disclosure. The method 700 includes, at block B702, initiating IR image acquisition. Fig. 1, the loop control component 110 may, for example, initiate and / or control an image capture loop in which the image capture component 120 uses the IR camera 105 to generate and / or otherwise access IR image data representing a portion (e.g., a palm, a hand, a neck, a forearm, a face, a fingertip, an eye, etc.) of an authenticating user.

[0059] The method 700 includes, at block B704, detecting a hand or palm represented in a captured IR image. With respect to Fig. 1, the image capture component 120 (e.g., the palm capture network 122) may, for example, identify and / or locate a captured hand or a captured palm depicted in the captured IR image.

[0060] The method 700 tests in block B706 whether the detection accuracy is above a certain threshold. With respect to Fig. 1, the loop control component 110 may, for example, test a detection accuracy of the detected hand or palm detected in block B704. For example, the loop control component 110 may identify a prediction confidence (e.g., for the binary classification of the entire IR image predicted by the palm detection network 122, an aggregate and / or statistical value representing the prediction confidence of a plurality of per-pixel classifications, such as an average prediction confidence for pixels located within a predicted bounding box, etc.), and the loop control component 110 may use the identified prediction confidence as the detection accuracy and test the prediction confidence against a specified confidence threshold.If the detection accuracy does not meet the specified threshold, the method 700 returns to block B702 to acquire a new IR image. Otherwise, the method 700 proceeds to block B708.

[0061] The method 700 includes, in block B708, segmenting veins to capture a vein topology for the authenticating user. With respect to Fig. 1, for example, the palm vein detector 140 may segment a vein topology from the acquired IR image data and / or a portion thereof. In some embodiments, the region of interest generator 130 may identify and provide a representation of one or more subset regions (e.g., a predicted bounding box or other boundary shape) representing a detected hand or palm to the palm vein detector 140, and the palm vein detector 140 may process the detected hand or palm and / or the one or more subset regions to detect (e.g., segment) the veins represented in the detected hand or palm and / or the one or more subset regions.

[0062] The method 700 tests in block B710 whether the detection accuracy is above a certain threshold. With respect to Fig. 1, the loop control component 110 may, for example, test a detection accuracy of the detected veins segmented in block B708. The loop control component 110 may, for example, determine a prediction confidence (e.g., for the binary classification of the entire input IR image predicted by the vein detection network 144, an aggregated and / or statistical value representing the prediction confidence of a plurality of classifications per pixel, such as an average prediction confidence for the pixels within the tested region, etc.), and the loop control component 110 may use the identified prediction confidence as the detection accuracy and test the prediction confidence against a specified confidence threshold. If the detection accuracy does not meet the specified threshold, the method 700 returns to block B702 to acquire a new IR image.Otherwise, the method 700 proceeds to block B712.

[0063] The method 700 includes, in block B712, a comparison of the detected and the reference vein topology. With respect to Fig. 1, for example, the input generator 152 may apply a representation of the sensed and reference palm vein topology to the palm vein comparison network 154 to generate a representation of whether corresponding regions of the sensed and reference palm vein topology are predicted to match, such as a correspondence map that quantifies the probability that corresponding regions of the sensed and reference palm vein topology will match.

[0064] The method 700 includes, in block B714, determining whether the detected and reference vein topologies match. With respect to Fig. 1, for example, the matching component 156 may quantify the number of matching regions by applying a confidence threshold (e.g., 85%) to each element of the predicted output (e.g., a predicted correspondence map). Thus, the matching component 156 may determine the number of matching regions between the acquired and reference palm vein topologies, and the matching component 156 may compare this number to a specified threshold. If the number of matching regions does not meet the specified threshold, the method 700 may return to block B712 to test the acquired palm vein topology against another reference vein topology (if there are any remaining to be tested). If the number of matching regions meets the specified threshold, the method 700 includes determining that the user is authenticated at block B716.

[0065] The systems and methods described herein may be used, without limitation, by non-autonomous vehicles, semi-autonomous vehicles (e.g., in one or more adaptive driver assistance systems (ADAS)), guided and unguided robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, flying vessels, boats, shuttles, emergency vehicles, motorcycles, electric or motorized bicycles, aircraft, construction vehicles, trains, underwater vehicles, remotely operated vehicles such as drones, and / or other types of vehicles.Furthermore, the systems and methods described herein may be used for a variety of purposes, including, but not limited to, machine control, machine locomotion, machine propulsion, synthetic data generation, model training, perception, augmented reality, virtual reality, mixed reality, robotics, security and surveillance, simulation and digital twinning, autonomous or semi-autonomous machine applications, deep learning, environmental simulation, object or actor simulation and / or digital twinning, data center processing, conversational AI, light transport simulation (e.g., ray tracing, path tracing, etc.), collaborative content creation for 3D assets, cloud computing, and / or other suitable applications.

[0066] The embodiments presented here may include a variety of different systems, such as automotive systems (e.g., a control system for an autonomous or semi-autonomous machine, a perception system for an autonomous or semi-autonomous machine), systems implemented with a robot, aviation systems, media systems, boat systems, intelligent area surveillance systems, systems for performing deep learning operations, systems for performing simulation operations, systems for performing digital twinning operations, systems implemented using an edge device, systems containing one or more virtual machines (VMs), systems for performing operations for generating synthetic data, systems implemented at least partially in a data center, systems for performing AI operations in dialogue, systems for performing light transport simulations,Systems for performing collaborative content creation for 3D assets, systems implemented at least partially using cloud computing resources, and / or other types of systems. EXEMPLARY AUTONOMOUS VEHICLE

[0067] Fig. 8A is an illustration of an example autonomous vehicle 800, according to some embodiments of the present disclosure. The autonomous vehicle 800 (alternatively referred to herein as "vehicle 800") may include, without limitation, a passenger vehicle, such as a car, a truck, a bus, an emergency service vehicle, a shuttle, an electric or motorized bicycle, a motorcycle, a fire engine, a police vehicle, an ambulance, a boat, a construction vehicle, an underwater vehicle, a robotic vehicle, a drone, an aircraft, a vehicle coupled to a trailer (e.g., a semi-trailer used to transport cargo), and / or another type of vehicle (e.g., that is unmanned and / or accommodates one or more passengers).Autonomous vehicles are generally described in terms of levels of automation 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, published June 15, 2018, Standard No. J3016-201609, published September 30, 2016, and prior and future versions of this standard). The vehicle 800 may exhibit functionality according to one or more of the Level 3 - Level 5 autonomous driving levels. The vehicle 800 may exhibit functionality according to one or more of the Level 1 - Level 5 autonomous driving levels.For example, depending on the embodiment, the vehicle 800 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 "autonomous" as used herein may include any and / or all types of autonomy for the vehicle 800 or other machine, such as fully autonomous, highly autonomous, conditionally autonomous, partially autonomous, assisted autonomy, semi-autonomous, primarily autonomous, or another designation.

[0068] The vehicle 800 may include components such as a chassis, a vehicle body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other components of a vehicle. The vehicle 800 may include a propulsion system 850, such as an internal combustion engine, a hybrid electric power plant, a pure electric motor, and / or another type of propulsion system. The propulsion system 850 may be connected to a drivetrain of the vehicle 800, which may include a transmission to enable propulsion of the vehicle 800. The propulsion system 850 may be controlled in response to receiving signals from the throttle or accelerator 852.

[0069] A steering system 854, which may include a steering wheel, may be used to steer the vehicle 800 (e.g., along a desired path or route) when the propulsion system 850 is operating (e.g., when the vehicle is moving). The steering system 854 may receive signals from a steering actuator 856. The steering wheel may be optional for full automation (Level 5).

[0070] The brake sensor system 846 may be used to apply the vehicle brakes in response to receiving signals from the brake actuators 848 and / or the brake sensors.

[0071] The one or more controllers 836 that control one or more systems on chips (SoCs) 804 ( Fig. 8C) and / or GPUs may provide signals (e.g., representative of instructions) to one or more components and / or systems of the vehicle 800. For example, the one or more controllers may send signals to apply the vehicle brakes via one or more brake actuators 848, to apply the steering system 854 via one or more steering actuators 856, to apply the propulsion system 850 via one or more throttle / accelerator devices 852. The one or more controllers 836 may include one or more built-in (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and issue operational commands (e.g., signals representing commands) to enable autonomous driving and / or to assist a human driver in driving the vehicle 800.The one or more controllers 836 may include a first controller 836 for autonomous driving functions, a second controller 836 for functional safety functions, a third controller 836 for artificial intelligence functions (e.g., computer vision), a fourth controller 836 for infotainment functions, a fifth controller 836 for emergency redundancy, and / or other controllers. In some examples, a single controller 836 may perform two or more of the above functionalities, two or more controllers 836 may perform a single functionality, and / or any combination thereof.

[0072] The one or more controllers 836 may provide the signals to control one or more components and / or systems of the vehicle 800 in response to sensor data received from one or more sensors (e.g., sensor inputs). The sensor data may be received, for example and without limitation, from one or more of the following: global navigation satellite systems (“GNSS”) sensor(s) 858 (e.g., global positioning system sensor(s)), RADAR sensor(s) 860, ultrasonic sensor(s) 862, LIDAR sensor(s) 864, inertial measurement unit (“IMU”) sensor(s) 866 (e.g., accelerometer(s), gyroscope(s), magnetic compass(es), magnetometer(s), etc.), microphone(s) 896, stereo camera(s) 868, wide-angle camera(s) 870 (e.g., fisheye cameras), infrared camera(s) 872, ambient camera(s) 874 (e.g.,360-degree cameras), long-range and / or medium-range camera(s) 898, speed sensor(s) 844 (e.g., for measuring the speed of the vehicle 800), vibration sensor(s) 842, steering sensor(s) 840, brake sensor(s) (e.g., as part of the brake sensor system 846), one or more sensors of the occupant monitoring system (OMS) 801 (e.g., one or more interior cameras), and / or other sensor types.

[0073] One or more of the controllers 836 may receive inputs (e.g., in the form of input data) from an instrument cluster 832 of the vehicle 800 and provide outputs (e.g., in the form of output data, display data, etc.) via a human-machine interface (HMI) display 834, an audible annunciator, a speaker, and / or via other components of the vehicle 800. The outputs may include information such as vehicle speed, rpm, time, map data (e.g., the high-definition (“HD”) map 822 of Fig. 8C), location data (e.g., the location of the vehicle 800, e.g., on a map), direction, location of other vehicles (e.g., an occupancy grid), information about objects and status of objects as perceived by the one or more controllers 836, etc. For example, information about the presence of one or more objects (e.g., a road sign, a warning sign, a changing traffic light, etc.) and / or information about maneuvers that the vehicle has performed, is currently performing, or will perform (e.g., change lanes now, take exit 34B in two miles, etc.) may be displayed on the HMI display 834.

[0074] The vehicle 800 further includes a network interface 824 that may utilize one or more wireless antennas 826 and / or modems for communication over one or more networks. The network interface 824 may, for example, be capable of communication via 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), etc. The one or more wireless antennas 826 may also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using local area networks such as Bluetooth, Bluetooth Low Energy (LE), Z-Wave, ZigBee, etc.and / or Low Power Wide Area Networks (“LPWANs”) such as LoRaWAN, SigFox, etc.

[0075] Fig. Figure 8B is an example of camera locations and fields of view for the example autonomous vehicle 800 of Fig. 8A, according to some embodiments of the present disclosure. The cameras and the respective fields of view are an example of an embodiment and are not intended to be limiting. For example, additional and / or alternative cameras may be included and / or the cameras may be located at various locations on the vehicle 800.

