Object tracking for autonomous machine applications using LiDAR data

Abstract

In various examples, obstacle detectors can use LiDAR data to track the velocity state of detected objects or obstacles. For instance, using only LiDAR data, the Iterative Closest Point (ICP) algorithm can be used to determine the current state of detected objects in the current frame, and a Kalman filter can be used to maintain the tracking state of one or more detected objects over time. An obstacle detector can be configured to estimate the velocity of one or more detected objects, compare the estimated velocity with one or more previously tracked states of previously detected objects, determine if the detected object corresponds to a previously detected object, and update the tracking state for the previously detected object with the estimated velocity.

Description

Background Technology

[0001] LiDAR (Light Detection and Ranging) can be used to detect objects with high precision. For example, a LiDAR sensor measures the time it takes for a laser (or other source) to return from an object, and the accumulation of these measurements can be used to generate a three-dimensional (3D) point cloud indicating the object in the LiDAR sensor's sensing field or field of view. Therefore, each scan of the LiDAR sensor can be used to generate a 3D point cloud, which is a single capture of readings from the LiDAR sensor in time, and subsequent LiDAR scans can be spaced out at time intervals. In one example application area, LiDAR sensors are used in autonomous or semi-autonomous vehicles and other machine types to detect potential objects near the LiDAR sensor and extending into the vicinity of the vehicle or machine. For example, LiDAR sensors can be used to identify boundary shapes corresponding to objects in the environment with high precision. However, because many types of LiDAR sensors do not capture velocity information, the boundary shapes generated from LiDAR data typically lack velocity information. Without velocity information, these boundary shapes are traditionally not used for vehicle control. For example, when the ego machine moves forward, if an object appears at a certain distance in front of the ego machine, it should be considered if the object is moving towards the ego machine, rather than if the object is moving away from the ego machine (e.g., at the same or higher speed). Therefore, due to the lack of velocity information, alternative sources of velocity information have traditionally been needed to gather sufficient information about the object's motion to make accurate and reliable control decisions.

[0002] Brief Overview

[0003] Embodiments of this disclosure relate to object tracking, at least in part, using LiDAR data for autonomous machine applications (e.g., autonomous or semi-autonomous vehicles or machines and / or robotic platforms). Systems and methods are disclosed for identifying and tracking objects by estimating velocity and comparing the estimated velocity with one or more tracked object states. Based on this comparison, a tracked object state is identified as corresponding to a specific object and updated with newly estimated velocity data.

[0004] Compared to these traditional methods, current systems and methods are capable of tracking the velocity state of detected objects or obstacles using LiDAR data. For example, using only LiDAR data, the Iterative Closest Point (ICP) algorithm can be used to determine the current state of objects detected in the current frame, and a Kalman filter can be used to maintain the tracking state of one or more objects detected over time. Embodiments of this disclosure include a system configured to estimate the velocity of one or more detected objects, compare the estimated velocity with one or more previous tracking states of previously detected objects, determine that the detected object corresponds to a previously detected object, and update the tracking state of the previously detected object with the estimated velocity. In such an arrangement, during each LiDAR scan, the system can associate each detected object with objects detected from previous LiDAR scans and update known information about those detected objects. The tracking state of each detected object can be represented using a probability distribution function (PDF) generated using, for example, a Kalman filter, where the PDF represents the probability of the detected object's velocity (and / or its position within the environment determined using the velocity). As the system analyzes new data corresponding to subsequent frames, the tracking state can be improved over time. Attached Figure Description

[0005] The system and method for object tracking using LiDAR data for autonomous machine applications are described in detail below with reference to the accompanying drawings, wherein:

[0006] Figure 1A This is a data flow diagram illustrating an object tracking system according to some embodiments of the present disclosure;

[0007] Figure 1B This is a time flow diagram illustrating a process for object tracking according to some embodiments of the present disclosure;

[0008] Figure 2A This is a graphical representation of the first set of LiDAR data at time T1 according to some embodiments of the present disclosure, showing the object to be tracked;

[0009] Figure 2B This is a graphical representation of a second set of LiDAR data at time T2 after T1, according to some embodiments of the present disclosure, showing the updated location of the object being tracked;

[0010] Figure 3 It is a graphical representation of the probability distribution corresponding to the velocity of the detected object according to some embodiments of this disclosure;

[0011] Figure 4 This is an illustration showing how a vehicle passes by a parked vehicle and how the vehicle detects a parked vehicle according to some embodiments of the present disclosure;

[0012] Figure 5A These are illustrations of example autonomous vehicles according to some embodiments of the present disclosure;

[0013] Figure 5B According to some embodiments of this disclosure Figure 5A Examples of camera positions and fields of view for autonomous vehicles;

[0014] Figure 5C According to some embodiments of this disclosure Figure 5A A block diagram of an example system architecture for an example autonomous vehicle;

[0015] Figure 5D Cloud-based servers and according to some embodiments of this disclosure Figure 5A A system diagram illustrating communication between autonomous vehicles;

[0016] Figure 6 This is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure; and

[0017] Figure 7 This is a block diagram of an example data center applicable to implementing some embodiments of this disclosure. Detailed Implementation

[0018] Systems and methods relating to object tracking using LiDAR data are disclosed. While this disclosure may relate to an example autonomous vehicle 500 (also referred to herein as "vehicle 500" or "self-machine 500"), examples thereof are described herein. Figures 5A-5D The description herein is intended to be limiting. For example, the systems and methods described herein can be, but are not limited to, non-autonomous vehicles, semi-autonomous vehicles (e.g., in one or more adaptive driver assistance systems (ADAS)), driving and non-driving robots or robot-using platforms, warehouse vehicles, off-road vehicles, vehicles connected to one or more trailers, aircraft, boats, shuttles, emergency response vehicles, motorcycles, electric or motorized bicycles, aircraft, engineering vehicles, underwater vehicles, drones, and / or other vehicle types. Furthermore, while this disclosure may describe object tracking via LiDAR data in autonomous vehicles, this is not intended to be limiting, and the systems and methods described herein can be used in augmented reality, virtual reality, mixed reality, robotics, security and surveillance, autonomous or semi-autonomous machine applications, and / or any other technological space that may use security applications.

[0019] Embodiments of this disclosure relate to tracking objects, at least in part, using LiDAR data in various applications—such as autonomous or semi-autonomous machine applications. For example, one or more LiDAR sensors can generate 3D LiDAR data representing their field of view or sensory field—for example, the LiDAR data can represent a standard spherical projection or other projection type. LiDAR sensors can continuously generate LiDAR data, resulting in multiple sets of LiDAR data—each representing a single scan or rotation of the LiDAR sensor. LiDAR datasets can be separated by time intervals, and embodiments of this disclosure can use a first (e.g., current) LiDAR dataset and compare it with tracking states indicating one or more previous LiDAR datasets, as described herein.

[0020] In the initial state, the system can receive LiDAR data (e.g., in the form of a 3D point cloud) generated using one or more LiDAR sensors associated with an autonomous or semi-autonomous vehicle or other machine. Based on the LiDAR data, the system can identify one or more objects—for example, by detecting the boundaries corresponding to each object in the LiDAR data. In this initial state, the system may not have information about which objects (if any) are moving (and if so, the speed and direction of this movement). To address this lack of information, the system can create velocity tracking states for the initially empty objects. During subsequent LiDAR data generation, the system can estimate and refine the velocity tracking states of the detected one or more objects.

[0021] In embodiments of this disclosure, the system may determine, at least in part, the estimated velocity of an object detected in a subsequent scan (“current object detection”) based on a comparison between the current object detection, one or more previous object detections detected during a previous LiDAR scan, and time intervals between the respective LiDAR scans. Thus, the current object detection may have more than one estimated velocity based on one or more previous object detections in a previous scan to which the current detection can be compared. The distance traveled between the location of the current object detection and the location of a previous object detection (based on the physical distance between these two locations) may be calculated. A coordinate transformation—e.g., to the current coordinate system of the current frame—may also be applied so that the distance traveled takes into account the self-motion of the LiDAR sensor between scans (e.g., the self-movement of the machine between consecutive LiDAR scans). The estimated velocity may then be calculated by dividing the distance traveled by the elapsed time between the current object detection and the previous object detection, based on the time interval between LiDAR sensor scans. The estimated velocity may include a magnitude component and / or a direction component.

[0022] In embodiments of this disclosure, the system can compare each estimated velocity with a corresponding previous object detection for that estimated velocity. For example, each previous object detection may have a velocity tracking state (or more generally, an object tracking state) indicating the estimated velocity of the previous object detection, if any, as improved over time. This comparison can be performed between a first object boundary from a previous scan and a second object boundary from a current scan. This comparison can be performed using an Iterative Closest Point (ICP) algorithm, which provides discrete comparison points for LiDAR data corresponding to each object boundary. The system can then use the probability distribution function (“PDF”) of the object determined from the tracking state to determine the likelihood of the estimated velocity relative to the corresponding previously detected object.

[0023] In embodiments of this disclosure, the system can determine, at least in part, that one of the previously detected objects has the highest probability of corresponding to the current object based on a comparison of each estimated velocity with its corresponding previously detected object. Initially, to reduce computation and runtime, the system can determine a minimum threshold for comparison—for example, using Mahalanobis distance. For instance, if the estimated velocity for a particular object exceeds the minimum threshold with respect to the PDF, the system can remove that object from consideration. For the remaining objects, the estimated velocity can be compared to the PDF using a cost function that represents the likelihood that the currently detected object corresponds to one or more previously detected objects. The system can then evaluate the cost function and determine which object (if any) to associate the current velocity with, to update the tracking state corresponding to the associated object.

[0024] In embodiments of this disclosure, the system can update the tracked object state corresponding to the previous object detection with the highest probability corresponding to the current object detection to generate an updated tracked object state. For example, the currently estimated velocity can be applied to the tracked object state, and a Kalman filter can be used to update the tracked object state using the estimated velocity. This updated tracked object state can then serve as a comparison basis during subsequent LiDAR scans to continue improving the tracked object state over time using the Kalman filter. Therefore, LiDAR data can be used to directly identify velocity information corresponding to objects or obstacles in the environment, which allows for the direct application of LiDAR data to the control decisions of an autonomous machine.

[0025] refer to Figure 1A , Figure 1AThis is an example object tracking system 100 (also referred to herein as "System 100") according to some embodiments of this disclosure. It should be understood that such and other arrangements described herein are presented by way of example only. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or instead of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities that can be implemented as discrete or distributed components or in combination with other components, and implemented in any suitable combination and location. The various functions described herein as being performed by entities can be performed by hardware, firmware, and / or software. For example, various functions can be performed by a processor that executes instructions stored in memory. For example, in some embodiments, System 100 may include... Figures 5A-5D An example of an autonomous vehicle 500, Figure 6 Example computing device 600 and / or Figure 7 Similar features, functions, and / or components to the example data center 700.

[0026] System 100 can be a component of any of a variety of systems, or otherwise associated with a variety of systems. For example, system 100 may include or correspond to a control system for an autonomous or semi-autonomous machine, a perception system for an autonomous or semi-autonomous machine, a system for performing simulated operations, a system for performing deep learning operations, a system implemented using edge devices, a system implemented using robots, a system containing one or more virtual machines (VMs), a system implemented at least partially in a data center, and / or a system implemented at least partially using cloud computing resources.

[0027] like Figure 1A As shown, system 100 may include one or more depth sensing sensors 102 (e.g., LiDAR sensors, RADAR sensors, ultrasonic sensors, etc.). Depth sensing sensors 102 typically include a transmitter and a receiver and may include any suitable field of view or sensing field—e.g., a wide field of view (e.g., 180 to 360 degrees)—and may be movable (e.g., rotated), in embodiments, to obtain a view of a larger area relative to depth sensing sensor 102. For example, LiDAR signals may be reflected from objects near depth sensing sensor 102. The object may be movable relative to depth sensing sensor 102, and depth sensing sensor 102 may be movable relative to a lower surface (e.g., a road on which autonomous vehicle 500 is traveling). The receiver may receive indications (directly or indirectly) of these various reflected signals, and these indications may be stored and / or transmitted as data for later analysis.

[0028] The depth sensor 102 may have a sensor controller 104, which can be used to control the operation of the depth sensor 102 and interpret the results. For example, the sensor controller 104 or other processor may receive sensor data 106 and process, analyze, or otherwise perform calculations related to the sensor data 106. In some embodiments, the depth sensor 102 and / or sensor controller 104 may be similar to reference Figures 5A-5C The LiDAR sensor 564 described could also be another type of depth sensing sensor 102.

[0029] Sensor controller 104 can output sensor data 106 to computing system 108, for example in vehicle 500 and / or Figure 6 The example computing device 600 is a computing system that performs calculations. Sensor data 106 can be any of various forms, such as, but not limited to, (2D or 3D) LiDAR point clouds, projections or distance images, and / or other sensor data representations. Sensor data 106 can be analyzed to perform various functions associated with it and can be used with other sensor data 106 (e.g., Figures 5A-5C The various sensors shown and discussed herein are used in combination. In embodiments using LiDAR sensors, sensor data 106 may be referred to as LiDAR data; however, in other embodiments of this disclosure, sensor data 106 may be another type of depth data (e.g., from RADAR, ultrasound, etc.).

[0030] System 100 may include object detector 110. Object detector 110 may analyze the set of LiDAR data to identify one or more obstacle boundaries. Obstacle boundaries can be identified from a set of LiDAR data indicating physical obstacles near depth sensor 102. Obstacle boundaries may indicate a portion of a physical obstacle facing depth sensor 102. In an embodiment, obstacle boundaries and additional obstacle boundaries may be continuously occluded from depth sensor 102. Therefore, as machine 500 and / or obstacles move relative to each other, existing obstacle boundaries can be extended (e.g., because additional portions of the obstacle boundaries are now visible, such as...). Figure 4 (As shown and discussed herein) and new or updated obstacle boundaries can be created (e.g., because the corresponding physical obstacle is no longer occluded). Similarly, two or more obstacle boundaries may correspond to the same physical obstacle, for example, due to the irregular shape of the physical obstacle.