[0076] The camera types for the cameras may include, but are not limited to, digital cameras that may be configured for use with the components and / or systems of the vehicle 800. The one or more cameras may operate at Automotive Safety Integrity Level (ASIL) B and / or another ASIL. The camera types may have any image capture rate, such as 60 frames per second (fps), 120 fps, 240 fps, etc., depending on the embodiment. The cameras may use rolling shutters, global shutters, another type of shutter, or a combination thereof.In some examples, the color filter array may include a red clear clear clear (RCCC) color filter array, a red clear clear blue (RCCB) color filter array, a red blue green clear (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 an RCCC, an RCCB, and / or an RBGC color filter array, may be used to increase light sensitivity.

[0077] 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 multifunction mono camera can be installed to provide features such as lane departure warning, traffic sign assist, and intelligent headlight control. One or more of the cameras (e.g., all cameras) can simultaneously record and provide image data (e.g., video).

[0078] One or more of the cameras may be mounted in a bracket, e.g., a specially designed (three-dimensional (“3D”) printed) bracket, to eliminate stray light and reflections from inside the vehicle (e.g., reflections from the dashboard reflected in the windshield mirrors) that could interfere with the camera's image data acquisition. With reference to the mounting of exterior mirrors, the exterior mirrors may be custom 3D printed so that the camera mounting plate is adapted to the shape of the exterior mirror. In some examples, the one or more cameras may be integrated into the exterior mirror. For side-facing cameras, the one or more cameras may also be integrated into the four pillars at each corner of the cabin.

[0079] Cameras with a field of view that includes portions of the environment in front of the vehicle 800 (e.g., forward-facing cameras) can be used for the surrounding view to help identify forward paths and obstacles, and to provide information critical to establishing an occupancy grid and / or determining preferred vehicle paths with the assistance of one or more controllers 836 and / or control SoCs. Forward-facing cameras can be used to perform many of the same ADAS functions as LIDAR, including emergency braking, pedestrian detection, and collision avoidance. Forward-facing cameras can also be used for ADAS features and systems that include lane departure warnings (LDW), autonomous cruise control (ACC), and / or other features such as traffic sign detection.

[0080] A variety of cameras can be used in a forward-facing configuration, including, for example, a monocular camera platform that incorporates a complementary metal oxide semiconductor (CMOS) color imager. Another example is the wide-angle cameras 870, which can be used to capture objects that enter the field of view from the periphery (e.g., pedestrians, crossing vehicles, or bicycles). Although Fig. 8B illustrates only one wide-angle camera, the vehicle 800 may include any number (including zero) of wide-angle cameras 870. Furthermore, any number of long-range cameras 898 (e.g., a long-range stereo camera pair) may be used for depth-based object detection, particularly for objects for which a neural network has not yet been trained. The one or more long-range cameras 898 may also be used for object detection and classification, as well as basic object tracking.

[0081] Any number of stereo cameras 868 may also be included in a forward-facing configuration. In at least one embodiment, one or more of the stereo cameras 868 may include an integrated control unit comprising a scalable processing unit that can provide programmable logic ("FPGA") and a multi-core microprocessor with an integrated controller area network ("CAN") or Ethernet interface on a single chip. Such a unit can be used to create a 3D map of the vehicle's surroundings that includes a distance estimate for all points in the image. An alternative stereo camera 868 may include a compact stereo vision sensor that may include two camera lenses (one each on the left and right) and an image processing chip that can measure the distance between the vehicle and the target object and process the generated information (e.g.,metadata) to activate the autonomous emergency braking and lane departure warning functions. Other types of stereo cameras 868 may be used in addition to or alternatively to those described here.

[0082] Cameras with a field of view that includes portions of the environment to the side of the vehicle 800 (e.g., side cameras) may be used for the environmental view and provide information used to create and update the occupancy grid and to generate collision warnings in the event of a side impact. For example, the one or more environmental cameras 874 (e.g., four environmental cameras 874, as in Fig. 8B) may be positioned on the vehicle 800. The one or more surround-view cameras 874 may include one or more wide-angle cameras 870, one or more fisheye cameras, one or more 360-degree cameras, and / or the like. For example, four fisheye cameras may be mounted on the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle may utilize three surround-view cameras 874 (e.g., left, right, and rear) and utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround-view camera.

[0083] Cameras with a field of view that includes portions of the environment behind the vehicle 800 (e.g., rearview cameras) may be used for parking assistance, surround view, rear collision warnings, and occupancy grid creation and updating. A variety of different cameras may be used, including, but not limited to, cameras that are also suitable as one or more forward-facing cameras (e.g., one or more long-range and / or medium-range cameras 898, one or more stereo cameras 868, one or more infrared cameras 872, etc.), as described herein.

[0084] Cameras with a field of view including portions of the interior environment within the cabin of the vehicle 800 (e.g., one or more OMS sensors 801) may be used as part of an occupant monitoring system (OMS), such as, but not limited to, a driver monitoring system (DMS). For example, OMS sensors (e.g., the one or more OMS sensors 801) may be used (e.g., by the one or more controllers 836) to track an occupant's and / or driver's gaze direction, head posture, and / or blinking. This gaze information may be used to determine the occupant's or driver's level of attention (e.g., to detect drowsiness, fatigue, and / or distraction) and / or to take appropriate actions to prevent harm to the occupant or driver.In some embodiments, data from OMS sensors may be used to enable gaze-controlled operations initiated by the driver and / or other occupants, such as adjusting cabin temperature and / or airflow, opening and closing windows, controlling cabin lighting, controlling entertainment systems, adjusting mirrors, adjusting seat positions, and / or other operations. In some embodiments, an OMS may be used for applications such as determining whether items and / or occupants have been left behind in a vehicle cabin (e.g., by detecting the presence of occupants after the driver has exited the vehicle).

[0085] Fig. 8C is a block diagram of an example system architecture for the example autonomous vehicle 800 of Fig. 8A, according to some embodiments of the present disclosure; It should be understood that these and other arrangements described herein are set forth as examples only. Other arrangements and elements (e.g., engines, interfaces, functions, arrangements, groupings of functions, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities that may be implemented as individual or distributed components, or in conjunction with other components, and in any suitable combination and location. Various functions performed by entities described herein may be performed by hardware, firmware, and / or software. Various functions may be performed, for example, by a processor executing instructions stored in memory.

[0086] Each of the vehicle’s components, features and systems 800 in Fig. 8C is illustrated as being connected via bus 802. Bus 802 may include a controller area network (CAN) data interface (alternatively referred to herein as a "CAN bus"). A CAN may be a network within vehicle 800 that is used to support the control of various features and functions of vehicle 800, such as the application of brakes, acceleration, braking, steering, windshield wipers, etc. A CAN bus may be configured to have dozens or even hundreds of nodes, each with its own unique identifier (e.g., a CAN ID). The CAN bus may be read to determine steering wheel angle, vehicle speed, engine speed (RPM), button positions, and / or other vehicle status indications. The CAN bus may be ASIL B compliant.

[0087] Although bus 802 is described herein as a CAN bus, this is not intended to be a limitation. For example, FlexRay and / or Ethernet may be used in addition to or alternatively to the CAN bus. Furthermore, while a single wire is used to represent bus 802, this is not intended to be a limitation. For example, there may be any number of buses 802, 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 that use a different protocol. In some examples, two or more buses 802 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 802 may be used for collision avoidance functionality and a second bus 802 may be used for actuation control.In each example, each bus 802 may communicate with one of the components of the vehicle 800, and two or more buses 802 may communicate with the same components. In some examples, each SoC 804, each controller 836, and / or each computer within the vehicle may have access to the same input data (e.g., inputs from sensors of the vehicle 800) and may be connected to a common bus, such as the CAN bus.

[0088] The vehicle 800 may include one or more controllers 836 as described herein with respect to Fig. 8A. The one or more controllers 836 may be used for a variety of functions. The one or more controllers 836 may be coupled to any of the various other components and systems of the vehicle 800 and may be used for control of the vehicle 800, artificial intelligence of the vehicle 800, infotainment for the vehicle 800, and / or the like.

[0089] The vehicle 800 may include a system (or more systems) on a chip (SoC) 804. The SoC 804 may include one or more CPUs 806, one or more GPUs 808, one or more processors 810, one or more caches 812, one or more accelerators 814, one or more data memories 816, and / or other unillustrated components and features. The one or more SoCs 804 may be used to control the vehicle 800 in a variety of platforms and systems. For example, the one or more SoCs 804 in a system (e.g., the system of the vehicle 800) may be combined with an HD card 822 that may be accessed via a network interface 824 from one or more servers (e.g., the one or more servers 878 of Fig. 8D) can receive map refreshes and / or updates.

[0090] The one or more CPUs 806 may include a CPU cluster or CPU complex (alternatively referred to herein as a "CCPLEX"). The one or more CPUs 806 may include multiple cores and / or L2 caches. For example, in some embodiments, the one or more CPUs 806 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the one or more CPUs 806 may include four dual-core clusters, each cluster having a dedicated L2 cache (e.g., a 2 MB L2 cache). The one or more CPUs 806 (e.g., the CCPLEX) may be configured to support concurrent operation of clusters such that any combination of the clusters of the one or more CPUs 806 may be active at any given time.

[0091] The one or more CPUs 806 may implement power management features that include one or more of the following: individual hardware blocks may be automatically clocked 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-locked; each core cluster may be independently clocked if all cores are clock-locked or power-locked; and / or each core cluster may be independently power-locked if all cores are power-locked. The one or more CPUs 806 may further implement an enhanced power state management algorithm in which allowable power states and expected wake-up times are established, and the hardware / microcode determines the best power state to enter for the core, cluster, and CCPLEX.The processing cores can support simplified sequences for entering the power state into software, offloading the work to the microcode.

[0092] The one or more GPUs 808 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The one or more GPUs 808 may be programmable and may be efficient for parallel workloads. The one or more GPUs 808 may, in some examples, use an extended Tensor instruction set. The one or more GPUs 808 may include one or more streaming microprocessors, where each streaming microprocessor may include an L1 cache (e.g., an L1 cache with at least 96 KB of memory capacity), and two or more of the streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512 KB of memory capacity). In some embodiments, the one or more GPUs 808 may include at least eight streaming microprocessors. The one or more GPUs 808 may use one or more application programming interfaces (APIs) for computations.In addition, the one or more GPUs 808 may use one or more parallel computing platforms and / or programming models (e.g., CUDA from NVIDIA).

[0093] The one or more GPUs 808 may be power-optimized for best performance in automotive and embedded use cases. For example, the one or more GPUs 808 may be fabricated on a fin field-effect transistor (FinFET). However, this is not a limitation, and the one or more GPUs 808 may also be fabricated using other semiconductor manufacturing processes. Each streaming microprocessor may include a number of mixed-precision processing cores divided into multiple blocks. For example, and without limitation, 64 PF32 cores and 32 PF64 cores may be divided into four processing blocks. In such an example, each processing block can be assigned 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two mixed-precision NVIDIA TENSOR COREs for deep learning matrix arithmetic, an L0 instruction cache, a warp scheduler, a dispatch unit, and / or a 64 KB register file.In addition, the streaming microprocessors can include independent parallel integer and floating-point datapaths to enable efficient execution of workloads with a mix of computations and addressing calculations. The streaming microprocessors can include an independent thread scheduling function to enable fine-grained synchronization and cooperation between parallel threads. The streaming microprocessors can include a combined L1 data cache and shared memory unit to improve performance while simplifying programming.

[0094] The one or more GPUs 808 may include high-bandwidth memory (HBM) and / or a 16 GB HBM2 memory subsystem to provide, in some examples, a peak memory bandwidth of approximately 900 GB / second. In some examples, synchronous graphics random-access memory (SGRAM), e.g., double data rate synchronous graphics random-access memory (GDDR5), may be used in addition to or alternatively to the HBM memory.