[0031] Examples of LiDAR data and identified obstacle boundaries in Figure 2A and Figure 2B As shown in the image. Figure 2A The first set of LiDAR data 200 shown in the figure Figure 2B The second set of LiDAR data 202 is shown. Obstacle boundary line 204 is shown in... Figure 2A-2B As shown, it corresponds to the detected obstacle boundary. Figure 2A-2B It also includes an unfilled central region 206 and corresponds to the location of the depth sensor 102 (although not in...). Figure 2A or Figure 2B (Shown in the diagram) and the radius extending from it. A machine representation 208 is set in a central region 206 to provide a reference for the approximate size and position of the machine (e.g., vehicle 500) relative to an obstacle. A series of concentric circles 210 are also shown, which can indicate areas where information is available but no obstacle is detected. In the case of an obstacle being detected, the obstacle may be present in the concentric circles 210 and an occluded area 212 (indicated by blank areas) may exist behind the obstacle boundary line 204 (from the perspective of the depth sensing sensor 102). As the obstacle and / or the depth sensing sensor 102 move relative to each other to remove the occlusion, the occluded object may become visible in additional iterations of the process. Figure 2B The obstacle boundary 204 in the middle has been relative to Figure 2A The obstacle boundary moves within the game.

[0032] System 100 may include a tracking state manager 112. The tracking state manager 112 can create and update various tracking states identified by the object detector 110. The tracking states can indicate obstacles tracked over time between subsequent iterations using the methods described herein. The tracking states can be used to determine the velocity and orientation of the obstacles (relative to the machine, relative to the ground, or relative to some other reference frame).

[0033] Because LiDAR data does not directly measure the velocity and orientation of obstacles, embodiments of this disclosure can use a Kalman filter to generate accurate estimates of the obstacle's velocity and orientation, and may include other variables associated with the obstacle. In some embodiments of this disclosure, a Kalman filter is used to generate and / or refine each of one or more tracked object states.

[0034] A Kalman filter, also known as a linear quadratic estimation (LQE), uses a set of inaccurate and / or indirect measurements over time to generate estimates of some unknown variables. In embodiments of this disclosure, the unknown variable being estimated may be the velocity of a detected obstacle, which may include orientation and / or velocity information. The Kalman filter may include a prediction operation that can generate estimates of the unknown variables based on inaccurate and / or indirect measurements. Furthermore, the Kalman filter estimates may be updated upon receiving a new set of inaccurate and / or indirect measurements. In embodiments, more weight may be given to data and / or estimates with higher determinism. The Kalman filter may include, or be any variant of, an extended Kalman filter or an unscented Kalman filter.

[0035] System 100 may include a velocity estimator 114. The velocity estimator can determine the estimated velocity of an object detected corresponding to one or more tracking states. Velocity estimator 114 can determine the estimated velocity of a newly detected obstacle boundary based on hypothetical correlations with one or more existing tracking states. The estimated velocity may include at least one of magnitude or orientation components. In some cases, determining which (if any) tracking state corresponds to the current obstacle boundary may be based on the magnitude component of the velocity, the orientation component of the velocity, or both.

[0036] Velocity estimator 114 can determine the estimated velocity by comparing the current detection of an object determined using one or more LiDAR sensors, the previous detection of an object or another object determined using one or more LiDAR sensors, and the elapsed time between the current detection and the previous detection. The physical distance between the two locations of each detection divided by the elapsed time can give the magnitude and / or direction of the estimated velocity. Velocity estimator 114 can pass the potential distance and / or estimated velocity through a gate function. If the distance, velocity, or other attribute is below a certain threshold, the gate function may only allow additional computation and / or consideration. The gate function can prevent unnecessary computation associated with potential matches that are unlikely to be actually relevant. In other embodiments, the gate function can be applied after the estimated velocity is determined and before the comparison is made to determine the correlation.

[0037] The velocity estimator 114 may include a coordinate transformer. The coordinate transformer can transform one or more tracked object states to a coordinate system corresponding to the currently detected object, where the comparison of the estimated velocity with one or more tracked object states is performed within the coordinate system. When comparing two (e.g., consecutive) frames (e.g., comparing two consecutive point clouds), one frame (e.g., the previous frame) can be transformed or corrected to the coordinate system of the subsequent or current frame (and vice versa) to compensate for ego-motion. By compensating for ego-motion, the analysis may only need to consider the object's motion relative to its own vehicle, since the ego-vehicle's motion has been decomposed—thus allowing the alignment of two frames (e.g., two LiDAR point clouds) in the same coordinate system.

[0038] Once the estimated velocity is determined, the tracking state manager 112 (or other component) can determine which tracking states correspond to the current object boundary and thus update these tracking states with the new information using a Kalman filter. The tracking state manager 112 can compare the estimated velocity with one or more tracking object states corresponding to one or more previously detected objects.

[0039] Looking at it now Figure 3 The tracking state manager 112 can create and / or analyze probability distributions corresponding to the velocities of each detected object among the previously detected objects. Figure 3 An example probability distribution plot 300 is shown. Typically, the probability distribution plot 300 has a probability axis 302 (e.g., the weights of the distribution) and a value axis 304. It should also be understood that some embodiments of this disclosure may be considered using higher dimensions (e.g., 3D) and... Figure 3 The simplified probability distribution diagram 300 is presented to facilitate reader understanding.

[0040] A first probability distribution 306 and a second probability distribution 308 are displayed on a probability distribution graph 300. A measurement 310 is displayed at a specific location relative to the first probability distribution 306 and the second probability distribution 308. Embodiments of this disclosure disclose determining which of the probability distributions 306 and 308 (if any) is most likely to correspond to the measurement 310.

[0041] The first probability distribution 306 may have an associated first gate function 312. A relatively wide first probability distribution 306 may have a corresponding wide gate function, indicating that it can correspond to multiple measurements of the first probability distribution. The second probability distribution 308 may have an associated second gate function 314. A relatively sharp second probability distribution 308 may have a corresponding narrow gate function, indicating that a narrow range of measurements can correspond to the second probability distribution. Figure 3As can be seen in the example, measurement 310 is within both the first gate function 312 and the second gate function 314. If measurement 310 is outside one or more gate functions, the corresponding probability distribution can be disregarded.

[0042] Because, in Figure 3 In the example, measurement 310 is passed through (e.g., within) two gate functions 312, 314, and the tracking state analyzer 112 (or other component) can determine a probability distribution (indicating a previously detected object among a set of previously detected objects) with the highest probability corresponding to measurement 310 (indicating the current obstacle boundary) (or one of them). The tracking state analyzer 112 can analyze a certain distance between measurement 310 and respective probability distributions 306, 308. In some embodiments of this disclosure, this certain distance can be based on a highest likelihood, which is represented by the height of the distance relative to the probability distribution of the measurement. Therefore, since measurement 310 is most likely to be the second probability distribution 308 shown, the tracking state analyzer 112 can select the second probability distribution as the object most likely to correspond to measurement 310.

[0043] In other embodiments of this disclosure, the specific distance may be the Mahalanobis distance as a cost function. However, the Mahalanobis distance may be more favorable for a wider range of probability distributions. This is because the Mahalanobis distance for the first probability distribution 306 may be smaller. Therefore, in embodiments of this disclosure, the Mahalanobis distance may be used as a gate function, and the choice is based on likelihood (e.g., the vertical height of the respective probability distribution).

[0044] Once a corresponding tracking state is identified, the tracking state manager 112 (or other component) can use the estimated velocity to update the tracking object state corresponding to the previously detected object, which has the highest probability of corresponding to the object, to generate an updated tracking object state. The updated tracking state can be referred to as the improved tracking state. The tracking state manager 112 can use a Kalman filter to create the improved tracking state.

[0045] System 100 may include vehicle controller 116. For example... Figure 1A As shown, system 100 may include vehicle controller 116, which can analyze information in the tracking state to determine obstacles in the physical environment surrounding depth perception sensor 102. Vehicle controller 116 can then instruct or send information related to one or more vehicle actions based on the identified obstacles, such as activating the brakes or turning the vehicle. For example, vehicle controller 116 may execute (or otherwise associate with) an autonomous driving software stack that may include a perception layer, a world model management layer, a planning layer, a control layer, an actuation layer, an obstacle avoidance layer, and / or one or more other layers.

[0046] System 100 can be associated with computing system 600 in autonomous vehicle 500, for example, regarding Figures 5A-5D and Figure 6 As shown and described. For example, an autonomous vehicle 500 that operates autonomously may include a computer system 600 that makes extensive observations of surrounding obstacles and determines a number of actions that the vehicle 500 must perform to avoid those obstacles.

[0047] Now for reference Figure 1B Each block of method 150 described herein includes a computational process that can be performed using any combination of hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in memory. Method 150 can also be embodied as computer-usable instructions stored on a non-transient computer storage medium. Method 150 can be provided by a standalone application, service, or managed service (standalone or in combination with another managed service) or plug-in to another product, to name a few. Furthermore, by way of example, regarding Figure 1A System 100 and / or Figure 6 The computer system 600 describes method 150. However, the method may be performed additionally or alternatively by any system, or any combination of systems, including but not limited to those described herein.

[0048] Figure 1B This is a flowchart illustrating a method 150 for detecting and tracking objects between subsequent sets of LiDAR data (or other depth-sensing data) and improving those detections over time, according to some embodiments of this disclosure. At block B102, method 150 includes, for example, acquiring current detections from depth-sensing sensor 102 and / or sensor controller 104. At block B102, method 150 may include determining one or more object boundaries from the LiDAR data, wherein the object boundaries may define one or more objects detected in the LiDAR data, such as… Figure 2A As shown in and discussed herein, current detection can be received from the depth sensing sensor 102 (e.g., a LiDAR sensor). Figure 1B The current detection (A) in the depth sensing sensor 102 indicates a scan or other capture at a certain moment. One or more object boundaries in the current detection can be associated with a specific time for determining the estimated velocity.

[0049] At box B104, method 150 includes comparing one or more currently detected object boundaries with one or more previously detected and / or tracking states. First tracking state ( Figure 1B Tracking state 1) and second tracking state ( Figure 1BThe tracking states 2) can each be considered relative to the boundary of the currently detected object. The first tracking state and the second tracking state can be potential candidates related to the boundary of the currently detected object.

[0050] In embodiments of this disclosure, at block B104, method 150 includes using an Iterative Closest Point (ICP) algorithm to compare the current obstacle boundary with a first obstacle boundary corresponding to a first tracking state to generate a first velocity estimate and with a second obstacle boundary corresponding to a second tracking state to generate a second velocity estimate. The Iterative Closest Point algorithm can identify an offset vector such that the offset vector combined with the first tracking state will align with the second tracking state. In embodiments, the ICP algorithm can calculate an offset vector indicating the magnitude and direction of the obstacle's movement (if any). Various distance and velocity measurements can be calculated based on the offset vector.

[0051] Examples of the ICP algorithm are in Figure 4The diagram is illustrated graphically, showing the self-vehicle 400 moving relative to the parked vehicle 402. As the self-vehicle 400 moves past the parked vehicle 402, a series of four detections 404 (displayed sequentially) are captured. A corresponding set of four obstacle boundaries 406A-D are detected. Detection 1 (performed as the self-vehicle 400 moves away from and approaches the parked vehicle 402) includes a first obstacle boundary 406A, which is a straight line indicating the first end of the parked vehicle 402. The centroid 408A of Detection 1 is located at the center of this straight line. Detection 2 (performed as the self-vehicle 400 approaches and nears the parked vehicle 402) includes a second obstacle boundary 406B, which is "L"-shaped and indicates the portion of the parked vehicle 402 visible to the depth sensing sensor 102 on the self-vehicle 400. The centroid 408B of Detection 2 is located at the average position of the "L" shape. The ICP algorithm can correlate the first obstacle boundary 406A with the second obstacle boundary 406B to provide a more complete obstacle boundary. Relative to detection 2, the first obstacle boundary 406A is shown as a dashed line relative to the second obstacle boundary 406B. Detection 3 (performed when the ego vehicle 400 approaches and passes the parked vehicle 402) includes a third obstacle boundary 406C, which is "L"-shaped and indicates the portion of the parked vehicle 402 visible to the ego vehicle 400's depth sensing sensor 102. The centroid 408C of detection 3 is located at the average position of the "L" shape. The ICP algorithm can associate the first obstacle boundary 406A and / or the second obstacle boundary 406B (or at least partially indicate a combined obstacle boundary of both the first obstacle boundary 406A and the second obstacle boundary 406B) with the third obstacle boundary 406C to provide a more complete obstacle boundary. Detection 4 (performed when the ego vehicle 400 passes the parked vehicle 402) includes a fourth obstacle boundary 406D, which is a straight line indicating the second end of the parked vehicle 402. The centroid 408D of detection 4 is located at the center of this straight line. The ICP algorithm can associate a first obstacle boundary 406A, a second obstacle boundary 406B, and / or a third obstacle boundary 406D (or a combination of obstacle boundaries that at least partially indicate some combination thereof) with a fourth obstacle boundary 406D to provide a more complete set of obstacle boundaries. It should be understood that detections 3 and 4 are considered as self-vehicle 400 passing over parked vehicle 402 and respectively include substantially mirror images of the obstacle boundaries of detections 2 and 1 (although this is for illustrative purposes only). Typically, the individual detections will be performed over a time interval, rather than at a relative position to any particular obstacle. Figure 4 The corresponding tests are shown for illustrative purposes only.

[0052] It should be understood that the previous system only used the centroid 408A-D to determine the location of the obstacle in each iteration (e.g., detections 1-4). Figure 4As shown, when the ego vehicle 400 passes the parked vehicle 402, the centroid 408A-D moves. This would result in detected motion in prior art systems. The ICP algorithm reduces or eliminates perceived motion of the parked vehicle 402 by comparing the respective obstacle boundaries 406A-D rather than relying on the centroid 408A-D. The ICP algorithm can be used to calculate offset vectors to align two or more obstacle boundaries 406A-D. For Figure 4 For parked vehicle 402, the offset vector will be zero or close to zero, resulting in almost no detected movement of the parked vehicle. For moving vehicles ( Figure 4 (Not shown in the image), the translation vector will indicate the direction and magnitude of the vehicle's movement. The magnitude of the movement can be determined at least in part based on the translation removal of obstacles between two corresponding detections to the time interval between detections. The direction of the offset vector can be determined at least in part based on the direction of the translation.