[0095] The one or more GPUs 808 may include unified memory technology that includes access counters to enable more accurate migration of memory pages to the processor that accesses them most frequently, thereby improving efficiency for processor-shared memory areas. In some examples, Address Translation Services (ATS) support may be used to allow the one or more GPUs 808 to directly access the page tables of the one or more CPUs 806. In such examples, if the memory management unit (MMU) of the one or more GPUs 808 experiences a failure, an address translation request may be sent to the one or more CPUs 806.In response, the one or more CPUs 806 may look up the virtual-to-physical mapping for the address in their page tables and send the translation back to the one or more GPUs 808. Thus, the unified memory technology may enable a single unified virtual address space for the memory of both the one or more CPUs 806 and the one or more GPUs 808, thereby simplifying the programming of the one or more GPUs 808 and the porting of applications to the one or more GPUs 808.

[0096] Additionally, the one or more GPUs 808 may include an access counter that can track the frequency of access by the one or more GPUs 808 to memory of other processors. The access counter can help move memory pages to the physical memory of the processor that accesses the pages most frequently.

[0097] The one or more SoCs 804 may include any number of caches 812, including those described herein. For example, the one or more caches 812 may include an L3 cache available to both the one or more CPUs 806 and the one or more GPUs 808 (e.g., connected to both the one or more CPUs 806 and the one or more GPUs 808). The one or more caches 812 may include a write-back cache that may track line states, e.g., by using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). The L3 cache may include 4 MB or more, depending on the embodiment, although smaller cache sizes may be used.

[0098] The one or more SoCs 804 may include one or more arithmetic logic units (ALUs) that may be utilized in performing processing related to any of the various tasks or operations of the vehicle 800, such as processing DNNs. Additionally, the one or more SoCs 804 may include one or more floating-point units (FPUs)—or other mathematical co-processors or numerical co-processors—for performing mathematical operations within the system. For example, the one or more SoCs 804 may include one or more FPUs integrated as execution units within one or more CPUs 806 and / or one or more GPUs 808.

[0099] The one or more SoCs 804 may include one or more accelerators 814 (e.g., hardware accelerators, software accelerators, or a combination thereof). The one or more SoCs 804 may, for example, include a hardware acceleration 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 acceleration cluster to accelerate neural networks and other computations. The hardware acceleration cluster may be used to complement the one or more GPUs 808 and offload some of the tasks of the one or more GPUs 808 (e.g., to free up more cycles of the one or more GPUs 808 to perform other tasks). The one or more accelerators 814 may, for example, be used for targeted workloads (e.g.,Perception, convolutional neural networks (CNNs), etc.) that are stable enough to be suitable for acceleration can be used. The term "CNN" as used here can include all types of CNNs, including region-based or regional convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0100] The one or more accelerators 814 (e.g., the hardware acceleration cluster) may include a deep learning accelerator (DLA). The one or more DLAs may include one or more tensor processing units (TPUs) configured to provide an additional tens of trillion operations per second for deep learning applications and inferencing. The TPUs may be accelerators configured and optimized to perform image processing functions (e.g., for CNNs, RCNNs, etc.). The one or more DLAs may also be optimized for a specific set of neural network types and floating-point operations, as well as for inferencing. The design of the one or more DLAs may provide more performance per millimeter than a general-purpose GPU, far exceeding the performance of a CPU.The one or more TPUs can perform multiple functions, including a single-instance convolution function that supports, for example, INT8, INT16, and FP16 data types for both features and weights, as well as post-processing functions.

[0101] The one or more DLAs can quickly and efficiently execute neural networks, particularly CNNs, on processed or unprocessed data for a variety of functions, including, for example and without limitation: a CNN for object identification and detection using data from camera sensors; a CNN for distance estimation using data from camera sensors; a CNN for emergency vehicle detection and identification using data from microphones; a CNN for facial recognition and vehicle owner identification using data from camera sensors; and / or a CNN for safety and / or security events.

[0102] The one or more DLAs can perform any function of the one or more GPUs 808, and by using an inference accelerator, for example, a developer can dedicate either the one or more DLAs or the one or more GPUs 808 to each function. For example, the developer can focus the processing of CNNs and floating-point operations on the one or more DLAs and leave other functions to the one or more GPUs 808 and / or other accelerators 814.

[0103] The one or more accelerators 814 (e.g., the hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may also be referred to herein as a computer vision accelerator. The one or more PVAs 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 one or more PVAs may provide a balance between performance and flexibility. Each PVA may, for example and without limitation, include any number of reduced instruction set computer (RISC) cores, direct memory access (DMA) cores, and / or any number of vector processors.

[0104] The RISC cores may interact with image sensors (e.g., the image sensors of one of the cameras described herein), image signal processors, and / or the like. Each of the RISC cores may include any amount of memory. The RISC cores may use any number of protocols, depending on the embodiment. In some examples, the RISC cores may execute a real-time operating system (RTOS). The RISC cores may be implemented with one or more integrated circuits, application-specific integrated circuits (ASICs), and / or memory devices. The RISC cores may include, for example, an instruction cache and / or tightly coupled RAM.

[0105] The DMA may enable components of the PVA(s) to access the system's memory independently of the one or more CPUs 806. The DMA may support any number of features designed to optimize the PVA, including, but not limited to, support for multi-dimensional addressing and / or circular addressing. In some examples, the DMA may support up to six or more dimensions of addressing, which may include block width, block height, block depth, horizontal block stepping, vertical block stepping, and / or depth stepping.

[0106] The vector processors may be programmable processors that can be designed to efficiently and flexibly execute programming for computer vision algorithms and provide signal processing functions. In some examples, the PVA may include a PVA core and two vector processing subsystem partitions. The PVA core may include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripheral devices. The vector processing subsystem may operate as the primary processing engine of the PVA and may include a vector processing unit (VPU), an instruction cache, and / or a memory (e.g., VMEM). A VPU core may include a digital signal processor, such asA digital signal processor with single instruction multiple data (SIMD) and very long instruction words (VLIW). The combination of SIMD and VLIW can increase throughput and speed.

[0107] Each of the vector processors may include an instruction cache and may be coupled to dedicated memory. Therefore, in some examples, each of the vector processors may be configured to operate independently of the other vector processors. In other examples, the vector processors included in a particular PVA may be configured to use data parallelism. For example, in some embodiments, the plurality of vector processors included in a single PVA may execute the same computer vision algorithm, but on different regions of an image. In other examples, the vector processors included in a particular PVA may concurrently execute different computer vision algorithms on the same image, or even execute different algorithms on consecutive images or portions of an image.Among other things, any number of PVAs can be included in the hardware acceleration cluster, and any number of vector processors can be included in each of the PVAs. Furthermore, the one or more PVAs can contain additional memory for error correcting code (ECC) to increase the overall security of the system.

[0108] The one or more accelerators 814 (e.g., the hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high-bandwidth, low-latency SRAM to the one or more accelerators 814. In some examples, the on-chip memory may include at least 4 MB of SRAM, consisting of, for example, and without limitation, eight field-configurable memory blocks accessible 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 the DLA may access the memory through a backbone that provides the PVA and the DLA with high-speed access to the memory.The backbone may include an on-chip computer vision network connecting the PVA and DLA to the memory (e.g., using the APB).

[0109] The on-chip computer vision network can include an interface that determines that both the PVA and the DLA are delivering ready and valid signals before transmitting control signals / addresses / data. Such an interface can provide separate phases and channels for transmitting control signals / addresses / data, as well as bursty communication for continuous data transmission. This type of interface can conform to ISO 26262 or IEC 61508, although other standards and protocols can also be used.

[0110] In some examples, the one or more SoCs 804 may include a real-time ray tracing hardware accelerator, as described in U.S. Patent Application No. 16 / 101,232, filed August 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the positions and extents of objects (e.g., within a world model) to generate real-time visualization simulations, for RADAR signal interpretation, for sound propagation synthesis and / or analysis, for simulation of SONAR systems, for general wave propagation simulation, for comparison with LIDAR data for localization, and / or for other functions, and / or for other purposes. In some embodiments, one or more tree traversal units (TTUs) may be used to perform one or more operations related to ray tracing.

[0111] The one or more accelerators 814 (e.g., the hardware accelerator cluster) have a wide range of uses for autonomous driving. The PVA can be a programmable image processing accelerator that can be used for critical processing steps in ADAS and autonomous vehicles. The capabilities of the PVA are well suited to algorithmic areas that require predictable processing with low power and low latency. In other words, the PVA is well suited for semi-dense or dense regular computations, even on small datasets, that require predictable runtimes with low latency and low power. In the context of autonomous vehicle platforms, the PVAs are therefore designed to execute classical computer vision algorithms because they are efficient at object detection and operate on integer mathematics.

[0112] According to one embodiment of the technology, the PVA is used, for example, to perform computer stereo vision. In some examples, a semi-global matching-based algorithm may be used, although this is not intended as a limitation. Many applications for Level 3-5 autonomous driving require on-the-fly motion estimation or stereo matching (e.g., structure from motion, pedestrian detection, lane detection, etc.). The PVA can perform a computer stereo vision function on inputs from two monocular cameras.

[0113] In some examples, the PVA can be used to perform dense optical flow processing on raw radar data (e.g., using a 4D Fast Fourier Transform) to provide processed radar. In other examples, the PVA is used for time-of-flight depth processing, e.g., by processing raw flight data to provide processed time-of-flight data.

[0114] The DLA can be used to power any type of network to improve control and driving safety; for example, this includes a neural network that outputs a confidence value for each object detection. Such a confidence value can be interpreted as a probability or as a relative "weight" of each detection compared to other detections. This confidence value allows the system to make further decisions about which detections should be considered true positives rather than false positives. For example, the system can set a confidence threshold and consider only those detections that exceed the threshold as true positives.In an automatic emergency braking (AEB) system, false positive detections would cause the vehicle to automatically perform emergency braking, which is obviously undesirable. Therefore, only the most certain detections should be considered as triggers for AEB. The DLA may employ a neural network to regress the confidence value. The neural network may take as input at least a subset of parameters, such as the bounding box dimensions, the ground plane estimate obtained (e.g., from another subsystem), output of the momentum measurement unit (IMU) sensor 866 correlated with the orientation of the vehicle 800, distance, 3D position estimates of the object obtained by the neural network and / or other sensors (e.g., LiDAR sensors 864 or RADAR sensors 860), among others.

[0115] The one or more SoCs 804 may include one or more data stores 816 (e.g., memory). The one or more data stores 816 may be on-chip memory on the one or more SoCs 804 in which neural networks to be executed on the GPU and / or the DLA may be stored. In some examples, the one or more data stores 816 may be large enough to store multiple instances of neural networks for redundancy and security. The one or more data stores 816 may include L2 or L3 cache(s) 812. The reference to the one or more data stores 816 may include a reference to memory associated with the PVA, the DLA, and / or one or more other accelerators 814, as described herein.

[0116] The one or more SoCs 804 may include one or more processors 810 (e.g., embedded processors). The one or more processors 810 may include a boot and power management processor, which may be a dedicated processor and subsystem to handle the boot power and management functions and associated security enforcement. The boot and power management processor may be part of the boot sequence of the one or more SoCs 804 and may provide runtime power management services. The boot and power management processor may provide clock and voltage programming, assistance with system transitions to a low-power state, management of the thermals and temperature sensors of the one or more SoCs 804, and / or management of the one or more SoCs 804 power states.Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and the one or more SoCs 804 may use the ring oscillators to sense the temperatures of the one or more CPUs 806, the one or more GPUs 808, and / or the one or more accelerators 814. If it is determined that the temperatures exceed a threshold, the boot and power management processor may enter a temperature fault routine and place the one or more SoCs 804 into a lower power state and / or place the vehicle 800 into a chauffeur-to-safe-stop mode (e.g., bring the vehicle 800 to a safe stop).

[0117] The one or more processors 810 may also include a set of embedded processors that can serve as an audio processing engine. The audio processing engine may be an audio subsystem that enables full hardware support for multi-channel audio across multiple interfaces and a wide and flexible range of audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core with a digital signal processor with dedicated RAM.