[0053] In embodiments of this disclosure, the comparison between the current obstacle detection and one or more previous detections is based at least in part on the similarity of the shapes of the corresponding obstacles. This is likely because shapes are unlikely to change drastically between single frames of LiDAR data. The first obstacle boundary may include a first shape metric indicating a first geometry, and the current obstacle boundary may include a second shape metric indicating a second geometry. Thus, the determination that the current obstacle boundary most closely corresponds to the first obstacle boundary can be further based at least in part on the first and second geometries.

[0054] The ICP algorithm can be used, at least in part, to perform the comparison discussed above between the currently detected first object boundary and the second object boundary in the tracking state. Although the ICP position changes in... Figure 4 In the example, the motion is dramatic; in embodiments of this disclosure, consecutive LiDAR datasets can be placed closer together. Therefore, the shape difference of the obstacle boundaries 406A-D between two consecutive scans of the LiDAR data can be minimized, and an accurate measurement of the motion of the physical object between consecutive scans can be provided. It should be understood that in embodiments of this disclosure, the ICP algorithm may consider or otherwise transform one or more tracked object states into a coordinate system corresponding to the current detection of the object, and compare the estimated velocity with one or more tracked object states that may be in the updated coordinate system. This transformation to the updated coordinate system may be based on the measured motion of the ego machine 500 (e.g., ego motion).

[0055] At box B106, method 150 includes determining whether a correlation exists between the current detection and any previous detection and / or tracking state. This operation can be performed by determining a cost function (or other probability distribution) associated with the estimated velocity of each potential correlation. The estimated velocity of a potential match can be determined, and then the cost function indicating the likelihood of correlation can be calculated based at least in part on the estimated velocity. The estimated velocity of the object can be determined at least in part based on the current detection of an object determined using one or more LiDAR sensors, the previous detection of an object determined using one or more LiDAR sensors or another object, and a comparison between the time elapsed between the current detection and the previous detection. Both the current detection of the object and the previous detection of the object can use the ICP as a reference location, and the previous detection of the object can be adjusted by coordinate transformation.

[0056] At box B106, method 150 may further include comparing the estimated velocity with one or more tracked object states corresponding to one or more previously detected objects, each tracked object state representing a probability distribution of the velocity corresponding to a corresponding detected object among the previously detected objects. The probability distribution may indicate the likelihood of the estimated velocity corresponding to the tracked object state. Based on the analysis of the probability distribution, one of the previously detected objects may be identified as having the highest probability corresponding to that object or being among those with the highest probability corresponding to that object.

[0057] At box B106, method 150 may further include calculating a cost function corresponding to the probability distribution. Figure 1B In the example, the current detection A is compared with tracking state 1 to create a first probability distribution, and the current detection A is compared with tracking state 2 to create a second probability distribution. The cost function can be at least partially based on, as... Figure 3 The corresponding probability distributions are shown. The first cost function can be calculated using the first probability distribution and the first velocity estimate. The second cost function can be calculated using the second probability distribution and the second velocity estimate.

[0058] A cost function can be evaluated to determine which (if any) most likely corresponds to the current obstacle boundary. Typically, a lower cost function indicates a higher relevance likelihood. At box B106, method 150 may further include determining that the current obstacle boundary most closely corresponds to the first obstacle boundary based on the first cost function being less than the second cost function.

[0059] Upon discovering a correlation with a first tracking state, method 150 includes updating the first tracking state at box B108. Alternatively, upon discovering a correlation with a second tracking state, method 150 includes updating the second tracking state at box B110. The corresponding updated tracking state may include information regarding previous obstacle boundaries and current obstacle boundaries, which are found to correspond to the respective tracking states.

[0060] If no correlation is found, method 150 includes creating a new tracking state for the object at box B112. If no correlation is found with the previous tracking state, the newly discovered object may have entered the field of view from a distance or from behind an occlusion. The new tracking state can then be used for future comparisons to determine whether a subsequent current obstacle boundary corresponds to the new tracking state. Thus, the new tracking state can be improved in subsequent scans by the depth sensing sensor 102.

[0061] Method 150, following box B108 in box B114, includes applying a Kalman filter to a first tracking state. At box B116, method 150 includes creating an improved first tracking state. An improved tracking state can be created by applying a Kalman filter to add additional information to the tracking state. In addition to one or more previous obstacle detections for the same physical obstacle, the improved tracking state can (directly or indirectly) indicate the current obstacle detection. Similarly, method 150, following box B110 in box B118, includes applying a Kalman filter to a second tracking state. At box B120, method 150 includes creating an improved second tracking state.

[0062] At block B120, method 150 includes controlling the machine based at least in part on an improved and / or new tracking state. For example, the tracking state may indicate a physical object that could directly or indirectly affect the autonomous vehicle 500 or other machines. Thus, obstacles may be moving, such as other vehicles or pedestrians, or stationary, such as buildings and trees. Obstacles may have a position relative to the vehicle, a 3D motion vector (which may include acceleration, rotation, or other motion indicators), and a relative size (based on the number of pixels corresponding to the obstacle). In some embodiments, based at least in part on information about the obstacle, system 100 may determine the likelihood that the obstacle may affect the vehicle, and the system may determine one or more remedial actions to be performed to avoid the obstacle. For example, vehicle controller 118 may determine that the vehicle should brake, turn, or accelerate to avoid the obstacle. If the obstacle cannot be avoided, system 100 may determine that one or more remedial actions should be taken, such as minimizing damage or activating other safety features. The control at block B120 may be performed to avoid the identified obstacle by taking one or more remedial actions. Controlling the vehicle may include sending commands to any of a plurality of vehicle systems, such as regarding Figures 5A-5D Those described.

[0063] Example autonomous vehicles

[0064] Figure 5A This is an illustration of an example autonomous vehicle 500 according to some embodiments of the present disclosure. The autonomous vehicle 500 (or, alternatively, referred to herein as “vehicle 500”) may include, but is not limited to, passenger vehicles such as automobiles, trucks, buses, ambulances, shuttles, electric or motorized bicycles, motorcycles, fire trucks, police cars, ambulances, boats, engineering vehicles, underwater vessels, drones, trailer-mounted vehicles, and / or other types of vehicles (e.g., driverless and / or capable of accommodating one or more passengers). Autonomous vehicles are generally described according to the level 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 previous and future versions of this standard). Vehicle 500 may be able to perform one or more functions that meet the requirements of Level 3 to Level 5 autonomous driving. For example, depending on the embodiment, vehicle 500 may be able to perform conditional automation (Level 3), high automation (Level 4), and / or full automation (Level 5).

[0065] Vehicle 500 may include components such as chassis, body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other vehicle components. Vehicle 500 may include a propulsion system 550, such as an internal combustion engine, a hybrid power plant, an all-electric motor, and / or another type of propulsion system. Propulsion system 550 may be connected to the drivetrain of vehicle 500, which may include a transmission, to enable propulsion of vehicle 500. Propulsion system 550 may be controlled in response to receiving a signal from throttle / accelerator 552.

[0066] A steering system 554, which may include a steering wheel, can be used to steer the vehicle 500 (e.g., along a desired path or route) when the propulsion system 550 is operating (e.g., when the vehicle is in motion). The steering system 554 may receive signals from the steering actuator 556. For fully automatic (level 5) functionality, the steering wheel may be optional.

[0067] The brake sensor system 546 can be used to operate the vehicle brakes in response to receiving signals from the brake actuator 548 and / or the brake sensor.

[0068] It may include one or more System-on-a-Chip (SoC) 504 ( Figure 5C One or more controllers 536, including and / or one or more GPUs, may provide signals (e.g., signals representing commands) to one or more components and / or systems of vehicle 500. For example, one or more controllers may send signals to operate vehicle brakes via one or more brake actuators 548, to operate steering system 554 via one or more steering actuators 556, and to operate propulsion system 550 via one or more throttles / accelerators 552. One or more controllers 536 may include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operating commands (e.g., signals representing commands) to enable autonomous driving and / or assist a human driver in driving vehicle 500. One or more controllers 536 may include a first controller 536 for autonomous driving functions, a second controller 536 for functional safety functions, a third controller 536 for artificial intelligence functions (e.g., computer vision), a fourth controller 536 for infotainment functions, a fifth controller 536 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 536 can handle two or more of the functions described above, and two or more controllers 536 can handle a single function, and / or any combination thereof.

[0069] One or more controllers 536 may provide signals for controlling one or more components and / or systems of vehicle 500 in response to sensor data (e.g., sensor inputs) received from one or more sensors. Sensor data may be received from, for example, but not limited to, a global navigation satellite system sensor 558 (e.g., a Global Positioning System sensor), a RADAR sensor 560, an ultrasonic sensor 562, a LIDAR sensor 564, an inertial measurement unit (IMU) sensor 566 (e.g., an accelerometer, gyroscope, magnetic compass, magnetometer, etc.), a microphone 596, a stereo camera 568, a wide-angle camera 570 (e.g., a fisheye camera), an infrared camera 572, a surround camera 574 (e.g., a 360-degree camera), a long-range and / or medium-range camera 598, a speed sensor 544 (e.g., for measuring the rate of vehicle 500), a vibration sensor 542, a steering sensor 540, a braking sensor (e.g., as part of a braking sensor system 546), and / or other sensor types.

[0070] One or more of the controllers 536 may receive inputs (e.g., represented by input data) from the instrument cluster 532 of the vehicle 500 and provide outputs (e.g., represented by output data, display data, etc.) via a human-machine interface (HMI) display 534, an auditory signaling device, a speaker, and / or via other components of the vehicle 500. These outputs may include information such as vehicle speed, rate, time, map data (e.g., [missing information]). Figure 5C Information such as the HD map 522, location data (e.g., the location of vehicle 500 on the map), direction, and the location of other vehicles (e.g., occupying a grid), as well as information about objects and their states perceived by controller 536, etc. For example, HMI display 534 may display information about the existence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information about driving maneuvers that the vehicle has made, is making, or can make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0071] Vehicle 500 further includes a network interface 524, which can communicate via one or more networks using one or more wireless antennas 526 and / or a modem. For example, network interface 524 may be able to communicate via LTE, WCDMA, UMTS, GSM, CDMA2000, etc. One or more wireless antennas 526 may also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using one or more local area networks such as Bluetooth, Bluetooth LE, Z-Wave, ZigBee, etc. and / or one or more low-power wide area networks (LPWAN) such as LoRaWAN, SigFox, etc.

[0072] Figure 5BFigure 5 shows an example of the camera position and field of view of an example autonomous vehicle 500. (Refer to...) Figure 5A According to some embodiments of this disclosure, the camera and corresponding field of view are an exemplary 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 different positions on the vehicle 500.

[0073] The camera type used for the camera may include, but is not limited to, a digital camera suitable for use with components and / or systems of vehicle 500. The camera may operate at Automotive Safety Integrity Level (ASIL) B and / or another ASIL. The camera type may have any image capture rate, such as 60 frames per second (fps), 120 fps, 240 fps, etc., depending on the embodiment. The camera may be able to use a rolling shutter, a global shutter, another type of shutter, or a combination thereof. In some examples, the color filter array may include a red-white-white-white (RCCC) color filter array, a red-white-white-blue (RCCB) color filter array, a red-blue-green-white (RBGC) color filter array, a Foveon X3 color filter array, a Bayer sensor (RGGB) color filter array, a monochrome sensor color filter array, and / or another type of color filter array. In some embodiments, a sharp-pixel camera, such as a camera with RCCC, RCCB, and / or RBGC color filter arrays, may be used in efforts to improve light sensitivity.

[0074] In some examples, one or more of the cameras can be used to perform advanced driver assistance system (ADAS) functions (e.g., as part of a redundant or fail-safe design). For example, a multi-function monocular camera can be installed to provide functions including lane departure warning, traffic sign assistance, and intelligent headlight control. One or more of the cameras (e.g., all cameras) can simultaneously record and provide image data (e.g., video).

[0075] One or more of the cameras can be mounted in mounting components such as custom-designed (3-D printed) parts to cut off stray light and reflections from inside the vehicle (e.g., reflections from the dashboard in the windshield mirror) that may interfere with the camera's image data capture capabilities. Regarding the wing mirror mounting components, the wing mirror components can be custom-3-D printed so that the camera mounting plate matches the shape of the wing mirror. In some examples, one or more cameras can be integrated into the wing mirror. For side-view cameras, one or more cameras can also be integrated into the four pillars at each corner of the cab.

[0076] A camera with a field of view that includes the environment in front of the vehicle 500 (e.g., a front-facing camera) can be used for surround view to help identify forward paths and obstacles, and, with the assistance of one or more controllers 536 and / or control SoCs, to provide information crucial for generating an occupancy grid and / or determining a preferred vehicle path. The front-facing camera can be used to perform many of the same ADAS functions as LiDAR, including emergency braking, pedestrian detection, and collision avoidance. The front-facing camera can also be used in ADAS functions and systems, including Lane Departure Warning (LDW), Autonomous Cruise Control (ACC), and / or other functions such as traffic sign recognition.

[0077] A variety of cameras can be used in front-facing configurations, including monocular camera platforms such as CMOS (Complementary Metal-Oxide-Semiconductor) color imagers. Another example could be a wide-angle camera 570, which can be used to perceive objects entering the field of view from the periphery (such as pedestrians, traffic at intersections, or bicycles). Although Figure 5B The middle image shows only one wide-angle camera, but any number of wide-angle cameras 570 can be present on the vehicle 500. Furthermore, remote cameras 598 (e.g., a pair of long-view stereo cameras) can be used for depth-based object detection, especially for objects for which neural networks have not yet been trained. Remote cameras 598 can also be used for object detection and classification, as well as basic object tracking.

[0078] One or more stereo cameras 568 may also be included in a front-mounted configuration. The stereo camera 568 may include an integrated control unit comprising a scalable processing unit that can provide a multi-core microprocessor and programmable logic (FPGA) with an integrated CAN or Ethernet interface on a single chip. Such a unit can be used to generate a 3D map of the vehicle environment, including distance estimates for all points in the image. Alternative stereo cameras 568 may include a compact stereo vision sensor that may include two camera lenses (one on each side) and an image processing chip capable of measuring the distance from the vehicle to a target object and using the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning functions. Other types of stereo cameras 568 may be used in addition to those described herein, or alternatively.

[0079] Cameras with a field of view including the side portion of the environment of vehicle 500 (e.g., side-view cameras) can be used for surround view, providing information for creating and updating occupancy grids and generating side-impact collision warnings. For example, surround camera 574 (e.g., ... Figure 5BThe four surround cameras 574 shown can be mounted on the vehicle 500. The surround cameras 574 can include a wide-angle camera 570, a fisheye camera, a 360-degree camera, and / or the like. Four examples are provided; the four fisheye cameras can be positioned at the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle can use three surround cameras 574 (e.g., left, right, and rear) and can utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround-view camera.

[0080] Cameras with a field of view that includes the environment behind vehicle 500 (e.g., rear-view cameras) can be used for parking assistance, surround view, rear collision warning, and creating and updating occupancy grids. A wide variety of cameras can be used, including but not limited to those also suitable as front-facing cameras as described herein (e.g., long-range and / or mid-range cameras 598, stereo cameras 568, infrared cameras 572, etc.).

[0081] Figure 5C For use in accordance with some embodiments of this disclosure Figure 5A The example autonomous vehicle 500 is illustrated in the block diagram of an example system architecture. It should be understood that this arrangement, and other arrangements described herein, are merely illustrative. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities, which may be implemented as discrete or distributed components or in combination with other components, and in any suitable combination and location. The various functions described herein as being performed by these entities can be implemented via hardware, firmware, and / or software. For example, the various functions can be implemented by a processor executing instructions stored in memory.

[0082] Figure 5C Each component, feature, and system in vehicle 500 is illustrated as being connected via bus 502. Bus 502 may include a Controller Area Network (CAN) data interface (or, alternatively, referred to herein as the "CAN bus"). CAN may be a network within vehicle 500 used to assist in the control of various features and functions of vehicle 500, such as the actuation of brakes, acceleration, braking, steering, windshield wipers, etc. The CAN bus may be configured to have dozens or even hundreds of nodes, each with its own unique identifier (e.g., CAN ID). The CAN bus can be read to find steering wheel angle, ground speed, engine speed per minute (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0083] Although bus 502 is described herein as a CAN bus, this is not intended to be limiting. For example, FlexRay and / or Ethernet may be used in addition to or alternatively to a CAN bus. Furthermore, although bus 502 is represented by a single line, this is not intended to be limiting. For example, any number of buses 502 may exist, which may include one or more CAN buses, one or more FlexRay buses, one or more Ethernet buses, and / or one or more other types of buses using different protocols. In some examples, two or more buses 502 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 502 may be used for a collision avoidance function, and a second bus 502 may be used for drive control. In any example, each bus 502 may communicate with any component of vehicle 500, and two or more buses 502 may communicate with the same component. In some examples, each SoC 504, each controller 536, and / or each computer within the vehicle may have access to the same input data (e.g., input from sensors of vehicle 500) and may be connected to a common bus such as a CAN bus.

[0084] Vehicle 500 may include one or more controllers 536, such as those described herein. Figure 5A The controllers described herein. Controller 536 can be used for a wide variety of functions. Controller 536 can be coupled to any other different components and systems of vehicle 500 and can be used for the control of vehicle 500, artificial intelligence of vehicle 500, infotainment and / or the like for vehicle 500.

[0085] Vehicle 500 may include one or more System-on-Chip (SoC) 504s. SoC 504 may include a CPU 506, GPU 508, processor 510, cache 512, accelerator 514, data storage 516, and / or other components and features not shown. SoC 504 can be used to control vehicle 500 across a wide variety of platforms and systems. For example, one or more SoCs 504s may be combined with an HD map 522 in a system (e.g., the system of vehicle 500), the HD map being transmitted via a network interface 524 from one or more servers (e.g., [server name missing]). Figure 5D One or more servers (578) receive map refresh and / or updates.

[0086] CPU 506 may include CPU clusters or CPU complexes (alternatively referred to herein as "CCPLEX"). CPU 506 may include multiple cores and / or L2 cache. For example, in some embodiments, CPU 506 may include eight cores in a coherent multiprocessor configuration. In some embodiments, CPU 506 may include four dual-core clusters, each with a dedicated L2 cache (e.g., 2MB L2 cache). CPU 506 (e.g., CCPLEX) may be configured to support simultaneous cluster operation, such that any combination of clusters of CPU 506 can be active at any given time.

[0087] CPU 506 can implement power management capabilities including one or more of the following features: automatic clock gating of hardware blocks when idle to conserve dynamic power; clock gating of each core when the core is not actively executing instructions due to the execution of WFI / WFE instructions; independent power gating of each core; independent clock gating of each core cluster when all cores are clock-gated or power-gated; and / or independent power gating of each core cluster when all cores are power-gated. CPU 506 can further implement enhanced algorithms for managing power states, wherein allowed power states and desired wake-up times are specified, and the hardware / microcode determines the optimal power state to enter for the core, cluster, and CCPLEX. The processing core can support simplified power state entry sequences in software, with this work offloaded to the microcode.

[0088] GPU 508 may include an integrated GPU (or, alternatively, referred to herein as an "iGPU"). GPU 508 may be programmable and efficient for parallel workloads. In some examples, GPU 508 may use an enhanced tensor instruction set. GPU 508 may include one or more streaming microprocessors, wherein each streaming microprocessor may include an L1 cache (e.g., an L1 cache with at least 96KB of storage capacity), and two or more of these streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB of storage capacity). In some embodiments, GPU 508 may include at least eight streaming microprocessors. GPU 508 may use a computation application programming interface (API). Furthermore, GPU 508 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0089] In automotive and embedded applications, the GPU 508 can be power-optimized for optimal performance. For example, the GPU 508 can be fabricated on FinFETs. However, this is not intended to be limiting, and the GPU 508 can be fabricated using other semiconductor manufacturing processes. Each streaming microprocessor can combine several mixed-precision processing cores divided into multiple blocks. For example, and without limitation, 64 PF32 cores and 32 PF64 cores can be divided into four processing blocks. In such an example, each processing block can be allocated 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two mixed-precision NVIDIA Tensor cores for deep learning matrix arithmetic, an L0 instruction cache, a warp scheduler, dispatch units, and / or a 64KB register file. Furthermore, the streaming microprocessor can include independent parallel integer and floating-point data paths to leverage the mixture of computation and addressing computations for efficient execution of workloads. The streaming microprocessor can include independent thread scheduling capabilities to allow for finer-grained synchronization and cooperation between parallel threads. Streaming microprocessors can include a combination of L1 data cache and shared memory units to improve performance while simplifying programming.

[0090] The GPU 508 may include, in some examples, a High Bandwidth Memory (HBM) and / or a 16GB HBM2 memory subsystem providing a peak memory bandwidth of approximately 900GB / s. In some examples, in addition to HBM memory or alternatively, Synchronous Graphics Random Access Memory (SGRAM), such as Generation 5 Graphics Double Data Rate Synchronous Random Access Memory (GDDR5), may be used.

[0091] The GPU 508 may include unified memory technology, which includes access counters to allow memory pages to be migrated more precisely to the processors that access them most frequently, thereby improving the efficiency of shared memory ranges between processors. In some examples, Address Translation Service (ATS) support can be used to allow the GPU 508 to directly access the CPU 506 page tables. In such examples, when the GPU 508 Memory Management Unit (MMU) experiences a miss, an address translation request can be transferred to the CPU 506. In response, the CPU 506 can look up the virtual-physical mapping for the address in its page tables and transfer the translation back to the GPU 508. Thus, unified memory technology can allow a single unified virtual address space for the memory of both the CPU 506 and the GPU 508, simplifying GPU 508 programming and porting applications to the GPU 508.

[0092] In addition, the GPU 508 may include access counters that track how frequently the GPU 508 accesses the memory of other processors. These access counters help ensure that memory pages are moved to the physical memory of the processor that accesses those pages most frequently.

[0093] SoC 504 may include any number of caches 512, including those described herein. For example, cache 512 may include an L3 cache available to both CPU 506 and GPU 508 (e.g., it is connected to both CPU 506 and GPU 508). Cache 512 may include a write-back cache, which can track the state of rows, for example, using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). Depending on the embodiment, the L3 cache may include 4MB or more, but a smaller cache size may also be used.

[0094] SoC 504 may include one or more arithmetic logic units (ALUs) that can be used to perform processing of any of a variety of tasks or operations related to vehicle 500, such as processing a DNN. Additionally, SoC 504 may include a floating-point unit (FPU) or other mathematical coprocessor or digital coprocessor type for performing mathematical operations within the system. For example, SoC 104 may include one or more FPUs integrated as execution units within CPU 506 and / or GPU 508.

[0095] SoC 504 may include one or more accelerators 514 (e.g., hardware accelerators, software accelerators, or combinations thereof). For example, SoC 504 may include a hardware acceleration cluster, which may include optimized hardware accelerators and / or large on-chip memory. This large on-chip memory (e.g., 4MB SRAM) can enable the hardware acceleration cluster to accelerate neural networks and other computations. The hardware acceleration cluster can be used to supplement GPU 508 and offload some tasks from GPU 508 (e.g., freeing up more cycles of GPU 508 to perform other tasks). As an example, accelerator 514 can be used for targeted workloads (e.g., perceptrons, convolutional neural networks (CNNs), etc.) that are stable enough to be easily controlled for acceleration. When used herein, the term "CNN" can include all types of CNNs, including region-based or region convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0096] Accelerator 514 (e.g., a hardware acceleration cluster) may include a Deep Learning Accelerator (DLA). The DLA may include one or more Tensor Processing Units (TPUs) that can be configured to provide an additional 10 trillion operations per second for deep learning applications and inference. The TPU may be an accelerator configured to perform image processing functions (e.g., for CNNs, RCNNs, etc.) and optimized for performing image processing functions. The DLA may be further optimized for a specific set of neural network types and floating-point operations and inference. The DLA is designed to provide higher performance per millimeter than a general-purpose GPU and significantly outperform CPUs. The TPU can perform several functions, including single-instance convolution functions, support for INT8, INT16, and FP16 data types for both features and weights, and post-processor functions.

[0097] DLA can execute neural networks, especially CNNs, quickly and efficiently on processed or unprocessed data for any function across a wide variety of applications, such as, but not limited to: CNNs for object recognition and detection using data from camera sensors; CNNs for distance estimation using data from camera sensors; CNNs for emergency vehicle detection and recognition using data from microphones; CNNs for face recognition and vehicle owner recognition using data from camera sensors; and / or CNNs for safety and / or safety-related events.

[0098] The DLA can perform any function of the GPU 508, and by using inference accelerators, for example, a designer can make either the DLA or the GPU 508 target any function. For example, a designer can focus the CNN processing and floating-point operations on the DLA and leave other functions to the GPU 508 and / or other accelerators 514.

[0099] Accelerator 514 (e.g., a hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may alternatively be referred to herein as a computer vision accelerator. The PVA may be designed and configured to accelerate computer vision algorithms for advanced driver assistance systems (ADAS), autonomous driving, and / or augmented reality (AR) and / or virtual reality (VR) applications. The PVA can provide a balance between performance and flexibility. For example, each PVA may include, for example, but not limited to, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0100] RISC cores can interact with image sensors (such as the image sensor of any camera described herein), image signal processors, and / or the like. Each of these RISC cores may include any amount of memory. Depending on the embodiment, the RISC core may use any of several protocols. In some examples, the RISC core may execute a real-time operating system (RTOS). RISC cores may be implemented using one or more integrated circuit devices, application-specific integrated circuits (ASICs), and / or memory devices. For example, a RISC core may include an instruction cache and / or tightly coupled RAM.

[0101] DMA enables PVA components to access system memory independently of the CPU 506. DMA can support any number of features to provide optimizations to the PVA, including but not limited to support for multidimensional addressing and / or circular addressing. In some examples, DMA can support addressing in up to six or more dimensions, which can include block width, block height, block depth, horizontal block step, vertical block step, and / or depth step.

[0102] A vector processor can be a programmable processor designed to efficiently and flexibly execute programming for computer vision algorithms and provide signal processing capabilities. In some examples, a PVA may include a PVA core and two vector processing subsystem partitions. The PVA core may include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem may operate as the main processing engine of the PVA and may include a vector processing unit (VPU), an instruction cache, and / or vector memory (e.g., VMEM). The VPU core may include a digital signal processor, such as, for example, a Single Instruction Multiple Data (SIMD) or Very Long Instruction Word (VLIW) digital signal processor. The combination of SIMD and VLIW can enhance throughput and speed.

[0103] Each of the vector processors may include an instruction cache and may be coupled to dedicated memory. Consequently, in some examples, each of the vector processors may be configured to execute independently of other vector processors. In other examples, the vector processors included in a particular PVA may be configured to employ data parallelization. For example, in some embodiments, multiple vector processors included in a single PVA may execute the same computer vision algorithm, but on different regions of an image. In other examples, vector processors included in a particular PVA may execute different computer vision algorithms simultaneously on the same image, or even different algorithms on a sequence of images or portions of an image. Among other things, any number of PVAs may be included in a hardware acceleration cluster, and any number of vector processors may be included in each of these PVAs. Furthermore, the PVA may include additional error correction code (ECC) memory to enhance overall system security.

[0104] Accelerator 514 (e.g., a hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high-bandwidth, low-latency SRAM for accelerator 514. In some examples, on-chip memory may include at least 4MB of SRAM consisting of, for example, but not limited to, eight field-configurable memory blocks, accessible by both the PVA and DLA. Each pair of memory blocks may include an Advanced Peripheral Bus (APB) interface, configuration circuitry, a controller, and a multiplexer. Any type of memory may be used. The PVA and DLA may access memory via a backbone that provides high-speed memory access to the PVA and DLA. The backbone may include (e.g., using an APB) an on-chip computer vision network that interconnects the PVA and DLA to memory.