[0118] The one or more processors 810 may also include an always-on processor engine that can provide the necessary hardware functions 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 (e.g., timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0119] The one or more processors 810 may also include a safety cluster engine containing a dedicated processor subsystem for safety management of automotive applications. The safety cluster engine may include two or more processor cores, tightly coupled RAM, supporting peripherals (e.g., timers, an interrupt controller, etc.), and / or routing logic. In a safety mode, the two or more cores may operate in a lockstep mode, functioning as a single core with comparison logic that captures any differences between their operations.

[0120] The one or more processors 810 may also include a real-time camera engine, which may include a dedicated processor subsystem for managing the real-time camera.

[0121] The one or more processors 810 may further include a high dynamic range signal processor, which may include an image signal processor, which is a hardware engine that is part of the camera processing pipeline.

[0122] The one or more processors 810 may include a video image compositor, which may be a processing block (e.g., implemented on a microprocessor) that implements video post-processing functions required by a video playback application to generate the final image for the player window. The video image compositor may perform lens distortion correction on the one or more wide-angle cameras 870, the one or more surround cameras 874, and / or the in-cabin surveillance camera sensors. The in-cabin surveillance camera sensor is preferably monitored by a neural network running on another instance of the enhanced SoC and configured to detect events in the cabin and respond accordingly.An in-cabin system can perform lip reading to activate cellular service and make a call, dictate emails, change the destination, activate or change the vehicle's infotainment system and settings, or enable voice-activated web browsing. Certain functions are available to the driver only when the vehicle is operating in autonomous mode and are disabled otherwise.

[0123] The video image compositor can incorporate enhanced temporal noise reduction for both spatial and temporal noise reduction. For example, when motion occurs in a video, the noise reduction weights the spatial information accordingly, reducing the weight of information provided by neighboring frames. If a frame or section of a frame contains no motion, the temporal noise reduction performed by the video image compositor can use information from the previous frame to reduce noise in the current frame.

[0124] The video image compositor may also be configured to perform stereo distortion correction on the input stereo lens images. The video image compositor may also be used for user interface design when the operating system desktop is in use and the GPU(s) 808 do not need to constantly render new surfaces. Even when the GPU(s) 808 are powered on and actively performing 3D rendering, the video image compositor may be used to offload the GPU(s) 808, thereby improving performance and responsiveness.

[0125] The one or more SoCs 804 may also include a Mobile Industry Processor Interface (MIPI) serial camera interface for receiving video and input from cameras, a high-speed interface, and / or a video input block that may be used for camera and related pixel input functions. The one or more SoCs 804 may also include one or more input / output controllers, one or more of which may be software-controlled and used to receive I / O signals that are not associated with a specific role.

[0126] The one or more SoCs 804 may further include a wide range of peripheral interfaces to enable communication with peripheral devices, audio codecs, power management, and / or other devices. The one or more SoCs 804 may be used to process data from cameras (e.g., via Gigabit Multimedia Serial Link and Ethernet), sensors (e.g., one or more LIDAR sensors 864, one or more RADAR sensors 860, etc., which may be connected via Ethernet), data from bus 802 (e.g., vehicle speed 800, steering wheel position, etc.), and data from one or more GNSS sensors 858 (e.g., connected via Ethernet or CAN bus).The one or more SoCs 804 may further include dedicated high-performance mass storage controllers, which may include their own DMA engines and which may be used to offload routine data management tasks from the one or more CPUs 806.

[0127] The one or more SoCs 804 may be an end-to-end platform with a flexible architecture spanning automation levels 3-5, thereby providing a comprehensive functional safety architecture, leveraging computer vision and ADAS techniques for diversity and redundancy, and providing a platform for a flexible, reliable driving software stack along with deep learning tools. The one or more SoCs 804 may be faster, more reliable, and even more power and space efficient than conventional systems. For example, the one or more accelerators 814, in combination with the one or more CPUs 806, the one or more GPUs 808, and the one or more data memories 816, may form a fast, efficient platform for Level 3-5 autonomous vehicles.

[0128] The technology thus offers capabilities and functions that cannot be achieved by conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using a high-level programming language, such as the C programming language, to execute a variety of processing algorithms on a wide variety of visual data. However, CPUs are often unable to meet the performance requirements of many computer vision applications, such as execution time and power consumption requirements. In particular, many CPUs are unable to execute complex object detection algorithms in real time, which is a prerequisite for in-vehicle ADAS applications and a requirement for practical Level 3-5 autonomous vehicles.

[0129] In contrast to conventional systems, the technology described here enables the simultaneous and / or sequential execution of multiple neural networks and the combination of the results to enable Level 3-5 autonomous driving functionality by providing a CPU complex, a GPU complex, and a hardware acceleration cluster. For example, a CNN running on the DLA or dGPU (e.g., the one or more GPUs 820) may include text and word recognition, allowing the supercomputer to read and understand traffic signs, including signs for which the neural network has not been specifically trained. The DLA may further include a neural network capable of identifying, interpreting, and providing semantic understanding of the sign, and passing this semantic understanding to the path planning modules running on the CPU complex.

[0130] Another example is that multiple neural networks can run simultaneously, as required for Level 3, 4, or 5 driving. For example, a warning sign reading "Caution: Flashing lights indicate black ice," along with an electric light, can be interpreted independently or jointly by multiple neural networks. The sign itself can be identified as a traffic sign by a first deployed neural network (e.g., a trained neural network), and the text "Flashing lights indicate black ice" can be interpreted by a second deployed neural network, which informs the vehicle's path-planning software (preferably running on the CPU complex) that, if flashing lights are detected, black ice is present.The turn signal can be identified by running a third neural network across multiple images, which informs the vehicle's path planning software about the presence (or absence) of turn signals. All three neural networks can run simultaneously, e.g., within the DLA and / or on one or more GPUs 808.

[0131] In some examples, a facial recognition and vehicle owner identification CNN may use data from camera sensors to identify the presence of an authorized driver and / or owner of the vehicle 800. The always-on sensor processing engine may be used to unlock the vehicle when the owner approaches the driver's door and turn on the lights, and, in security mode, to disable the vehicle when the owner exits the vehicle. In this way, the one or more SoCs 804 provide security against theft and / or carjacking.

[0132] In another example, a CNN for detecting and identifying emergency vehicles may use data from microphones 896 to detect and identify emergency vehicle sirens. Unlike conventional systems that use general classifiers to detect sirens and manually extract features, the one or more SoCs 804 use the CNN to classify environmental and urban sounds, as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to detect the relative approach speed of the emergency vehicle (e.g., by using the Doppler effect). The CNN may also be trained to identify emergency vehicles specific to the local area in which the vehicle operates, as identified by one or more GNSS sensors 858.For example, if the CNN is operating in Europe, it will attempt to detect European sirens, and if it is operating in the United States, the CNN will attempt to identify only North American sirens. Once an emergency vehicle is detected, a controller can be used to execute an emergency vehicle safety routine, slow the vehicle, pull over to the side of the road, park the vehicle, and / or idle the vehicle, using ultrasonic sensors 862, until the one or more emergency vehicles pass by.

[0133] The vehicle may include one or more CPUs 818 (e.g., one or more discrete CPUs or one or more dCPUs) that may be coupled to the one or more SoCs 804 via a high-speed connection (e.g., PCIe). The CPUs 818 may, for example, include an x86 processor. The CPUs 818 may, for example, be used to perform a variety of functions, including reconciling potentially inconsistent results between ADAS sensors and the one or more SoCs 804 and / or monitoring the status and health of the one or more controllers 836 and / or the infotainment SoC 830.

[0134] The vehicle 800 may include one or more GPUs 820 (e.g., one or more discrete GPUs or one or more dGPUs) that may be coupled to the one or more SoCs 804 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The one or more GPUs 820 may provide additional artificial intelligence capabilities, e.g., by executing redundant and / or distinct neural networks, and may be used to train and / or update neural networks based on inputs (e.g., sensor data) from sensors of the vehicle 800.

[0135] The vehicle 800 may further include the network interface 824, which may include one or more wireless antennas 826 (e.g., one or more wireless antennas for various communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 824 may be used to enable a wireless connection via the Internet to the cloud (e.g., to the one or more servers 878 and / or other network devices), to other vehicles, and / or to computing devices (e.g., passenger client devices). To communicate with other vehicles, a direct connection between the two vehicles and / or an indirect connection may be established (e.g., via networks and the Internet). Direct connections may be established via vehicle-to-vehicle communication.Vehicle-to-vehicle communication may provide vehicle 800 with information about vehicles in the vicinity of vehicle 800 (e.g., vehicles in front of, beside, and / or behind vehicle 800). This functionality may be part of a cooperative adaptive cruise control function of vehicle 800.

[0136] The network interface 824 may include an SoC that provides modulation and demodulation functions and enables the controller(s) 836 to communicate over wireless networks. The network interface 824 may include a radio frequency front-end for upconversion from baseband to radio frequency and downconversion from radio frequency to baseband. The frequency conversions may be performed using known methods and / or super-heterodyne methods. In some examples, the radio frequency 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.

[0137] The vehicle 800 may further include one or more data stores 828, which may be located off-chip (e.g., outside the SoCs 804). The one or more data stores 828 may include one or more memory elements, including RAM, SRAM, DRAM, VRAM, flash, hard drives, and / or other components and / or devices capable of storing at least one bit of data.

[0138] The vehicle 800 may further include one or more GNSS sensors 858. The one or more GNSS sensors 858 (e.g., GPS, assisted GPS sensors, differential GPS (DGPS) sensors, etc.) assist in mapping, sensing, occupancy grid creation, and / or path planning. Any number of GNSS sensors 858 may be used, including, for example and without limitation, a GPS using a USB port with an Ethernet-to-serial (RS-232) bridge.

[0139] The vehicle 800 may also include one or more RADAR sensors 860. The one or more RADAR sensors 860 may be used by the vehicle 800 for long-range vehicle detection, even in darkness and / or poor weather conditions. The functional safety level of the RADAR may be ASIL B. The one or more RADAR sensors 860 may use the CAN and / or bus 802 (e.g., to transmit the data generated by the one or more RADAR sensors 860) for control and access to object tracking data, with access to the raw data via Ethernet in some examples. A variety of RADAR sensor types may be used. The one or more RADAR sensors 860 may be suitable for, for example, front-, rear-, and side-facing RADAR without limitation. In some examples, one or more Pulse Doppler RADAR sensors are used.

[0140] The one or more RADAR sensors 860 may include various configurations, such as long range with a narrow field of view, short range with a wide field of view, short range side coverage, etc. In some examples, long range RADAR may be used for the adaptive cruise control function. Long range RADAR systems may provide a wide field of view realized by two or more independent scans, e.g., within a range of 250 m. The one or more RADAR sensors 860 may assist in distinguishing between static and moving objects and may be used by ADAS systems for emergency braking and forward collision warning. Long range RADAR sensors may include a monostatic multi-modal RADAR with multiple (e.g., six or more) fixed RADAR antennas and a high-speed CAN and FlexRay interface.In a six-antenna example, the middle four antennas can create a focused beam pattern designed to detect the surroundings of vehicle 800 at higher speeds with minimal interference from traffic in adjacent lanes. The other two antennas can expand the field of view so that vehicles entering or exiting the lane of vehicle 800 can be quickly detected.

[0141] Medium-range radar systems, for example, can have a range of up to 860 m (front) or 80 m (rear) and a field of view of up to 42 degrees (front) or 850 degrees (rear). Short-range radar systems can include, among other features, radar sensors designed for installation at both ends of the rear bumper. When installed at both ends of the rear bumper, such a radar sensor system can generate two beams that continuously monitor the blind spot area to the rear and side of the vehicle.

[0142] Short-range radar systems can be used in an ADAS system to detect blind spots and / or as lane change assistants.