[0105] On-chip computer vision networks can include interfaces that ensure both the PVA and DLA provide ready and valid signals before transmitting any control signals / addresses / data. Such interfaces can provide separate phases and channels for transmitting control signals / addresses / data, as well as burst communication for continuous data transmission. This type of interface can conform to ISO 26262 or IEC 61508 standards, but other standards and protocols can also be used.

[0106] In some examples, SoC 504 may include, for example, a real-time ray tracing hardware accelerator as described in U.S. Patent Application No. 16 / 101,232, filed August 10, 2018. This real-time ray tracing hardware accelerator can be used to quickly and efficiently determine the location and extent of objects (e.g., within a world model) to generate real-time visualization simulations for RADAR signal interpretation, sound propagation synthesis and / or analysis, SONAR system simulation, general wave propagation simulation, comparison with LiDAR data for localization and / or other functional purposes, and / or for other uses. In some embodiments, one or more Tree Traversal Units (TTUs) may be used to perform one or more ray tracing-related operations.

[0107] Accelerators 514 (e.g., hardware accelerator clusters) have broad applications in autonomous driving. PVAs can be programmable vision accelerators used in critical processing stages of ADAS and autonomous vehicles. The capabilities of PVAs are a good match for algorithmic domains requiring predictable processing, low power, and low latency. In other words, PVAs perform well in semi-dense or dense rule computation, even on small datasets requiring predictable runtimes with low latency and low power. Therefore, in the context of platforms for autonomous vehicles, PVAs are designed to run classical computer vision algorithms because they are efficient in object detection and integer mathematical operations.

[0108] For example, according to one embodiment of this technology, PVA is used to perform computer stereo vision. In some examples, semi-global matching-based algorithms may be used, but this is not intended to be limiting. Many applications for Level 3-5 autonomous driving require instantaneous motion estimation / stereo matching (e.g., from moving structures, pedestrian recognition, lane detection, etc.). PVA can perform computer stereo vision functions on input from two monocular cameras.

[0109] In some examples, PVA can be used to perform intensive optical flow, providing processed RADAR data from the raw RADAR data (e.g., using 4D Fast Fourier Transform). In other examples, PVA is used for time-of-flight depth processing, which, for example, involves processing raw time-of-flight data to provide processed time-of-flight data.

[0110] DLA can be used to run any type of network to enhance control and driving safety, including, for example, neural networks that output a confidence metric for each object detection. Such a confidence value can be interpreted as a probability or as providing a relative “weight” for each detection compared to other detections. This confidence value allows the system to make further decisions about which detections should be considered true positives rather than false positives. For example, the system can set a threshold for the confidence and only consider detections exceeding the threshold as true positives. In an Automatic Emergency Braking (AEB) system, false positives can cause the vehicle to automatically perform emergency braking, which is clearly undesirable. Therefore, only the most confident detections should be considered as triggers for AEB. DLA can run a neural network to regress the confidence value. This neural network can take at least a subset of parameters as its input, such as bounding box dimensions, ground plane estimates (e.g., from another subsystem), inertial measurement unit (IMU) sensor 566 outputs related to vehicle orientation and distance, 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LiDAR sensor 564 or RADAR sensor 560), etc.

[0111] SoC 504 may include one or more data storage units 516 (e.g., memory). Data storage units 516 may be on-chip memory of SoC 504, which may store neural networks to be executed on the GPU and / or DLA. In some examples, for redundancy and security, data storage units 516 may be large enough to store multiple instances of the neural network. Data storage units 512 may include L2 or L3 cache 512. References to data storage units 516 may include references to memory associated with PVA, DLA, and / or other accelerators 514 as described herein.

[0112] SoC 504 may include one or more processors 510 (e.g., embedded processors). Processor 510 may include a startup and power management processor, which may be a dedicated processor and subsystem for handling startup power and management functions, as well as safety implementation. The startup and power management processor may be part of the SoC 504 startup sequence and may provide runtime power management services. The startup power and management processor may provide clock and voltage programming, auxiliary system low-power state transitions, SoC 504 thermal and temperature sensor management, and / or SoC 504 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to the temperature, and SoC 504 may use the ring oscillator to detect the temperature of CPU 506, GPU 508, and / or accelerator 514. If it is determined that the temperature exceeds a threshold, the startup and power management processor may enter a temperature fault routine and place SoC 504 into a lower power state and / or place vehicle 500 into a driver-safe parking mode (e.g., safely stop vehicle 500).

[0113] The processor 510 may further include a set of embedded processors that can be used as an audio processing engine. The audio processing engine can be an audio subsystem that allows for full hardware support for multi-channel audio via multiple interfaces, as well as a wide and flexible range of audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core with a digital signal processor and dedicated RAM.

[0114] The processor 510 may further include an always-on-processor engine that can provide the necessary hardware features to support low-power sensor management and wake-up use cases. This always-on-processor engine may include a processor core, tightly coupled RAM, support for peripherals (such as timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0115] Processor 510 may further include a security cluster engine, which includes a dedicated processor subsystem for handling security management for automotive applications. The security cluster engine may include two or more processor cores, tightly coupled RAM, support for peripheral devices (e.g., timers, interrupt controllers, etc.), and / or routing logic. In secure mode, the two or more cores may operate in lockstep mode and function as a single core with comparison logic that detects any differences between their operations.

[0116] The processor 510 may further include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0117] The processor 510 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.

[0118] Processor 510 may include a video image compositer, which may be (e.g., implemented on a microprocessor) a processing block, implementing video post-processing functions required by the video playback application to generate the final image for the player window. The video image compositer may perform lens distortion correction on the wide-angle camera 570, the surround camera 574, and / or the in-cabin monitoring camera sensor. The in-cabin monitoring camera sensor is preferably monitored by a neural network running on another instance of an advanced SoC, configured to recognize in-cabin events and respond accordingly. The in-cabin system may perform lip reading to activate mobile phone services and make calls, dictate emails, change vehicle destinations, activate or change the vehicle's infotainment system and settings, or provide voice-activated web browsing. Some functions are only available to the driver when the vehicle is operating in autonomous mode and are disabled in other situations.

[0119] Video image compositers can include enhanced temporal denoising for both spatial and temporal noise reduction. For example, in the case of motion in the video, denoising appropriately weights spatial information, reducing the weight of information provided by neighboring frames. In cases where the image or part of the image does not contain motion, the temporal denoising performed by the video image compositer can use information from previous images to reduce noise in the current image.

[0120] The video image compositer can also be configured to perform stereo correction on input stereo lens frames. When the operating system desktop is in use and the GPU 508 does not need to continuously render new surfaces, the video image compositer can be further used for user interface components. Even when the GPU 508 is powered on and activated, performing 3D rendering, the video image compositer can be used to offload the GPU 508 to improve performance and responsiveness.

[0121] SoC 504 may further include a Mobile Industry Processor Interface (MIPI) camera serial interface, a high-speed interface, and / or a video input block that can be used for camera and related pixel input functions for receiving video and input from a camera. SoC 504 may further include an input / output controller that can be software-controlled and can be used to receive I / O signals not assigned to a specific role.

[0122] SoC 504 may further include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management and / or other devices. SoC 504 can be used to process data from cameras and sensors (e.g., LIDAR sensor 564, RADAR sensor 560, etc., which can be connected via Gigabit Multimedia Serial Link and Ethernet), data from bus 502 (e.g., vehicle 500 speed, steering wheel position, etc.), and data from GNSS sensor 558 (connected via Ethernet or CAN bus). SoC 504 may further include a dedicated high-performance, high-capacity memory controller, which may include its own DMA engine, and which can be used to free up CPU 506 from routine data management tasks.

[0123] SoC 504 can be an end-to-end platform with a flexible architecture spanning Automation Levels 3-5, providing a comprehensive functional safety architecture that leverages and efficiently utilizes computer vision and ADAS technologies for diversity and redundancy, along with deep learning tools to deliver a flexible and reliable driving software stack. SoC 504 can be faster, more reliable, and even more energy- and space-efficient than conventional systems. For example, when combined with CPU 506, GPU 508, and data storage 516, accelerator 514 can provide a fast and efficient platform for Level 3-5 autonomous vehicles.

[0124] Therefore, this technology offers capabilities and functionalities that cannot be achieved through conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using high-level programming languages ​​such as C to execute a wide variety of processing algorithms across a diverse range of visual data. However, CPUs often cannot meet the performance requirements of many computer vision applications, such as those related to execution time and power consumption. In particular, many CPUs cannot execute complex object detection algorithms in real time, which is a requirement for automotive ADAS applications and practical Level 3-5 autonomous vehicles.

[0125] In contrast to conventional systems, the techniques described in this paper, by providing CPU complexes, GPU complexes, and hardware acceleration clusters, allow multiple neural networks to be executed simultaneously and / or sequentially, and the results to be combined to achieve Level 3–5 autonomous driving capabilities. For example, a CNN executed on a DLA or dGPU (e.g., GPU 520) could include text and word recognition, allowing a supercomputer to read and understand traffic signs, including those for which neural networks have not yet been specifically trained. The DLA could further include a neural network capable of recognizing, interpreting, and providing semantic understanding of the signs, and passing that semantic understanding to a path planning module running on the CPU complex.

[0126] As another example, multiple neural networks can operate simultaneously, as required for Level 3, 4, or 5 driving. For instance, a warning sign consisting of "Caution: Flashing lights indicate icy conditions," along with a light, can be interpreted independently or jointly by several neural networks. The sign itself can be recognized as a traffic sign by a deployed first neural network (e.g., a trained neural network), and the text "Flashing lights indicate icy conditions" can be interpreted by a deployed second neural network, which informs the vehicle's path planning software (preferably executing on a CPU complex) that icy conditions exist when the flashing lights are detected. The flashing lights can be identified by a deployed third neural network operating across multiple frames, which informs the vehicle's path planning software of the presence (or absence) of the flashing lights. All three neural networks can operate simultaneously, for example, within a DLA and / or on a GPU 508.

[0127] In some examples, the CNN used for facial recognition and owner identification can use data from camera sensors to identify the presence of an authorized driver and / or owner of vehicle 500. A processing engine always on the sensors can be used to unlock the vehicle and turn on the lights when the owner approaches the driver's door, and in safe mode, to disable the vehicle when the owner leaves. In this way, SoC 504 provides security against theft and / or carjacking.

[0128] In another example, the CNN used for emergency vehicle detection and identification can use data from microphone 596 to detect and identify emergency vehicle siren. In contrast to conventional systems that use a general classifier to detect siren and manually extract features, SoC 504 uses a CNN to classify environmental and urban sounds as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative shut-off rate of emergency vehicles (e.g., by using the Doppler effect). The CNN can also be trained to identify emergency vehicles specific to the localized area in which the vehicle operates, as identified by GNSS sensor 558. Thus, for example, when operating in Europe, the CNN could seek to detect European siren, and when operating in the United States, the CNN could seek to identify siren only in North America. Once an emergency vehicle is detected, with the assistance of ultrasonic sensor 562, the control program can be used to execute emergency vehicle safety routines, causing the vehicle to slow down, pull over to the side of the road, stop, and / or idle until the emergency vehicle passes.

[0129] The vehicle may include a CPU 518 (e.g., a discrete CPU or dCPU) that can be coupled to the SoC 504 via a high-speed interconnect (e.g., PCIe). The CPU 518 may include, for example, an x86 processor. The CPU 518 can be used to perform any of a wide variety of functions, including, for example, arbitrating the results of potential inconsistencies between ADAS sensors and the SoC 504, and / or monitoring the status and health of the controller 536 and / or the infotainment SoC 530.

[0130] Vehicle 500 may include a GPU 520 (e.g., a discrete GPU or dGPU) that can be coupled to SoC 504 via a high-speed interconnect (e.g., NVIDIA's NVLINK). GPU 520 may provide additional artificial intelligence capabilities, for example by executing redundant and / or different neural networks, and can be used to train and / or update neural networks based on inputs from sensors of vehicle 500 (e.g., sensor data).

[0131] Vehicle 500 may further include a network interface 524, which may include one or more wireless antennas 526 (e.g., one or more wireless antennas for different communication protocols, such as cellular antennas, Bluetooth antennas, etc.). Network interface 524 can be used to enable wireless connectivity via the Internet to the cloud (e.g., with server 578 and / or other network devices), with other vehicles, and / or with computing devices (e.g., a passenger's client device). For communication with other vehicles, a direct link can be established between the two vehicles, and / or an indirect link can be established (e.g., across networks and via the Internet). A direct link can be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link can provide vehicle 500 with information about vehicles approaching vehicle 500 (e.g., vehicles in front, to the side, and / or behind vehicle 500). This functionality can be part of vehicle 500's cooperative adaptive cruise control function.

[0132] Network interface 524 may include a SoC that provides modulation and demodulation functions and enables controller 536 to communicate via a wireless network. Network interface 524 may include an RF front-end for up-conversion from baseband to RF and down-conversion from RF to baseband. Frequency conversion can be performed using known processes and / or using a superheterodyne process. In some examples, the RF front-end functionality may be provided by a separate chip. The network interface may include wireless functions for communication via LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-Wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0133] Vehicle 500 may further include data storage 528, which may include off-chip (e.g., off-chip SoC 504) storage devices. Data storage 528 may include one or more storage elements, including RAM, SRAM, DRAM, VRAM, flash memory, hard disk, and / or other components and / or devices capable of storing at least one bit of data.

[0134] Vehicle 500 may further include GNSS sensor 558. GNSS sensor 558 (e.g., GPS, assisted GPS sensor, differential GPS (DGPS) sensor, etc.) is used for assisted mapping, sensing, occupancy grid generation, and / or path planning functions. Any number of GNSS sensors 558 can be used, including, for example, but not limited to, GPS using a USB connector with an Ethernet-to-serial (RS-232) bridge.

[0135] Vehicle 500 may further include a RADAR sensor 560. The RADAR sensor 560 can be used by vehicle 500 for remote vehicle detection even in dark and / or inclement weather conditions. The RADAR functional safety level may be ASIL B. The RADAR sensor 560 can use CAN and / or bus 502 (e.g., to transmit data generated by the RADAR sensor 560) for control and access to object tracking data, and in some examples, Ethernet access for accessing raw data. A wide variety of RADAR sensor types can be used. For example, and without limitation, the RADAR sensor 560 can be adapted for front, rear, and side RADAR use. In some examples, a pulse Doppler RADAR sensor is used.

[0136] The RADAR sensor 560 can include different configurations, such as long-range with a narrow field of view, short-range with a wide field of view, short-range side coverage, etc. In some examples, the long-range RADAR can be used for adaptive cruise control functions. The long-range RADAR system can provide a wide field of view (e.g., within 250m) achieved through two or more independent scans. The RADAR sensor 560 can help distinguish between stationary and moving objects and can be used by ADAS systems for emergency braking assist and forward collision warning. The long-range RADAR sensor can include a single-site multi-mode RADAR with multiple (e.g., six or more) fixed RADAR antennas and high-speed CAN and FlexRay interfaces. In an example with six antennas, the four central antennas can create a focused beam pattern designed to record the vehicle 500's surroundings at higher rates with minimal traffic interference from adjacent lanes. The other two antennas can extend the field of view, enabling rapid detection of vehicles entering or leaving the vehicle 500's lane.

[0137] As an example, a mid-range RADAR system can include a range of up to 560m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 550 degrees (rear). Short-range RADAR systems can include, but are not limited to, RADAR sensors designed to be mounted at both ends of the rear bumper. When mounted at both ends of the rear bumper, such a RADAR sensor system can create two beams that continuously monitor blind spots behind and beside the vehicle.

[0138] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0139] Vehicle 500 may further include ultrasonic sensors 562. Ultrasonic sensors 562, which may be positioned at the front, rear, and / or sides of vehicle 500, can be used for parking assistance and / or creating and updating occupancy grids. A wide variety of ultrasonic sensors 562 can be used, and different ultrasonic sensors 562 can be used for different detection ranges (e.g., 2.5m, 4m). Ultrasonic sensors 562 can operate at functional safety level ASIL B.

[0140] Vehicle 500 may include a LIDAR sensor 564. The LIDAR sensor 564 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The LIDAR sensor 564 may be of functional safety level ASIL B. In some examples, vehicle 500 may include multiple LIDAR sensors 564 (e.g., two, four, six, etc.) that can use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0141] In some examples, the LiDAR sensor 564 may be able to provide a list of objects and their distances within a 360-degree field of view. Commercially available LiDAR sensors 564 may have an advertising range of, for example, approximately 500m, with an accuracy of 2cm-3cm, and support for 500Mbps Ethernet connectivity. In some examples, one or more non-protruding LiDAR sensors 564 may be used. In such examples, the LiDAR sensor 564 may be implemented as a small device that can be embedded in the front, rear, sides, and / or corners of a vehicle 500. In such examples, the LiDAR sensor 564 may provide a horizontal field of view of up to 120 degrees and a vertical field of view of 35 degrees, even for low-reflectivity objects, with a range of 200m. Front-mounted LiDAR sensors 564 may be configured for a horizontal field of view between 45 degrees and 135 degrees.

[0142] In some examples, LiDAR technologies such as 3D flash LiDAR can also be used. 3D flash LiDAR uses a flash of laser light as the emission source to illuminate the vehicle's surroundings up to approximately 200 meters. A flash LiDAR unit includes a receiver that records the laser pulse propagation time and reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LiDAR allows for the generation of highly accurate and distortion-free images of the surrounding environment using each laser flash. In some examples, four flash LiDAR sensors can be deployed, one on each side of the vehicle. Available 3D flash LiDAR systems include solid-state 3D staring array LiDAR cameras (e.g., non-browsing LiDAR devices) without moving parts other than a fan. Flash LiDAR devices can use 5 nanosecond Class I (eye-safe) laser pulses per frame and can capture reflected laser light in the form of a 3D range point cloud and co-registered intensity data. By using a flash LiDAR, and because a flash LiDAR is a solid-state device with no moving parts, the LiDAR sensor 564 is less susceptible to motion blur, vibration, and / or shock.

[0143] The vehicle may further include an IMU sensor 566. In some examples, the IMU sensor 566 may be located at the center of the rear axle of the vehicle 500. The IMU sensor 566 may include, for example, but not limited to, an accelerometer, a magnetometer, a gyroscope, a magnetic compass, and / or other sensor types. In some examples, such as in a six-axis application, the IMU sensor 566 may include an accelerometer and a gyroscope, while in a nine-axis application, the IMU sensor 566 may include an accelerometer, a gyroscope, and a magnetometer.

[0144] In some embodiments, the IMU sensor 566 can be implemented as a miniature, high-performance GPS-assisted inertial navigation system (GPS / INS) that combines a microelectromechanical system (MEMS) inertial sensor, a high-sensitivity GPS receiver, and an advanced Kalman filter algorithm to provide estimates of position, velocity, and attitude. Thus, in some examples, the IMU sensor 566 can enable the vehicle 500 to estimate heading by directly observing and correlating velocity changes from GPS to the IMU sensor 566 without input from a magnetic sensor. In some examples, the IMU sensor 566 and the GNSS sensor 558 can be combined into a single integrated unit.

[0145] The vehicle may include a microphone 596 placed in and / or around the vehicle 500. Among other things, the microphone 596 may be used for emergency vehicle detection and identification.

[0146] The vehicle may further include any number of camera types, including stereo camera 568, wide-angle camera 570, infrared camera 572, surround camera 574, long-range and / or mid-range camera 598, and / or other camera types. These cameras can be used to capture image data around the entire perimeter of the vehicle 500. The camera types used depend on the embodiment and the requirements of the vehicle 500, and any combination of camera types can be used to provide the necessary coverage around the vehicle 500. Furthermore, the number of cameras may vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. As an example and without limitation, these cameras may support Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the cameras is described herein with respect to... Figure 5A and Figure 5B It was described in more detail.

[0147] Vehicle 500 may further include vibration sensor 542. Vibration sensor 542 can measure vibrations of vehicle components such as axles. For example, changes in vibration can indicate changes in the road surface. In another example, when two or more vibration sensors 542 are used, differences between vibrations can be used to determine friction or slippage on the road surface (e.g., when there is a vibration difference between a power drive shaft and a free-rotating shaft).

[0148] Vehicle 500 may include ADAS system 538. In some examples, ADAS system 538 may include SoC. ADAS system 538 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward collision warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keeping assist (LKA), blind spot warning (BSW), rear cross traffic warning (RCTW), collision warning system (CWS), lane centering (LC) and / or other features and functions.

[0149] The ACC system can use a RADAR sensor 560, a LIDAR sensor 564, and / or a camera. The ACC system can include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to vehicles immediately in front of vehicle 500 and automatically adjusts the vehicle speed to maintain a safe distance. Lateral ACC performs distance holding and, if necessary, advises vehicle 500 to change lanes. Lateral ACC is associated with other ADAS applications such as LCA and CWS.

[0150] CACC uses information from other vehicles, which can be received indirectly from other vehicles via a wireless link or network connection (e.g., via the Internet) through network interface 524 and / or wireless antenna 526. Direct links can be provided by vehicle-to-vehicle (V2V) communication links, while indirect links can be infrastructure-to-vehicle (I2V) communication links. Typically, the V2V communication concept provides information about vehicles immediately ahead (e.g., vehicles immediately in front of vehicle 500 and in the same lane), while the I2V communication concept provides information about traffic further ahead. A CACC system can include either or both of these I2V and V2V information sources. Given information about vehicles ahead of vehicle 500, CACC can be more reliable, and it has the potential to improve traffic flow and reduce road congestion.

[0151] The Forward Collision Warning (FCW) system is designed to alert the driver to hazards, enabling the driver to take corrective action. The FCW system uses a front-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components. The FCW system can provide warnings in the form of, for example, audible, visual, haptic, and / or rapid braking pulses.

[0152] An AEB (Autonomous Emergency Braking) system detects an impending forward collision with another vehicle or other object and can automatically apply the brakes if the driver does not take corrective action within a specified time or distance parameter. The AEB system can use a front-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC. When the AEB system detects a hazard, it typically first alerts the driver to take corrective action to avoid a collision, and if the driver does not take corrective action, the AEB system can automatically apply the brakes to attempt to prevent or at least mitigate the effects of the predicted collision. The AEB system may include technologies such as dynamic brake support and / or collision proximity braking.

[0153] The Lane Departure Warning (LDW) system provides visual, auditory, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle crosses lane markings. When the driver indicates intentional lane departure, the LDW system is deactivated by activating a turn signal. The LDW system can utilize a front-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0154] The LKA system is a variation of the LDW system. If vehicle 500 begins to leave the lane, the LKA system provides corrective steering input or braking to vehicle 500.

[0155] The BSW system detects and warns the driver of vehicles in the vehicle's blind spot. The BSW system can provide visual, auditory, and / or tactile alerts to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses turn signals. The BSW system can utilize a rear-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0156] RCTW systems can provide visual, auditory, and / or tactile notifications when an object is detected outside the range of a rear-view camera while the vehicle is reversing. Some RCTW systems include AEB (Autonomous Emergency Braking) to ensure the application of the vehicle's brakes to avoid a collision. RCTW systems may use one or more rear-view RADAR sensors 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components.

[0157] Conventional ADAS systems can be prone to false positives, which can be annoying and distracting for the driver, but typically not catastrophic, as they alert the driver and allow them to determine whether a safe condition truly exists and take appropriate action. However, in an autonomous vehicle 500, in the event of conflicting results, the vehicle 500 itself must decide whether to heed the results from the main computer or auxiliary computer (e.g., the first controller 536 or the second controller 536). For example, in some embodiments, the ADAS system 538 may be a backup and / or auxiliary computer used to provide perception information to a backup computer rationality module. The backup computer rationality monitor may run redundant and varied software on hardware components to detect faults in perception and dynamic driving tasks. Outputs from the ADAS system 538 may be provided to a supervisory MCU. If the outputs from the main computer and the auxiliary computer conflict, the supervisory MCU must determine how to reconcile the conflict to ensure safe operation.

[0158] In some examples, the master computer can be configured to provide a confidence score to the supervisory MCU, indicating the master computer's confidence level in the selected result. If the confidence score exceeds a threshold, the supervisory MCU can follow the master computer's direction regardless of whether the auxiliary computer provides conflicting or inconsistent results. If the confidence score does not meet the threshold and the master and auxiliary computers indicate different results (e.g., conflict), the supervisory MCU can arbitrate between these computers to determine the appropriate result.

[0159] The supervisory MCU can be configured to run a neural network trained and configured to determine the conditions under which the auxiliary computer provides a false alarm based on outputs from both the host and auxiliary computers. Thus, the neural network in the supervisory MCU can learn when the output of the auxiliary computer can be trusted and when it cannot. For example, when the auxiliary computer is a RADAR-based FCW system, the neural network in the supervisory MCU can learn when the FCW system is identifying a metallic object that is not actually dangerous, such as a drain grid or manhole cover that triggers an alarm. Similarly, when the auxiliary computer is a camera-based LDW system, the neural network in the supervisory MCU can learn to ignore the LDW when a cyclist or pedestrian is present and lane departure is actually the safest strategy. In embodiments that include a neural network running on the supervisory MCU, the supervisory MCU may include at least one of a DLA or GPU suitable for running the neural network using associated memory. In a preferred embodiment, the supervisory MCU may include components of SoC 504 and / or be included as components of SoC 504.

[0160] In other examples, ADAS system 538 may include an auxiliary computer that performs ADAS functions using conventional computer vision rules. This allows the auxiliary computer to use classic computer vision rules (if-then), and the presence of neural networks in the supervising MCU can improve reliability, safety, and performance. For example, diverse implementations and intentional non-identity make the entire system more fault-tolerant, especially for failures caused by software (or software-hardware interface) functionality. For instance, if a software vulnerability or bug exists in the software running on the host computer and non-identical software code running on the auxiliary computer provides the same overall result, the supervising MCU can be more confident that the overall result is correct and that the vulnerability in the software or hardware on the host computer does not cause a substantial error.

[0161] In some examples, the output of ADAS system 538 can be fed to the perception block and / or the dynamic driving task block of the main computer. For example, if ADAS system 538 issues a forward collision warning because an object is immediately in front, the perception block can use this information when recognizing the object. In other examples, the assistance computer can have its own neural network, which is trained and thus reduces the risk of false positives as described herein.

[0162] Vehicle 500 may further include an infotainment SoC 530 (e.g., an in-vehicle infotainment system (IVI)). Although illustrated and described as an SoC, the infotainment system may not be an SoC and may include two or more discrete components. The infotainment SoC 530 may include a combination of hardware and software that can be used to provide vehicle 500 with audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming media, etc.), telephone (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.) and / or information services (e.g., navigation system, rear parking assistance, radio data system, vehicle-related information such as fuel level, total coverage distance, brake fuel level, fuel level, door opening / closing, air filter information, etc.). For example, the infotainment SoC 530 may include a radio, disc player, navigation system, video player, USB and Bluetooth connectivity, in-vehicle computer, in-vehicle entertainment, Wi-Fi, steering wheel audio controls, hands-free voice controls, head-up display (HUD), HMI display 534, telematics device, control panel (e.g., for controlling and / or interacting with various components, features, and / or systems) and / or other components. The infotainment SoC 530 may further be used to provide information (e.g., visual and / or auditory) to the vehicle's users, such as information from the ADAS system 538, autonomous driving information such as planned vehicle maneuvers, trajectories, surrounding environment information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0163] The infotainment SoC 530 may include GPU functionality. The infotainment SoC 530 can communicate with other devices, systems, and / or components of the vehicle 500 via bus 502 (e.g., CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 530 may be coupled to a supervisory MCU, allowing the GPU of the infotainment system to perform some autonomous driving functions in the event of a failure of the main controller 536 (e.g., the primary and / or backup computer of the vehicle 500). In such an example, the infotainment SoC 530 may place the vehicle 500 into a driver-safe parking mode as described herein.

[0164] Vehicle 500 may further include an instrument cluster 532 (e.g., a digital instrument panel, electronic instrument cluster, digital instrument panel, etc.). The instrument cluster 532 may include a controller and / or a supercomputer (e.g., a discrete controller or supercomputer). The instrument cluster 532 may include a set of instruments such as a speedometer, fuel level, oil pressure, tachometer, odometer, turn indicator, shift position indicator, seatbelt warning light, parking brake warning light, engine malfunction indicator, airbag (SRS) system information, lighting controls, safety system controls, navigation information, etc. In some examples, information may be displayed and / or shared between the infotainment SoC 530 and the instrument cluster 532. In other words, the instrument cluster 532 may be included as part of the infotainment SoC 530, or vice versa.