[0143] The vehicle 800 may also include one or more ultrasonic sensors 862. The one or more ultrasonic sensors 862, which may be mounted on the front, rear, and / or sides of the vehicle 800, may be used for parking assistance and / or for creating and updating an occupancy grid. A plurality of ultrasonic sensors 862 may be used, and different ultrasonic sensors 862 may be used for different sensing ranges (e.g., 2.5 m, 4 m). The one or more ultrasonic sensors 862 may operate at functional safety levels of ASIL B.

[0144] The vehicle 800 may include one or more LIDAR sensors 864. The one or more LIDAR sensors 864 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The one or more LIDAR sensors 864 may conform to ASIL B functional safety level. In some examples, the vehicle 800 may include multiple LIDAR sensors 864 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to deliver data to a Gigabit Ethernet switch).

[0145] In some examples, the one or more LIDAR sensors 864 may be capable of providing a list of objects and their distances for a 360-degree field of view. For example, commercially available LIDAR sensors 864 may have an advertised range of approximately 800 m, with an accuracy of 2 cm - 3 cm, and support for an 800 Mbps Ethernet connection. In some examples, one or more non-prominent LIDAR sensors 864 may be used. In such examples, the one or more LIDAR sensors 864 may be implemented as a small device that may be embedded in the front, rear, sides, and / or corners of the vehicle 800. In such examples, the one or more LIDAR sensors 864 may provide a horizontal field of view of up to 120 degrees and a vertical field of view of up to 35 degrees, with a range of 200 m, even for objects with low reflectivity.The one or more front-mounted LIDAR sensors 864 may be configured for a horizontal field of view between 45 degrees and 135 degrees.

[0146] In some examples, LIDAR technologies such as 3D flash LIDAR may also be used. 3D flash LIDAR uses a laser flash as the transmission source to illuminate the vehicle's surroundings up to approximately 200 m. A flash LIDAR unit contains a receptor that records the time of flight of the laser pulse and the reflected light at each pixel, which in turn corresponds to the distance between the vehicle and the objects. Flash LIDAR can enable highly accurate and distortion-free images of the surroundings to be created with each laser flash. In some examples, four flash LIDAR sensors may be deployed, one on each side of the vehicle. Available 3D flash LIDAR systems incorporate a solid-state 3D star array LIDAR camera that contains no moving parts other than a fan (e.g., a non-scanning LIDAR device).The flash lidar device can use a 5-nanosecond Class I (eye-safe) laser pulse per image and collect the reflected laser light as 3D range point clouds and co-registered intensity data. By using flash lidar, and because flash lidar is a solid-state device with no moving parts, the one or more lidar sensors 864 can be less susceptible to motion blur, vibration, and / or shock.

[0147] The vehicle may also include one or more IMU sensors 866. The one or more IMU sensors 866 may, in some examples, be located at the center of the rear axle of the vehicle 800. The one or more IMU sensors 866 may, for example and without limitation, include one or more accelerometers, one or more magnetometers, one or more gyroscopes, one or more magnetic compasses, and / or other types of sensors. In some examples, such as in six-axis applications, the one or more IMU sensors 866 may include accelerometers and gyroscopes, while in nine-axis applications, the one or more IMU sensors 866 may include accelerometers, gyroscopes, and magnetometers.

[0148] In some embodiments, the one or more IMU sensors 866 may be implemented as a miniaturized, high-performance GPS-aided inertial navigation system (GPS / INS) that combines micro-electro-mechanical systems (MEMS) inertial sensors, a high-sensitivity GPS receiver, and advanced Kalman filter algorithms to provide estimates of position, velocity, and attitude. Thus, in some examples, the one or more IMU sensors 866 may enable the vehicle 800 to estimate heading without requiring input from a magnetic sensor by directly observing and correlating GPS velocity changes with the one or more IMU sensors 866.In some examples, the one or more IMU sensors 866 and the one or more GNSS sensors 858 may be combined into a single integrated unit.

[0149] The vehicle may include one or more microphones 896 mounted in and / or around the vehicle 800. The one or more microphones 896 may be used, among other things, to detect and identify emergency vehicles.

[0150] The vehicle may further include any number of camera types, including one or more stereo cameras 868, one or more wide-angle cameras 870, one or more infrared cameras 872, one or more surround cameras 874, one or more long-range and / or medium-range cameras 898, and / or other camera types. The cameras may be used to capture image data around the entire perimeter of the vehicle 800. The types of cameras used depend on the embodiments and requirements for the vehicle 800, and any combination of camera types may be used to provide the necessary coverage around the vehicle 800. Furthermore, the number of cameras may vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or a different number of cameras.The cameras may support, by way of example and without limitation, Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the one or more cameras is referred to herein with reference to . Fig. 8A and Fig. 8B is described in more detail.

[0151] The vehicle 800 may further include one or more vibration sensors 842. The one or more vibration sensors 842 may measure vibrations from components of the vehicle, such as the one or more axles. For example, changes in vibrations may indicate a change in the road surface. In another example, when two or more vibration sensors 842 are used, the differences between the vibrations may be used to determine the friction or slippage of the road surface (e.g., when the difference in vibration is between a driven axle and a free-spinning axle).

[0152] The vehicle 800 may include an ADAS system 838. The ADAS system 838 may include an SoC in some examples. The ADAS system 838 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward crash warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keep assist (LKA), blind spot warning (BSW), rear cross-traffic warning (RCTW), collision warning systems (CWS), lane centering (LC), and / or other features and functions.

[0153] The ACC systems may use one or more radar sensors 860, one or more lidar sensors 864, and / or one or more cameras. The ACC systems may include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to the vehicle immediately in front of the vehicle 800 and automatically adjusts the vehicle speed to maintain a safe distance from preceding vehicles. Lateral ACC performs follow-through and advises the vehicle 800 to change lanes if necessary. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0154] CACC utilizes information from other vehicles that may be received via the network interface 824 and / or the one or more wireless antennas 826 from other vehicles over a wireless connection or indirectly via a network connection (e.g., over the Internet). Direct connections may be provided via a vehicle-to-vehicle (V2V) communication link, while indirect connections may be an infrastructure-to-vehicle (I2V) communication link. In general, the V2V communication concept provides information about the immediately preceding vehicles (e.g., vehicles immediately in front of and in the same lane as vehicle 800), while the I2V communication concept provides information about traffic further ahead. CACC systems may include both I2V and V2V information sources.Given the information about the vehicles ahead of vehicle 800, CACC can be more reliable and has the potential to improve traffic flow and reduce congestion on the road.

[0155] FCW systems are designed to warn the driver of a hazard so they can take corrective action. FCW systems utilize a forward-facing camera and / or one or more radar sensors 860 coupled with a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to feedback to the driver, e.g., a display, speaker, and / or vibrating component. FCW systems can provide a warning, e.g., in the form of a sound, a visual warning, a vibration, and / or a rapid braking pulse.

[0156] AEB systems detect an impending forward collision with another vehicle or object and can automatically apply the brakes if the driver does not take corrective action within a specified time or distance parameter. AEB systems may utilize one or more forward-facing cameras and / or one or more radar sensors 860 coupled with a dedicated processor, DSP, FPGA, and / or ASIC. When the AEB system detects a hazard, it typically first alerts the driver to take corrective action to avoid the collision; if the driver does not take corrective action, the AEB system can automatically apply the brakes to prevent or at least mitigate the effects of the predicted collision. AEB systems may incorporate techniques such as dynamic brake support and / or imminent crash braking.

[0157] LDW systems provide visual, audible, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle 800 crosses lane markings. An LDW system will not activate if the driver indicates an intentional lane departure by activating a turn signal. LDW systems may utilize forward-facing cameras coupled with a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to feedback to the driver, e.g., a display, speaker, and / or vibrating component.

[0158] LKA systems are a variant of LDW systems. LKA systems provide steering inputs or braking to correct the vehicle 800 when the vehicle 800 begins to drift out of its lane.

[0159] BSW systems detect and warn the driver of vehicles in the car's blind spot. BSW systems may provide a visual, audible, and / or tactile warning signal to indicate that merging or changing lanes is unsafe. The system may provide an additional warning when the driver activates a turn signal. BSW systems may utilize one or more rear-facing cameras and / or radar sensors 860 coupled with a dedicated processor, DSP, FPGA, and / or ASIC electrically coupled to feedback to the driver, such as a display, speaker, and / or vibrating component.

[0160] RCTW systems can provide visual, audible, and / or tactile notification when an object outside the range of the rearview camera is detected when the vehicle is reversing. Some RCTW systems include AEB to ensure the vehicle brakes are applied to avoid a crash. RCTW systems can utilize one or more rear-facing radar sensors coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to feedback to the driver, e.g., a display, speaker, and / or vibrating component.

[0161] Conventional ADAS systems can experience false positives, which can be annoying and distracting for the driver, but are generally not catastrophic because ADAS systems warn the driver and give them the opportunity to decide whether a safety condition truly exists and act accordingly. However, in an autonomous vehicle 800, in the event of conflicting results, the vehicle 800 must decide for itself whether to consider the result of a primary computer or a secondary computer (e.g., a first controller 836 or a second controller 836). In some embodiments, the ADAS system 838 may, for example, be a backup and / or secondary computer that provides perception information to a rationality module of the backup computer.The backup computer rationality monitor can run redundant, diverse software on hardware components to detect errors in perception and dynamic driving tasks. The outputs of the ADAS system 838 can be provided to a supervising MCU. If the outputs of the primary computer and the secondary computer conflict, the supervising MCU must determine how to resolve the conflict to ensure safe operation.

[0162] In some examples, the primary computer may be configured to provide the monitoring MCU with a confidence value indicating the primary computer's confidence in the chosen outcome. If the confidence value exceeds a threshold, the monitoring MCU may follow the primary computer's instruction regardless of whether the secondary computer provides a conflicting or inconsistent result. If the confidence value does not meet the threshold and the primary and secondary computers indicate different results (e.g., a conflict), the monitoring MCU may arbitrate between the computers to determine the appropriate outcome.

[0163] The monitoring MCU may be configured to run one or more neural networks trained and configured to determine the conditions under which the secondary computer triggers false alarms based on the output from the primary and secondary computers. This allows the one or more neural networks in the monitoring MCU to learn when the output of the secondary computer can be trusted and when it cannot. For example, if the secondary computer is a RADAR-based FCW system, a neural network in the monitoring MCU can learn when the FCW system identifies metallic objects that are not actually hazardous, such as a drain grate or manhole cover, which triggers an alarm.Similarly, if the secondary computer is a camera-based LDW system, a neural network in the monitoring MCU may learn to override the LDW system when cyclists or pedestrians are present and lane departure is indeed the safest maneuver. In embodiments including one or more neural networks running on the monitoring MCU, the monitoring MCU may include at least one DLA or GPU, one or more of which is suitable for executing the one or more neural networks with associated memory. In preferred embodiments, the monitoring MCU may comprise and / or be included as a component of the one or more SoCs 804.

[0164] In other examples, the ADAS system 838 may include a secondary computer that executes the ADAS functionality according to traditional computer vision rules. Thus, the secondary computer may use classic computer vision rules (if-then), and the presence of one or more neural networks in the supervising MCU may improve reliability, safety, and performance. For example, the diverse implementation and intentional non-identity makes the overall system more fault-tolerant, especially against errors caused by software (or software-hardware interfaces).For example, if a software bug or error occurs in the software on the primary computer and the non-identical software code on the secondary computer produces the same overall result, the monitoring MCU can have greater confidence that the overall result is correct and the bug in the software or hardware on the primary computer does not cause a significant error.

[0165] In some examples, the output of the ADAS system 838 may be fed into the perception block of the primary computer and / or the dynamic driving task block of the primary computer. For example, if the ADAS system 838 displays a forward collision warning due to an object immediately ahead, the perception block may use this information in identifying objects. In other examples, the secondary computer may have its own neural network trained to reduce the risk of false positives, as described herein.

[0166] The vehicle 800 may further include the infotainment SoC 830 (e.g., an in-vehicle infotainment (IVI) system). Although illustrated and described as an SoC, the infotainment system may not be an SoC and may include two or more discrete components. The infotainment SoC 830 may include a combination of hardware and software that can be used to provide audio (e.g., music, a personal digital assistant, navigation directions, news, radio, etc.), video (e.g., TV, movies, streaming, etc.), phone (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.), and / or information services (e.g., navigation systems, rear parking assist, a radio data system, vehicle-related information such as fuel level, total distance traveled, brake fuel level, oil level, door open / close, air filter information, etc.) to the vehicle 800.The infotainment SoC 830 may, for example, include radios, record players, navigation systems, video players, USB and Bluetooth connectivity, car computers, in-car entertainment, Wi-Fi, steering wheel audio controls, hands-free calling, a heads-up display (HUD), an HMI display 834, a telematics device, a control panel (e.g., for controlling and / or interacting with various components, functions, and / or systems), and / or other components. The infotainment SoC 830 may further be used to provide information (e.g., visual and / or audible) to one or more users of the vehicle, such as information from the ADAS system 838, autonomous driving information such as planned vehicle maneuvers, road layouts, environmental information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0167] The infotainment SoC 830 may include GPU functionality. The infotainment SoC 830 may communicate with other devices, systems, and / or components of the vehicle 800 via the bus 802 (e.g., CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 830 may be coupled to a supervisory MCU so that the GPU of the infotainment system may perform some self-driving functions if the one or more primary controllers 836 (e.g., the primary and / or backup computers of the vehicle 800) fail. In such an example, the infotainment SoC 830 may place the vehicle 800 into a chauffeur-to-safe-stop mode, as described herein.

[0168] The vehicle 800 may further include an instrument cluster 832 (e.g., a digital instrument panel, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 832 may include a controller and / or supercomputer (e.g., a discrete controller or supercomputer). The instrument cluster 832 may include a number of instruments, such as a speedometer, fuel level, oil pressure, tachometer, odometer, turn signals, shift position indicator, seat belt warning light(s), parking brake warning light(s), engine malfunction light(s), airbag system (SRS) information, lighting controls, safety system controls, navigation information, etc. In some examples, information from the infotainment SoC 830 and the instrument cluster 832 may be displayed and / or shared. In other words, the instrument cluster 832 may be included as part of the infotainment SoC 830, or vice versa.

[0169] Fig. 8D is a system diagram for communication between the one or more cloud-based servers and the example autonomous vehicle 800 of Fig. 8A, according to some embodiments of the present disclosure. The system 876 may include the one or more servers 878, the one or more networks 890, and the vehicles, including the vehicle 800. The server(s) 878 may include a plurality of GPUs 884(A)-884(H) (collectively referred to herein as GPUs 884), PCIe switches 882(A)-882(H) (collectively referred to herein as PCIe switches 882), and / or CPUs 880(A)-880(B) (collectively referred to herein as CPUs 880). The GPUs 884, the CPUs 880, and the PCIe switches may be interconnected with high-speed interconnects such as, without limitation, the NVIDIA-developed NVLink interfaces 888 and / or PCIe interconnects 886. In some examples, the GPUs 884 are connected via NVLink and / or NVSwitch SoC, and the GPUs 884 and the PCIe switches 882 are connected via PCIe links.Although eight GPUs 884, two CPUs 880, and two PCIe switches are illustrated, this is not intended to be limiting. Depending on the embodiment, each of the servers 878 may include any number of GPUs 884, CPUs 880, and / or PCIe switches. For example, the one or more servers 878 may each include eight, sixteen, thirty-two, and / or more GPUs 884.

[0170] The one or more servers 878 may receive, via the one or more networks 890 and from the vehicles, image data representative of images depicting unexpected or changed road conditions, such as recently commenced roadwork. The one or more servers 878 may transmit, via the one or more networks 890 and to the vehicles, neural networks 892, updated neural networks 892, and / or map information 894 containing information about traffic and road conditions. The updates to the map information 894 may include updates to the HD map 822, such as information about construction, potholes, detours, flooding, and / or other obstacles.In some examples, the neural networks 892, the updated neural networks 892, and / or the map information 894 may result from new training and / or experience represented in the data received from any number of vehicles in the environment and / or may be based on training performed in a data center (e.g., using the one or more servers 878 and / or other servers).

[0171] The one or more servers 878 may be used to train machine learning models (e.g., neural networks) based on training data. The training data may be generated by the vehicles and / or in a simulation (e.g., with a game machine). In some examples, the training data is tagged (e.g., when the neural network benefits from supervised learning) and / or undergoes other preprocessing, while in other examples, the training data is not tagged and / or preprocessed (e.g., when the neural network does not require supervised learning).Training may be performed using one or more classes of machine learning techniques, including, without limitation, supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, federated learning, transfer learning, feature learning (including principal component and cluster analysis), multilinear subspace learning, manifold learning, representation learning (including replacement dictionary learning), rule-based machine learning, anomaly detection, and any variations or combinations thereof. Once the machine learning models are trained, the machine learning models may be used by the vehicles (e.g., transmitted to the vehicles via the one or more networks 890) and / or the machine learning models may be used by the one or more servers 878 to remotely monitor the vehicles.

[0172] In some examples, the server(s) 878 may receive data from the vehicles and apply the data to real-time neural networks to perform intelligent inference in real time. The one or more servers 878 may include deep learning supercomputers and / or dedicated AI computers powered by GPUs 884, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, the one or more servers 878 may include a deep learning infrastructure using only CPU-powered data centers.

[0173] The deep learning infrastructure of the one or more servers 878 may be capable of performing fast, real-time inferencing and may utilize this capability to evaluate and verify the state of the processors, software, and / or associated hardware in the vehicle 800. For example, the deep learning infrastructure may receive periodic updates from the vehicle 800, such as a sequence of images and / or objects that the vehicle 800 has located in that sequence of images (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 the objects and compare them to the objects identified by the vehicle 800, and if the results do not match and the infrastructure concludes that the AI ​​in the vehicle 800 is not functioning properly, the one or more servers 878 may send a signal to the vehicle 800 instructing a fail-safe computer of the vehicle 800 to take control, notify the passengers, and perform a safe parking maneuver.

[0174] For inferencing, the one or more servers 878 may include GPUs 884 and one or more programmable inference accelerators (e.g., NVIDIA's TensorRT). The combination of GPU-driven servers and inference accelerators may enable real-time responsiveness. In other examples, e.g., where performance is less critical, servers powered by CPUs, FPGAs, and other processors may be used for inferencing. EXAMPLE COMPUTER DEVICE

[0175] Fig. 9 is a block diagram of an example computing device 900 suitable for use in implementing some embodiments of the present disclosure. Computing device 900 may include an interconnection system 902 that directly or indirectly couples the following devices: memory 904, one or more central processing units (CPUs) 906, one or more graphics processing units (GPUs) 908, a communications interface 910, input / output (I / O) ports 912, input / output components 914, a power supply 916, one or more presentation components 918 (e.g., displays), and one or more logic units 920. In at least one embodiment, the one or more computing devices 900 may include one or more virtual machines (VMs), and / or each of the components thereof may include virtual components (e.g., virtual hardware components).As non-limiting examples, one or more of the GPUs 908 may include one or more vGPUs, one or more of the CPUs 906 may include one or more vCPUs, and / or one or more of the logic units 920 may include one or more virtual logic units. Thus, a computing device 900 may include discrete components (e.g., a full GPU associated with the computing device 900), virtual components (e.g., a portion of a GPU associated with the computing device 900), or a combination thereof.

[0176] Although the various blocks of Figure 9 are illustrated as being connected by wires via the interconnect system 902, this is not intended to be limiting and is for clarity only. For example, in some embodiments, a presentation component 918, such as a display device, may be considered an I / O component 914 (e.g., if the display is a touchscreen). As another example, the CPUs 906 and / or GPUs 908 may include memory (e.g., memory 904 may represent a storage device in addition to the memory of the GPUs 908, the CPUs 906, and / or other components). In other words, the computing device of Fig. Figure 9 is merely illustrative. It does not distinguish between categories such as “workstation,” “server,” “laptop,” “desktop,” “tablet,” “client device,” “mobile device,” “handheld device,” “game console,” “electronic control unit (ECU),” “virtual reality system,” and / or other device or system types, since all devices within the scope of the computing device are Fig. 9 may be considered.

[0177] The interconnect system 902 may represent one or more interconnects or buses, such as an address bus, a data bus, a control bus, or a combination thereof. The interconnect system 902 may include one or more bus or interconnect types, such as an Industry Standard Architecture (ISA) bus, an Extended ISA (EISA) bus, a Video Electronics Standards Association (VESA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI Express (PCIe) bus, and / or another type of bus or interconnect. In some embodiments, there are direct connections between components. For example, the CPU 906 may be directly connected to the memory 904. Further, the CPU 906 may be directly connected to the GPU 908.For a direct or point-to-point connection between components, interconnect system 902 may include a PCIe link to establish the connection. In these examples, a PCI bus need not be included in computing device 900.

[0178] Memory 904 may include a variety of computer-readable media. The computer-readable media may be any available media accessible by the computing device 900. The computer-readable media may include both volatile and non-volatile media, as well as removable and non-removable media. By way of example and without limitation, the computer-readable media may include computer storage media and communication media.

[0179] The computer storage media may include both volatile and non-volatile media and / or removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, and / or other data types. For example, the memory 904 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 storage technologies, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage 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 900. As used herein, computer storage media does not per se include signals.

[0180] The computer storage media may embody computer-readable instructions, data structures, program modules, and / or other data types in a modulated data signal such as a carrier wave or other transport mechanism, and includes any media for conveying information. The term "modulated data signal" may refer to a signal having one or more of its characteristics adjusted or altered to encode information in the signal. The computer storage media may include, for example, and is not limited to, wired media, such as a wired network or a direct-wired connection, and wireless media, such as acoustic, RF, infrared, and other wireless media. Combinations of the above should also be included within the scope of computer-readable media.

[0181] The CPUs 906 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 900 to perform one or more of the methods and / or processes described herein. The CPUs 906 may each include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of concurrently executing a plurality of software threads. The CPUs 906 may include any type of processor and may include different types of processors depending on the type of computing device 900 implemented (e.g., processors with fewer cores for mobile devices and processors with more cores for servers).Depending on the type of computing device 900, the processor may be, for example, an Advanced RISC Machines (ARM) processor implemented with Reduced Instruction Set Computing (RISC) or an x86 processor implemented with Complex Instruction Set Computing (CISC). Computing device 900 may include one or more CPUs 906, in addition to one or more microprocessors or additional coprocessors, such as math coprocessors.

[0182] In addition to or alternatively to the one or more CPUs 906, the one or more GPUs 908 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 900 to perform one or more of the methods and / or processes described herein. One or more of the GPUs 908 may be an integrated GPU (e.g., with one or more of the CPUs 906) and / or one or more of the GPUs 908 may be a discrete GPU. In embodiments, one or more of the GPUs 908 may be a co-processor of one or more of the CPUs 906. The one or more GPUs 908 may be used by the computing device 900 to render graphics (e.g., 3D graphics) or perform general-purpose computations. The GPUs 908 may be used, for example, for general-purpose computing on GPUs (GPGPU).The one or more GPUs 908 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. The one or more GPUs 908 may generate pixel data for output images in response to rendering commands (e.g., rendering commands from the one or more CPUs 906 received via a host interface). The GPUs 908 may include graphics memory, such as display memory, for storing pixel data or other suitable data, such as GPGPU data. The display memory may be included as part of the memory 904. The GPUs 908 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 connect the GPUs via a switch (e.g., using NVSwitch).When combined, each GPU can generate 908 pixel data or GPGPU data for different sections of an output or for different outputs (e.g., a first GPU for a first image and a second GPU for a second image). Each GPU can contain its own memory or share memory with other GPUs.