[0165] Figure 5D For cloud-based servers and according to some embodiments of this disclosure Figure 5A This is a system diagram illustrating communication between example autonomous vehicles 500. System 576 may include server 578, network 590, and vehicles including vehicle 500. Server 578 may include multiple GPUs 584(A)-584(H) (collectively referred to herein as GPU 584), PCIe switches 582(A)-582(H) (collectively referred to herein as PCIe switch 582), and / or CPUs 580(A)-580(B) (collectively referred to herein as CPU 580). GPUs 584, CPUs 580, and PCIe switches may be interconnected with high-speed interconnects and / or PCIe connections 586, such as, but not limited to, NVLink interfaces 588 developed by NVIDIA. In some examples, GPUs 584 are connected via NVLink and / or NVSwitch SoCs, and GPUs 584 and PCIe switches 582 are connected via PCIe interconnects. Although eight GPUs 584, two CPUs 580, and two PCIe switches are shown in the diagram, this is not intended to be limiting. Depending on the embodiment, each of the servers 578 may include any number of GPUs 584, CPUs 580, and / or PCIe switches. For example, each of the servers 578 may include eight, sixteen, thirty-two, and / or more GPUs 584.

[0166] Server 578 can receive image data from vehicles via network 590, representing images of unexpected or changed road conditions such as recently commenced roadworks. Server 578 can also transmit neural network 592, updated neural network 592, and / or map information 594, including information about traffic and road conditions, to vehicles via network 590. Updates to map information 594 may include updates to HD map 522, such as information about construction sites, potholes, bends, floods, or other obstacles. In some examples, neural network 592, updated neural network 592, and / or map information 594 may have been generated from new training and / or data received from any number of vehicles in the environment, and / or based on experience gained from training performed at a data center (e.g., using server 578 and / or other servers).

[0167] Server 578 can be used to train machine learning models (e.g., neural networks) based on training data. Training data can be generated by the vehicle and / or generated in a simulation (e.g., using a game engine). In some examples, the training data is labeled (e.g., where the neural network benefits from supervised learning) and / or undergoes other preprocessing, while in other examples, the training data is not labeled and / or preprocessed (e.g., where the neural network does not require supervised learning). Training can be performed according to any one or more classes of machine learning techniques, including but not limited to: supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, joint learning, transfer learning, feature learning (including principal component analysis and cluster analysis), multilinear subspace learning, manifold learning, representation learning (including alternative dictionary learning), rule-based machine learning, anomaly detection, and any variations or combinations thereof. Once the machine learning model is trained, it can be used by the vehicle (e.g., transmitted to the vehicle via network 590), and / or the machine learning model can be used by server 578 to remotely monitor the vehicle.

[0168] In some examples, server 578 can receive data from a vehicle and apply that data to a state-of-the-art real-time neural network for real-time intelligent inference. Server 578 may include a deep learning supercomputer powered by GPU 584 and / or a dedicated AI computer, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 578 may include a deep learning infrastructure in a data center that uses only CPU power.

[0169] The deep learning infrastructure of server 578 may be capable of rapid, real-time inference and can be used to assess and verify the health status of the processor, software, and / or associated hardware in vehicle 500. For example, the deep learning infrastructure may receive periodic updates from vehicle 500, such as image sequences and / or objects located in those image sequences by vehicle 500 (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural network to identify objects and compare them with objects identified by vehicle 500. If the results do not match and the infrastructure concludes that the AI ​​in vehicle 500 has malfunctioned, then server 578 may transmit a signal to vehicle 500 instructing its fail-safe computer to take control, notify passengers, and complete a safe stopping operation.

[0170] For inference, server 578 may include GPU 584 and one or more programmable inference accelerators (such as NVIDIA's TensorRT 3). The combination of a GPU-powered server and inference acceleration enables real-time response. In other examples, such as where performance is less critical, CPU, FPGA, and other processor-powered servers can be used for inference.

[0171] Example computing device

[0172] Figure 6 The block diagram is provided for an example computing device 600 suitable for implementing some embodiments of the present disclosure. The computing device 600 may include an interconnect system 602 directly or indirectly coupled to the following devices: memory 604, one or more central processing units (CPUs) 606, one or more graphics processing units (GPUs) 608, a communication interface 610, input / output (I / O) ports 612, input / output components 614, a power supply 616, one or more presentation components 618 (e.g., a display), and one or more logic units 620. In at least one embodiment, the computing device 600 may include one or more virtual machines (VMs), and / or any component thereof may include virtual components (e.g., virtual hardware components). For a non-limiting example, one or more GPUs 608 may include one or more vGPUs, one or more CPUs 606 may include one or more vCPUs, and / or one or more logic units 620 may include one or more virtual logic units. Therefore, computing device 600 may include discrete components (e.g., a complete GPU dedicated to computing device 600), virtual components (e.g., a portion of the GPU dedicated to computing device 600), or a combination thereof.

[0173] although Figure 6The various boxes are shown connected via an interconnect system 602 with wiring, but this is not intended to be limiting and is merely for clarity. For example, in some embodiments, a presentation component 618, such as a display device, may be considered an I / O component 614 (e.g., if the display is a touchscreen). As another example, CPU 606 and / or GPU 608 may include memory (e.g., memory 604 may represent a storage device other than the memory of GPU 608, CPU 606, and / or other components). In other words, Figure 6 The computing devices mentioned are merely illustrative. No distinction is made between categories such as "workstation," "server," "laptop," "desktop," "tablet," "client device," "mobile device," "handheld device," "game console," "electronic control unit (ECU)," "virtual reality system," and / or other device or system types, as all of these are considered within the same category. Figure 6 Within the scope of computing devices.

[0174] Interconnect system 602 may represent one or more links or buses, such as address buses, data buses, control buses, or combinations thereof. Interconnect system 602 may include one or more link or bus types, such as Industry Standard Architecture (ISA) bus, Extended Industry Standard Architecture (EISA) bus, Video Electronics Standards Association (VESA) bus, Peripheral Component Interconnect (PCI) bus, Peripheral Component Interconnect Fast (PCIe) bus, and / or another type of bus or link. In some embodiments, there is a direct connection between components. As an example, CPU 606 may be directly connected to memory 604. Furthermore, CPU 606 may be directly connected to GPU 608. In cases where there is a direct or point-to-point connection between components, interconnect system 602 may include a PCIe link to perform the connection. In these examples, a PCI bus is not required in computing device 600.

[0175] Memory 604 may include any of a wide variety of computer-readable media. Computer-readable media can be any available medium that can be accessed by computing device 600. Computer-readable media may include volatile and non-volatile media, as well as removable and non-removable media. For example and without limitation, computer-readable media may include computer storage media and communication media.

[0176] Computer storage media may include volatile and non-volatile media and / or removable and non-removable media, implemented in any way or by any method or technique for storing information such as computer-readable instructions, data structures, program modules, and / or other data types. For example, memory 604 may store computer-readable instructions (e.g., representing programs and / or program elements, such as an operating system). Computer storage media may include, but is not limited to, RAM, ROM, EEPROM, flash memory or other storage technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage devices, magnetic tape cassettes, magnetic tape, disk storage devices or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by computing device 600. As used herein, computer storage media does not include the signal itself.

[0177] Computer storage media may contain computer-readable instructions, data structures, program modules, and / or other data types in modulated data signals such as carrier waves or other transmission mechanisms, and include any information transport medium. The term "modulated data signal" can refer to a signal whose characteristics are set or altered in a manner that encodes information into that signal. For example and without limitation, computer storage media may include wired media such as wired networks or direct wired connections, and wireless media such as sound, RF, infrared, and other wireless media. Any combination of the above should also be included within the scope of computer-readable media.

[0178] CPU 606 may be configured to execute at least some of computer-readable instructions to control one or more components of computing device 600 to perform one or more of the methods and / or processes described herein. Each of CPU 606 may include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of processing a large number of software threads simultaneously. CPU 606 may include any type of processor and may include different types of processors depending on the type of computing device 600 implemented (e.g., processors with fewer cores for mobile devices and processors with more cores for servers). For example, depending on the type of computing device 600, the processor may be an advanced RISC mechanism (ARM) processor implemented using Reduced Instruction Set Computing (RISC) or an x86 processor implemented using Complex Instruction Set Computing (CISC). In addition to one or more microprocessors or supplementary coprocessors such as math coprocessors, computing device 600 may also include one or more CPUs 606.

[0179] In addition to or replacing CPU 606, GPU 608 may also be configured to execute at least some computer-readable instructions to control one or more components of computing device 600 to perform one or more of the methods and / or processes described herein. One or more GPUs 608 may be integrated GPUs (e.g., having one or more CPUs 606) and / or one or more GPUs 608 may be discrete GPUs. In embodiments, one or more GPUs 608 may be coprocessors of one or more CPUs 606. Computing device 600 may use GPU 608 to render graphics (e.g., 3D graphics) or perform general-purpose computing. For example, GPU 608 may be used for general-purpose computing on a GPU (GPGPU). GPU 608 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. GPU 608 may generate pixel data for outputting an image in response to rendering commands (e.g., rendering commands received via a host interface from CPU 606). GPU 608 may include graphics memory, such as display memory, for storing pixel data or any other suitable data (e.g., GPGPU data). Display memory may be included as part of memory 604. GPU 608 may include two or more GPUs operating in parallel (e.g., via links). The links may connect the GPUs directly (e.g., using NVLINK) or via a switch (e.g., using NVSwitch). When combined, each GPU 608 may generate different portions of pixel data or GPGPU data for different outputs (e.g., the first GPU for the first image, the second GPU for the second image). Each GPU may include its own memory or may share memory with other GPUs.

[0180] In addition to or replacing CPU 606 and / or GPU 608, logic unit 620 may be configured to execute at least some computer-readable instructions to control one or more components of computing device 600 to perform one or more methods and / or processes described herein. In embodiments, CPU 606, GPU 608, and / or logic unit 620 may execute any combination of methods, processes, and / or portions thereof discretely or jointly. One or more logic units 620 may be part of and / or integrated into one or more CPUs 606 and / or one or more GPUs 608, and / or one or more logic units 620 may be discrete components of CPU 606 and / or GPU 608 or otherwise external thereto. In embodiments, one or more logic units 620 may be processors of one or more CPUs 606 and / or one or more GPUs 608.

[0181] Examples of logic unit 620 include one or more processing cores and / or components thereof, such as data processing unit (DPU), tensor core (TC), tensor processing unit (TPU), pixel vision core (PVC), vision processing unit (VPU), graphics processing cluster (GPC), texture processing cluster (TPC), streaming multiprocessor (SM), tree traversal unit (TTU), artificial intelligence accelerator (AIA), deep learning accelerator (DLA), arithmetic logic unit (ALU)), application-specific integrated circuit (ASIC), floating-point unit (FPU), input / output (I / O) element, peripheral component interconnect (PCI) or peripheral component interconnect fast (PCIe) element, etc.

[0182] The communication interface 610 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 600 to communicate with other computing devices via electronic communication networks, including wired and / or wireless communications. The communication interface 610 may include components and functions that enable communication via any of several different networks, such as wireless networks (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), wired networks (e.g., communication via Ethernet or InfiniBand), low-power wide area networks (e.g., LoRaWAN, SigFox, etc.), and / or the Internet. In one or more embodiments, the logic unit 620 and / or the communication interface 610 may include one or more data processing units (DPUs) to directly transmit data received via a network and / or via interconnect system 602 to (e.g., memory) one or more GPUs 608.

[0183] I / O port 612 enables computing device 600 to be logically coupled to other devices, including I / O component 614, presentation component 618, and / or other components, some of which may be built into (e.g., integrated into) computing device 600. Illustrative I / O component 614 includes microphones, mice, keyboards, joysticks, game pads, game controllers, satellite dish antennas, browsers, printers, wireless devices, and so on. I / O component 614 can provide a Natural User Interface (NUI) for processing user-generated air gestures, voice, or other physiological input. In some instances, the input may be transmitted to appropriate network elements for further processing. The NUI can implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, on-screen and adjacent-screen gesture recognition, air gestures, head and eye tracking, and touch recognition associated with the display of computing device 600 (described in more detail below). Computing device 600 may include depth cameras such as stereo camera systems, infrared camera systems, RGB camera systems, touchscreen technology, and combinations thereof for gesture detection and recognition. In addition, computing device 600 may include an accelerometer or gyroscope that enables motion detection (e.g., as part of an inertial measurement unit (IMU)). In some examples, the output of the accelerometer or gyroscope may be used by computing device 600 to render immersive augmented reality or virtual reality.

[0184] Power supply 616 may include a hard-wired power supply, a battery power supply, or a combination thereof. Power supply 616 may supply power to computing device 600 so that the components of computing device 600 can operate.

[0185] The presentation component 618 may include a display (such as a monitor, touch screen, television screen, head-up display (HUD), other display types, or combinations thereof), speakers, and / or other presentation components. The presentation component 618 may receive data from other components (such as GPU 608, CPU 606, DPU, etc.) and output that data (e.g., as an image, video, sound, etc.).

[0186] Example Data Center

[0187] Figure 7 An example data center 700 is shown, which can be used in at least one embodiment of this disclosure. The data center 700 may include a data center infrastructure layer 710, a framework layer 720, a software layer 730, and an application layer 740.

[0188] like Figure 7As shown, the data center infrastructure layer 710 may include a resource coordinator 712, grouped computing resources 714, and node computing resources (“nodes CR”) 716(1)-716(N), where “N” represents any complete positive integer. In at least one embodiment, nodes CR 716(1)-716(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 processing units or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memory), storage devices (e.g., solid-state drives or disk drives), network input / output (NW I / O) devices, network switches, virtual machines (VMs), power modules, and cooling modules, etc. In some embodiments, one or more nodes CR 716(1)-716(N) may correspond to servers having one or more of the aforementioned computing resources. In addition, in some embodiments, nodes CR716(1)-716(N) may include one or more virtual components, such as vGPU, vCPU, etc., and / or one or more of nodes CR716(1)-716(N) may correspond to virtual machines (VMs).

[0189] In at least one embodiment, the grouped computing resources 714 may include individual groups (not shown) of nodes CR716 housed in one or more racks, or a plurality of racks (also not shown) housed in data centers in various geographic locations. Individual groups of nodes CR716 within the grouped computing resources 714 may include computing, networking, memory, or storage resources that can be configured or allocated to support groups of one or more workloads. In at least one embodiment, several nodes CR716, including CPUs, GPUs, DPUs, and / or other processors, may be grouped within one or more racks to provide computing resources to support one or more workloads. One or more racks may also include any number of power modules, cooling modules, and / or network switches in any combination.

[0190] Resource coordinator 712 can be configured or otherwise controlled to control one or more nodes CR716(1)-716(N) and / or groups of computing resources 714. In at least one embodiment, resource coordinator 712 may include a Software Design Infrastructure (SDI) management entity for data center 700. Resource coordinator 712 may include hardware, software, or some combination thereof.

[0191] In at least one embodiment, such as Figure 7As shown, framework layer 720 may include a job scheduler 732, a configuration manager 734, a resource manager 736, and a distributed file system 738. Framework layer 720 may include a framework of software 732 supporting software layer 730 and / or one or more applications 742 of application layer 740. Software 732 or application 742 may respectively include web-based service software or applications, such as service software or applications provided by Amazon Web Services, Google Cloud, and Microsoft Azure. Framework layer 720 may be, but is not limited to, a free and open-source software web application framework, such as Apache Spark, which can utilize distributed file system 738 for large-scale data processing (e.g., "big data"). TM (Hereinafter referred to as "Spark"). In at least one embodiment, the job scheduler 732 may include a Spark driver for facilitating the scheduling of workloads supported by various layers of data center 700. In at least one embodiment, the configuration manager 734 may be able to configure different layers, such as software layer 730 and framework layer 720 including Spark and a distributed file system 738 for supporting large-scale data processing. The resource manager 736 is able to manage cluster or grouped computing resources mapped to or allocated for supporting distributed file system 738 and job scheduler 732. In at least one embodiment, cluster or grouped computing resources may include grouped computing resources 714 at data center infrastructure layer 710. The resource manager 736 may coordinate with resource coordinator 712 to manage these mapped or allocated computing resources.

[0192] In at least one embodiment, the software 732 included in the software layer 730 may include software used by at least a portion of the nodes CR716(1)-716(N), the grouped computing resources 714, and / or the distributed file system 738 of the framework layer 720. One or more types of software may include, but are not limited to, Internet web page search software, email virus browsing software, database software, and streaming video content software.

[0193] In at least one embodiment, the application layer 740 may include one or more applications 742 that can be used by at least a portion of the nodes CR716(1)-716(N), the grouped computing resources 714, and / or the distributed file system 738 of the framework layer 720. The one or more types of applications may include, but are not limited to, any number of genomics applications, cognitive computing and machine learning applications, including training or inference software, machine learning framework software (e.g., PyTorch, TensorFlow, Caffe, etc.), and / or other machine learning applications used in conjunction with one or more embodiments.

[0194] In at least one embodiment, any of the configuration manager 734, resource manager 736, and resource coordinator 712 can perform any number and type of self-modification actions based on any amount and type of data acquired in any technically feasible manner. Self-modification actions can alleviate the risk of data center operators of data center 700 making potentially poor configuration decisions and can prevent underutilization and / or skewed portions of the data center.

[0195] Data center 700 may include tools, services, software, or other resources for training one or more machine learning models or using one or more machine learning models to predict or infer information according to one or more embodiments described herein. For example, a machine learning model can be trained by calculating weight parameters based on a neural network architecture using the software and computing resources described above with respect to data center 700. In at least one embodiment, by using weight parameters calculated through one or more training techniques, information can be inferred or predicted using trained machine learning models corresponding to one or more neural networks, such as, but not limited to, those described herein, using the resources described above with respect to data center 700.

[0196] In at least one embodiment, the data center 700 may use a CPU, application-specific integrated circuit (ASIC), GPU, FPGA, and / or other hardware (or corresponding virtual computing resources) to perform training and / or inference. Furthermore, one or more of the aforementioned software and / or hardware resources may be configured as a service to allow a user to train or perform information inference, such as image recognition, speech recognition, or other artificial intelligence services.

[0197] Example network environment

[0198] A network environment suitable for implementing embodiments of this disclosure may include one or more client devices, servers, network-attached storage (NAS), other backend devices, and / or other device types. Client devices, servers, and / or other device types (e.g., each device) may... Figure 6 The implementation is carried out on one or more instances of computing device 600—for example, each device may include similar components, features, and / or functions of computing device 600. Furthermore, in the case of implementing back-end devices (e.g., servers, NAS, etc.), the back-end devices may be included as part of data center 700, examples of which are described herein. Figure 7 To describe in more detail.

[0199] Components of a network environment can communicate with each other via a network, which can be wired, wireless, or both. A network can include multiple networks, or networks within multiple networks. For example, a network can include one or more wide area networks (WANs), one or more local area networks (LANs), one or more public networks (such as the Internet and / or the Public Switched Telephone Network (PSTN)), and / or one or more private networks. In cases where the network includes a wireless telecommunications network, components such as base stations, communication towers, or even access points (and other components) can provide wireless connectivity.

[0200] A compatible network environment may include one or more peer-to-peer network environments (in which case the server may not be included in the network environment) and one or more client-server network environments (in which case one or more servers may be included in the network environment). In a peer-to-peer network environment, the server functionality described herein can be implemented on any number of client devices.

[0201] In at least one embodiment, the network environment may include one or more cloud-based network environments, distributed computing environments, combinations thereof, etc. A cloud-based network environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more servers, which may include one or more core network servers and / or edge servers. The framework layer may include a framework for supporting software at the software layer and / or one or more applications at the application layer. The software or applications may respectively include network-based service software or applications. In embodiments, one or more client devices may use the network-based service software or applications (e.g., by accessing the service software and / or applications via one or more application programming interfaces (APIs)). The framework layer may be, but is not limited to, a type of free and open-source software network application framework, such as one that can use a distributed file system for large-scale data processing (e.g., "big data").

[0202] A cloud-based network environment can provide cloud computing and / or cloud storage for any combination of the computing and / or data storage functions (or one or more portions thereof) described herein. Any of these various functions can be distributed across multiple locations from a central or core server (e.g., distributed across one or more data centers at the state, region, country, global, etc.). If the connection to a user (e.g., a client device) is relatively close to an edge server, the core server can assign at least a portion of the functionality to the edge server. A cloud-based network environment can be private (e.g., limited to a single organization), public (e.g., available to many organizations), and / or a combination thereof (e.g., a hybrid cloud environment).

[0203] Client devices may include those described in this article. Figure 6 The example computing device 600 described includes at least some components, features, and functions. By way of example and not limitation, the client device may be a personal computer (PC), laptop computer, mobile device, smartphone, tablet computer, smartwatch, wearable computer, personal digital assistant (PDA), MP3 player, virtual reality headset, global positioning system (GPS) or device, video player, camera, surveillance equipment or system, vehicle, ship, aircraft, virtual machine, drone, robot, handheld communication device, hospital equipment, gaming equipment or system, entertainment system, in-vehicle computer system, embedded system controller, remote control, electrical appliance, consumer electronics device, workstation, edge device, any combination of these described devices, or any other suitable device.

[0204] This disclosure can be described in the general context of machine-usable instructions or computer code, including computer-executable instructions such as program modules, which are executed by a computer or other machine such as a personal digital assistant or other handheld device. Typically, a program module, including routines, programs, objects, components, data structures, etc., refers to code that performs a specific task or implements a specific abstract data type. This disclosure can be practiced in a wide variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. This disclosure can also be practiced in distributed computing environments where tasks are performed by remote processing devices linked via a communication network.

[0205] As used herein, the phrase "and / or" relating to two or more elements should be interpreted as referring to only one element or a combination of elements. For example, "element A, element B, and / or element C" could include only element A, only element B, only element C, element A and element B, element A and element C, element B and element C, or elements A, B, and C. Furthermore, "at least one of element A or element B" could include at least one of element A, at least one of element B, or at least one of element A and at least one of element B. Further, "at least one of element A and element B" could include at least one of element A, at least one of element B, or at least one of element A and at least one of element B.

[0206] The subject matter of this disclosure is described in detail herein to satisfy legal requirements. However, the description itself is not intended to limit the scope of this disclosure. Rather, the inventors have envisioned that the claimed subject matter may also be embodied in other ways to include steps different from or similar combinations of steps described herein in conjunction with other current or future techniques. Moreover, although the terms “step” and / or “block” may be used herein to imply different elements of the method employed, these terms should not be construed as implying any particular order among or between the various steps disclosed herein, unless the order of the steps is explicitly described.

Claims

1. A system comprising: One or more LiDAR sensors; One or more processing units, for: The estimated velocity of the object is determined using the current and previous detections of the object, which are determined using the one or more LiDAR sensors; The estimated velocity is compared with one or more tracked object states corresponding to the velocities of one or more previously detected objects, each of the one or more tracked object states representing a probability distribution of the velocity of the corresponding detected object among the previously detected objects; Based at least in part on the states of the one or more tracked objects, determine one of the previously detected objects that is associated with the highest probability corresponding to the object; as well as The estimated velocity is used to update the state of the tracked object corresponding to the previously detected object with the highest probability corresponding to the object, to generate an updated tracked object state.

2. The system of claim 1, wherein the one or more processing units determine the estimated velocity of the object at least in part based on a comparison of a first object boundary and a second object boundary, the comparison comprising applying the Iterative Closest Point (ICP) algorithm.

3. The system of claim 1, wherein the estimated velocity includes at least one of magnitude or direction.

4. The system of claim 1, wherein each of the one or more tracked object states is generated using a Kalman filter.

5. The system of claim 1, further comprising transforming the states of the one or more tracked objects to a coordinate system corresponding to the current detection of the object, wherein the estimated velocity is compared with the states of the one or more tracked objects in the coordinate system.

6. The system of claim 5, wherein the conversion is at least in part based on the measured motion of the self-machine.

7. The system of claim 1 further includes using Mahalanobis distance to filter multiple tracked object states to determine the state of the one or more tracked objects.

8. The system of claim 1, wherein the system comprises at least one of the following: Control systems for autonomous or semi-autonomous machines; Sensing systems for autonomous or semi-autonomous machines; A system used to perform simulation operations; A system used to perform deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system containing one or more virtual machines (VMs); A system that is at least partially implemented in a data center; or A system that utilizes cloud computing resources at least in part.

9. A processor, comprising: Processing circuitry, used for: A first tracking state is created for a first object representing a first probability distribution, and a second tracking state is created for a second object representing a second probability distribution; The current obstacle boundary corresponding to the current detection of one of the first object or the second object is determined at least in part based on LiDAR readings; The current obstacle boundary is compared with a first obstacle boundary corresponding to the first tracking state to generate a first velocity estimate, and the current obstacle boundary is compared with a second obstacle boundary corresponding to the second tracking state to generate a second velocity estimate; A first cost function is calculated using the first probability distribution and the first velocity estimate, and a second cost function is calculated using the second probability distribution and the second velocity estimate; The current obstacle boundary is determined to be most closely related to the first obstacle boundary, based at least in part on the fact that the first cost function is less than the second cost function. as well as The first tracking state is updated using the first velocity estimate to generate an updated first tracking state corresponding to the first object.

10. The processor of claim 9, further comprising processing circuitry for converting the first tracking state and the second tracking state to a coordinate system corresponding to the currently detected state, taking into account the self-motion of the ego machine.

11. The processor according to claim 9, wherein: The first tracking state corresponds to the first Kalman filter on the first velocity vector of the first obstacle boundary, and The second tracking state corresponds to the second Kalman filter on the second velocity vector of the second obstacle boundary.

12. The processor according to claim 9, wherein: The first obstacle boundary includes a first shape metric indicating the first geometry. The current obstacle boundary includes a second shape metric indicating the second geometry, and The determination that the current obstacle boundary most closely corresponds to the first obstacle boundary is also based at least in part on the first geometry and the second geometry.

13. The processor of claim 9, wherein the updated first tracking state has lower uncertainty than the first tracking state.

14. The processor of claim 9, further comprising processing circuitry for performing a gate function using Mahalanobis distance to filter out one or more tracking states other than the first tracking state and the second tracking state.

15. The processor of claim 9, wherein the processor is included in at least one of the following: Control systems for autonomous or semi-autonomous machines; Sensing systems for autonomous or semi-autonomous machines; A system used to perform simulation operations; A system used to perform deep learning operations; Systems implemented using edge devices; Systems implemented using robots; A system containing one or more virtual machines (VMs); A system that is at least partially implemented in a data center; or A system that utilizes cloud computing resources at least in part.

16. A method comprising: The estimated velocity of the object is determined at least in part based on a comparison between the current detection of an object identified using one or more LiDAR sensors, a previous detection of the object or another object identified using the one or more LiDAR sensors, and the time elapsed between the current detection and the previous detection. The estimated velocity is compared with one or more tracked object states corresponding to one or more previously detected objects, each of the one or more tracked object states representing a probability distribution of the velocity of a corresponding detected object among the previously detected objects; Based at least in part on the comparison, determine the previously detected object with the highest probability corresponding to the previously detected object; as well as The estimated velocity is used to update the state of the tracked object corresponding to the previously detected object with the highest probability of corresponding to the object, in order to generate an updated tracked object state.

17. The method of claim 16, wherein the comparison is performed between a first object boundary and a second object boundary, and the comparison is performed using the Iterative Closest Point (ICP) algorithm.

18. The method of claim 16, wherein the estimated velocity includes at least one of magnitude or direction.

19. The method of claim 16, wherein each of the one or more tracked object states is generated using a Kalman filter.

20. The method of claim 16, further comprising transforming the one or more tracked object states to the coordinate system corresponding to the currently detected object, wherein the comparison of the estimated velocity with the one or more tracked object states is in the coordinate system.

Images