[0183] In addition to, or alternatively, the one or more CPUs 906 and / or the one or more GPUs 908, the one or more logic units 920 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 900 to perform one or more of the methods and / or processes described herein. In embodiments, the CPUs 906, the GPUs 908, and / or the one or more logic units 920 may discretely or jointly execute any combination of the methods, processes, and / or portions thereof. One or more of the logic units 920 may be part of and / or integrated with one or more of the CPUs 906 and / or one or more of the GPUs 908, and / or one or more of the logic units 920 may be discrete components or otherwise external to the CPUs 906 and / or the GPUs 908.In embodiments, one or more of the logic units 920 may be a co-processor of one or more of the CPUs 906 and / or one or more of the GPUs 908.

[0184] Examples of the one or more logic units 920 include one or more processing cores and / or components thereof, such as data processing units (DPUs), tensor cores (TCs), tensor processing units (TPUs), pixel visual cores (PVCs), vision processing units (VPUs), graphics processing clusters (GPCs), texture processing clusters (TPCs), streaming multiprocessors (SMs), tree traversal units (TTUs), artificial intelligence accelerators (AIAs), deep learning accelerators (DLAs), arithmetic logic units (ALUs), application-specific integrated circuits (Application-Specific Integrated Circuits, ASICs), floating point units (FPUs),Input / Output (I / O) elements, Peripheral Component Interconnect (PCI) or PCI Express (PCIe) elements, and / or similar.

[0185] The communication interface 910 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 900 to communicate with other computers over an electronic network, including wired and / or wireless communication. The communication interface 910 may include components and functions that enable communication over a variety of networks, such as wireless networks (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), wired networks (e.g., communication over Ethernet or InfiniBand), low-power wide-area networks (e.g., LoRaWAN, SigFox, etc.), and / or the Internet.In one or more embodiments, the one or more logic units 920 and / or the communication interface 910 may include one or more data processing units (DPUs) to transfer data received over a network and / or via the interconnect system 902 directly to one or more GPUs 908 (e.g., a memory thereof).

[0186] The I / O ports 912 may enable the computing device 900 to be logically coupled to other devices, including the I / O components 914, the one or more presentation components 918, and / or other components, some of which may be built into (e.g., integrated) the computing device 900. Illustrative I / O components 914 include a microphone, a mouse, a keyboard, a joystick, a gamepad, a game controller, a satellite dish, a scanner, a printer, a wireless device, etc. The I / O components 914 may provide a natural user interface (NUI) that processes air gestures, speech, or other physiological inputs generated by a user. In some cases, the inputs may be communicated to a suitable network element for further processing.An NUI may implement any combination of speech sensing, pen sensing, facial sensing, biometric sensing, both on-screen and off-screen gesture sensing, air gestures, head and eye tracking, and touch sensing (as described in more detail below) associated with a display of computing device 900. Computing device 900 may include depth cameras, such as stereoscopic camera systems, infrared camera systems, RGB camera systems, touchscreen technology, and combinations thereof, for gesture sensing and recognition. Additionally, computing device 900 may include accelerometers or gyroscopes (e.g., as part of an inertia measurement unit (IMU)) that enable the sensing of motion. In some examples, the output of the accelerometers or gyroscopes from computing device 900 may be used to present immersive augmented reality or virtual reality.

[0187] The power supply 916 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 916 may supply power to the computing device 900 to enable operation of the components of the computing device 900.

[0188] The one or more presentation components 918 may include a display (e.g., a monitor, a touchscreen, a television monitor, a heads-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The one or more presentation components 918 may receive data from other components (e.g., the one or more GPUs 908, the one or more CPUs 906, DPUs, etc.) and output the data (e.g., as an image, video, audio, etc.). EXEMPLARY DATA CENTER

[0189] Fig. 10 illustrates an exemplary data center 1000 that may be used in at least one embodiment of the present disclosure. Data center 1000 may include a data center infrastructure layer 1010, a framework layer 1020, a software layer 1030, and / or an application layer 1040.

[0190] As in Fig. 10, the data center infrastructure layer 1010 may include a resource orchestrator 1012, clustered computing resources 1014, and node computing resources (“node CRs”) 1016(1)-1016(N), where “N” represents any positive integer. In at least one embodiment, the node C.Rs 1016(1)-1016(N) may include, but are 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 memory), storage devices (e.g., solid-state or disk drives), network input / output (NW I / O) devices, network switches, virtual machines (VMs), power modules and / or cooling modules, etc.In some embodiments, one or more Node CRs among Node CRs 1016(1)-1016(N) may correspond to a server having one or more of the computing resources mentioned above. Furthermore, in some embodiments, Node CRs 1016(1)-1016(N) may include one or more virtual components, such as vGPUs, vCPUs, and / or the like, and / or one or more of Node CRs 1016(1)-1016(N) may correspond to a virtual machine (VM).

[0191] In at least one embodiment, the grouped computing resources 1014 may include separate groupings of node CRs 1016 housed in one or more racks (not shown) or in many racks in data centers in different geographical locations (also not shown). Separate groupings of node CRs 1016 within grouped computing resources 1014 may include grouped computing, network, memory, or storage resources that may be configured or allocated to support one or more workloads. In at least one embodiment, multiple node CRs 1016, including CPUs, GPUs, DPUs, and / or other processors, may be grouped in one or more racks to provide computing resources to support one or more workloads.The one or more racks may also contain any number of power modules, cooling modules, and / or network switches in any combination.

[0192] Resource orchestrator 1012 may configure or otherwise control one or more node CRs 1016(1)-1016(N) and / or clustered computing resources 1014. In at least one embodiment, resource orchestrator 1012 may include an entity for managing the software design infrastructure (SDI) for data center 1000. Resource orchestrator 1012 may include hardware, software, or a combination thereof.

[0193] In at least one embodiment, as in Fig. 10, the framework layer 1020 may include a job scheduler 1033, a configuration manager 1034, a resource manager 1036, and / or a distributed file system 1038. The framework layer 1020 may include a framework that supports the software 1032 of the software layer 1030 and / or one or more applications 1042 of the application layer 1040. The software 1032 or the one or more applications 1042 may each include web-based service software or applications such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. The framework layer 1020 may be some type of free and open source software web application framework such as, but not limited to, Apache Spark™ (hereinafter "Spark"), which may utilize a distributed file system 1038 for processing large amounts of data (e.g., "Big Data").In at least one embodiment, job scheduler 1033 may include a Spark driver to facilitate scheduling workloads supported by different layers of data center 1000. Configuration manager 1034 may be capable of configuring different layers, such as software layer 1030 and framework layer 1020, which includes Spark and distributed file system 1038, to support processing large amounts of data. Resource manager 1036 may be capable of managing clustered or grouped computing resources allocated or assigned to support distributed file system 1038 and job scheduler 1033. In at least one embodiment, the clustered or grouped computing resources may include clustered computing resource 1014 at infrastructure layer 1010 of the data center.The resource manager 1036 may coordinate with the resource orchestrator 1012 to manage these allocated or assigned computing resources.

[0194] In at least one embodiment, the software 1032 included in software layer 1030 may include software used by at least portions of node CRs 1016(1)-1016(N), clustered computing resources 1014, and / or distributed file system 1038 of framework layer 1020. One or more types of software may include, but are not limited to, web page searching software, email virus scanning software, database software, and streaming video content software.

[0195] In at least one embodiment, the applications 1042 included in the application layer 1040 may include one or more types of applications used by at least portions of the node CRs 1016(1)-1016(N), the clustered computing resources 1014, and / or the distributed file system 1038 of the framework layer 1020. One or more types of applications may include, but are not limited to, any number of a genome application, a cognitive computing application, and a machine learning application, including training or inference software, machine learning framework software (e.g., PyTorch, TensorFlow, Caffe, etc.), and / or other machine learning applications used in connection with one or more embodiments.

[0196] In at least one embodiment, the configuration manager 1034, the resource manager 1036, and / or the resource orchestrator 1012 may implement any number and type of self-modifying actions based on any amount and type of data collected in any technically feasible manner. Self-modifying actions may relieve a data center operator of the data center 1000 from making potentially poor configuration decisions and potentially avoid underutilized and / or poorly performing sections of a data center.

[0197] Data center 1000 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 according to one or more embodiments described herein. For example, one or more machine learning models may be trained by calculating weighting parameters according to a neural network architecture using software and / or computing resources described above with respect to data center 1000.In at least one embodiment, trained or deployed machine learning models corresponding to one or more neural networks may be used to infer or predict information using the resources described above relating to the data center 1000 by using weighting parameters calculated by one or more training techniques, such as, but not limited to, those described herein.

[0198] In at least one embodiment, the data center 1000 may utilize CPUs, application-specific integrated circuits (ASICs), GPUs, FPGAs, and / or other hardware (or corresponding virtual computing resources) to perform training and / or inferencing using the resources described above. Furthermore, one or more of the software and / or hardware resources described above may be configured as a service to enable users to train or infer information, such as image capture, speech capture, or other artificial intelligence services. EXAMPLE NETWORK ENVIRONMENTS

[0199] Network environments suitable for implementing embodiments of the 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 running on one or more instances of the one or more computing devices 900 of Fig. 9 - e.g., each device may include similar components, features, and / or functionality of the one or more computing devices 900. If backend devices (e.g., servers, NAS, etc.) are implemented, the backend devices may also be included as part of a data center 1000, an example of which is described herein with reference to Fig. 10 is described in more detail.

[0200] The components of a network environment can communicate with each other over one or more networks, which can be wired, wireless, or both. The network can contain multiple networks or a network of networks. For example, the network can contain 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 a public switched telephone network (PSTN), and / or one or more private networks. If the network contains a wireless telecommunications network, components such as a base station, a communications tower, or even access points (as well as other components) can provide wireless connectivity.

[0201] Compatible network environments may include one or more peer-to-peer network environments—in which case, a server cannot be included in a network environment—and one or more client-server network environments—in which case, one or more servers can be included in a network environment. In peer-to-peer network environments, the functionality described herein with respect to one or more servers may be implemented on any number of client devices.

[0202] In at least one embodiment, a network environment may include one or more cloud-based network environments, a distributed computing environment, a combination thereof, etc. A cloud-based network environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more servers, which may include one or more core network servers and / or edge servers. A framework layer may include a framework for supporting software of a software layer and / or one or more applications of an application layer. The software or the one or more applications may each include web-based service software or applications. In embodiments, one or more of the client devices may utilize the web-based service software or applications (e.g.,by accessing the service software and / or applications through one or more application programming interfaces (APIs). The framework layer may be, but is not limited to, some type of free and open-source software web application framework that uses, for example, a distributed file system for processing large amounts of data (e.g., "Big Data").

[0203] A cloud-based network environment may provide cloud computing and / or cloud storage that performs any combination of the computing and / or data storage functions (or one or more portions thereof) described herein. Each of these various functions may be distributed from central or core servers (e.g., from one or more data centers that may be located across a state, region, country, globe, etc.) across multiple locations. When a connection to a user (e.g., a client device) is relatively close to one or more edge servers, one or more core servers may offload at least a portion of the functionality to the one or more edge servers. A cloud-based network environment may 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).

[0204] The one or more client devices may include at least some of the components, features, and functions of the one or more described herein with respect to Fig.9. By way of example, and not limitation, a client device may be embodied as a personal computer (PC), a laptop, a mobile device, a smartphone, a tablet computer, a smart watch, a wearable computer, a personal digital assistant (PDA), an MP3 player, a virtual reality headset, a global positioning system (GPS) or a global positioning device, a video player, a video camera, a surveillance device or system, a vehicle, a boat, a flying boat, a virtual machine, a drone, a robot, a handheld communication device, a hospital device, a gaming device or system, an entertainment system, a vehicle computing system, an embedded system controller, a remote control, an appliance, a consumer electronics device, a workstation, an edge device,any combination of these described devices or any other suitable device.

[0205] The disclosure of this application also includes the following numbered clauses: Clause 1. Processing unit comprising: one or more processing units for the following: Generating infrared (IR) image data representing a region of an authenticating user; Generating a representation of one or more detected subcutaneous features based at least on the segmentation of at least a portion of the IR image data, and Performing one or more actions based at least on comparing the representation of the one or more detected subcutaneous features with a corresponding representation of one or more subcutaneous reference features stored in a database. Clause 2. The processor of Clause 1, wherein the one or more processing units further trigger an IR image capture loop using an IR camera observing an interior space based on at least one of the following: detecting one or more doors to the interior space being unlocked, detecting an authorized key via a proximity sensor, or detecting a body posture indicating one or more limbs, hands, wrists, or palms of the authenticating user within a threshold distance of the IR camera. Clause 3. The processor of Clause 1 or Clause 2, wherein the one or more processing units further trigger an IR image capture loop using an IR camera observing an exterior space based on at least one of the following: detecting a request to unlock one or more doors or detecting an authorized key via a proximity sensor. Clause 4. The processor of any preceding clause, wherein the one or more processing units further execute an IR image acquisition loop configured to generate frames of the IR image data and test a detection accuracy of corresponding instances of the one or more detected subcutaneous features represented in the frames of the IR image data against a threshold confidence. Clause 5. The processor of any preceding clause, wherein the one or more processing units further execute an IR image capture loop configured to repeatedly generate instances of the IR image data, apply at least one of a hand capture and a palm capture to capture a hand region or palm region depicted in the IR image data, apply a vein capture to the hand region or the palm region, and test a capture accuracy of at least one of the hand capture, the palm capture, and the vein capture. Clause 6. The processor of any preceding clause, wherein the one or more processing units further generate a representation of a detected hand or a detected palm represented by the IR image data, normalize a pose of the detected hand or the detected palm based at least on warping the IR image data, and detect veins in one or more regions of the detected hand or the detected palm. Clause 7. The processor of any preceding clause, wherein the one or more processing units further detect a hand region of the IR image data predicted to represent a hand, identify a palm region that is a subset of the hand region, and apply the vein detection to the palm region. Clause 8. The processor of any preceding clause, wherein the representation of the one or more detected subcutaneous features comprises a binary representation of a detected vein topology of the authenticating user. Clause 9. The processor of any preceding clause, wherein the one or more detected subcutaneous features of the authenticating user comprise a vein topology of a palm, neck, forearm, finger, or eye of the authenticating user. Clause 10. The processor of any preceding clause, wherein the representation of the one or more detected subcutaneous features comprises a detected palm vein topology of the authenticating user, the corresponding representation of one or more reference subcutaneous features comprises a reference palm vein topology of an authorized user, and the matching comprises applying representations of the detected palm vein topology and the reference palm vein topology to a discriminator network. Clause 11. A processor according to any preceding clause, wherein the one or more actions comprise updating one or more of a seat position, a mirror orientation, a steering wheel position, a heating or cooling setting, an audio or volume preset, a drive mode, a console display setting, a driver assistance setting, or vehicle lighting to reflect one or more corresponding stored values ​​associated with a user profile of the authenticating user. Clause 12. A processor according to any preceding clause, wherein the one or more actions comprise unlocking one or more doors. Clause 13. A processor according to any preceding clause, wherein the processor comprises 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 conducting digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources. Clause 14. System comprising one or more processing units for performing one or more actions based at least on comparing a representation of a detected vein topology with a corresponding representation of a reference vein topology stored in a database. Clause 15. The system of Clause 14, wherein the one or more processing units further execute an IR image capture loop configured to repeatedly generate instances of IR image data, apply a hand capture or a palm capture to capture a hand region or palm region depicted in the IR image data, apply a vein capture to the hand region or the palm region to segment the captured vein topology, and test a capture accuracy of at least one of the hand capture, the palm capture, or the vein capture. Clause 16. The system of Clause 14 or Clause 15, wherein the one or more processing units further generate a representation of a detected hand or a detected palm represented by the IR image data, normalize a pose of the detected hand or the detected palm based at least on warping the IR image data, and detect detected vein topology in one or more regions of the detected hand or the detected palm. Clause 17. The system of any of clauses 14-16, wherein the one or more processing units further detect a hand region represented in IR image data that is predicted to represent a hand, identify a palm region that is a subset of the hand region, and apply the vein detection to the palm region to segment the detected vein topology. Clause 18. A system as defined in any of Clauses 14 to 17, which system includes 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 conducting digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources. Clause 19. Procedure which includes: Generating a representation of a detected vein topology of an authenticating user based at least on the segmentation of at least a portion of IR image data representing a region of the authenticating user; Authenticating the authenticating user based at least on comparing the representation of the captured vein topology with a corresponding representation of a reference vein topology stored in a database; and Perform one or more actions based on the authentication of the authenticating user. Clause 20. Procedures under Clause 19, where the procedure is carried out by 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 carrying out digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources.

[0206] The disclosure may be described in the general context of computer code or machine-usable instructions, including computer-executable instructions such as program modules executed by a computer or other machine, such as a personal data assistant or other portable device. Generally, program modules include routines, programs, objects, components, data structures, etc., and refer to code that performs specific tasks or implements specific abstract data types. The disclosure may be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. The disclosure may also be practiced in distributed computing environments where tasks are performed by remotely controlled devices interconnected by a network for communication.

[0207] As used herein, any reference to "and / or" in reference to two or more elements should be interpreted to mean only one element or a 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. Furthermore, "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.

[0208] The subject matter of the present disclosure is specifically described herein to satisfy legal requirements. However, the description itself is not intended to limit the scope of the present disclosure. Rather, the inventors have contemplated that the claimed subject matter may be embodied in other ways to include various steps or combinations of steps similar to those described herein, in conjunction with other present or future technologies. Although the terms "step" and / or "block" may be used herein to refer to various elements of the methods employed, the terms should not be interpreted to imply any particular ordering among or between the various steps disclosed herein, unless and except that the order of each step is expressly described.

[0209] It is to be understood that the aspects and embodiments described above are purely exemplary and that changes in detail may be made within the scope of the claims.

[0210] Each device, method, and feature disclosed in the description and (optionally) in the claims and drawings may be presented independently or in any suitable combination.

[0211] Reference signs in the claims are for illustrative purposes only and do not limit the scope of the claims. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] US 16 / 101,232

[0110] Cited non-patent literature

[0000] Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles (Standard No. J3016-201806, published on 15 June 2018, Standard No. J3016-201609, published on 30 September 2016

[0067]

Claims

[1] Processor comprising: one or more processing units for the following: Generating infrared (IR) image data representing a region of an authenticating user; Generating a representation of one or more detected subcutaneous features based at least on the segmentation of at least a portion of the IR image data; and Performing one or more actions based at least on comparing the representation of the one or more detected subcutaneous features with a corresponding representation of one or more subcutaneous reference features stored in a database. [2] The processor of claim 1, wherein the one or more processing units further trigger an IR image capture loop using an IR camera observing an interior space based on at least one of the following: detecting one or more doors to the interior space being unlocked, detecting an authorized key via a proximity sensor, or detecting a posture indicating one or more limbs, hands, wrists, or palms of the authenticating user within a threshold distance of the IR camera. [3] The processor of claim 1 or claim 2, wherein the one or more processing units further initiate an IR image capture loop using an IR camera observing an exterior space based on at least one of: detecting a request to unlock one or more doors or detecting an authorized key via a proximity sensor. [4] The processor of any preceding claim, wherein the one or more processing units further execute an IR image acquisition loop configured to generate frames of the IR image data and test a detection accuracy of corresponding instances of the one or more detected subcutaneous features represented in the frames of the IR image data against a threshold confidence. [5] The processor of any preceding claim, wherein the one or more processing units further execute an IR image capture loop configured to repeatedly generate instances of the IR image data, apply at least one of a hand capture or a palm capture to capture a hand region or palm region depicted in the IR image data, apply a vein capture to the hand region or the palm region, and test a capture accuracy of at least one of the hand capture, the palm capture, or the vein capture. [6] The processor of any preceding claim, wherein the one or more processing units further generate a representation of a detected hand or a detected palm represented by the IR image data, normalize a pose of the detected hand or the detected palm based at least on warping the IR image data, and detect veins in one or more regions of the detected hand or the detected palm. [7] The processor of any preceding claim, wherein the one or more processing units further detect a hand region of the IR image data predicted to represent a hand, identify a palm region that is a subset of the hand region, and apply the vein detection to the palm region. [8] A processor according to any preceding claim, wherein the representation of the one or more detected subcutaneous features comprises a binary representation of a detected vein topology of the authenticating user. [9] The processor of any preceding claim, wherein the one or more detected subcutaneous features of the authenticating user comprise a vein topology of a palm, a neck, a forearm, a finger, or an eye of the authenticating user. [10] The processor of any preceding claim, wherein the representation of the one or more detected subcutaneous features comprises a detected palm vein topology of the authenticating user, the corresponding representation of one or more reference subcutaneous features comprises a reference palm vein topology of an authorized user, and the matching comprises applying representations of the detected palm vein topology and the reference palm vein topology to a discriminator network. [11] The processor of any preceding claim, wherein the one or more actions comprise updating one or more of a seat position, a mirror orientation, a steering wheel position, a heating or cooling setting, an audio or volume preset, a drive mode, a console display setting, a driver assistance setting, or vehicle lighting to reflect one or more corresponding stored values ​​associated with a user profile of the authenticating user. [12] A processor according to any preceding claim, wherein the one or more actions comprise unlocking one or more doors. [13] Processor according to one of the preceding claims, wherein the processor comprises 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 conducting digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources. [14] System comprising one or more processing units for performing one or more actions based at least on comparing a representation of a detected vein topology with a corresponding representation of a reference vein topology stored in a database. [15] The system of claim 14, wherein the one or more processing units further execute an IR image capture loop configured to repeatedly generate instances of IR image data, apply a hand capture or a palm capture to capture a hand region or palm region depicted in the IR image data, apply a vein capture to the hand region or the palm region to segment the captured vein topology, and test a capture accuracy of at least one of the hand capture, the palm capture, or the vein capture. [16] The system of claim 14 or claim 15, wherein the one or more processing units further generate a representation of a detected hand or a detected palm represented by the IR image data, normalize a posture of the detected hand or the detected palm based at least on the warping of the IR image data, and detect the detected vein topology in one or more regions of the detected hand or the detected palm. [17] The system of any of claims 14-16, wherein the one or more processing units further detect a hand region represented in IR image data that is predicted to represent a hand, identify a palm region that is a subset of the hand region, and apply the vein detection to the palm region to segment the detected vein topology. [18] A system according to any one of claims 14-17, wherein the system comprises 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 conducting digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources. [19] Procedure comprising: Generating a representation of a detected vein topology of an authenticating user based at least on the segmentation of at least a portion of IR image data representing a region of the authenticating user; Authenticating the authenticating user based at least on comparing the representation of the captured vein topology with a corresponding representation of a reference vein topology stored in a database; and Perform one or more actions based on the authentication of the authenticating user. [20] The method of claim 19, wherein the method is performed by 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 carrying out digital twinning operations; a system for performing light transport simulations; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system for performing real-time streaming; a system for generating or presenting one or more of augmented reality content, virtual reality content, or mixed reality content; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system for generating synthetic data using AI; a system that contains one or more virtual machines (VMs); a system that is at least partially implemented in a data center; or a system implemented at least in part using cloud computing resources.

Images