Landmark Detection Using Curve Fitting for Autonomous Driving Applications

Abstract

In various examples, one or more deep neural networks (DNNs) may be implemented to regress on the control points of the curves, and the control points may be used to perform curve-fitting operations—e.g., Bézier curve fitting—to identify landmark locations and geometries within the environment. Thus, the output of the DNNs may indicate two-dimensional (2D) image space and / or three-dimensional (3D) world space control point locations, and post-processing techniques—e.g., clustering and temporal smoothing—may be implemented to accurately and in real time determine landmark locations and poses. As such, the reconstructed curves corresponding to landmarks—e.g., lane lines, road lines, crosswalks, poles, text, etc.—may be used by a vehicle to perform one or more actions for navigating the environment.

Description

[Technical field]

[0001] The present invention relates to landmark detection using curve fitting for autonomous driving applications. [Background technology]

[0002] The ability to accurately and precisely detect landmarks within an environment - e.g., lane markings, lane outer lines, road boundaries, text, poles or signs, and / or other features - is essential for autonomous machine applications at all levels of autonomy - e.g., from driver / pilot assisted machines or vehicles to semi-autonomous and fully autonomous machines and vehicles. The variety of lane markings that may be encountered within an environment while driving is significant due to variations in road marking quality, geographic or regional differences in lane and road marking practices, and obstruction, degradation, and / or occlusion of road markings due to wear and tear, weather conditions, lighting conditions, temporary markings (e.g., due to construction or disaster recovery), and / or the like.

[0003] Some conventional approaches to lane or line detection use deep neural network (DNN) processing, where high-resolution images of the driving surface and associated annotations of lanes and lines are used to train a DNN (e.g., a convolutional neural network (CNN)) to recognize lane lines. These conventional approaches train the DNN to generate a segmentation mask that indicates the general location of lane lines in an image by classifying each pixel of the image as being part of a lane line or not. However, these conventional approaches require extensive post-processing—e.g., dynamic programming—to convert the output into a usable representation of lane lines, road boundaries, poles, and / or other features. For example, because the output of these conventional DNNs includes an identification of each pixel that belongs to each landmark feature, correlations between each pixel that belongs to the same landmark as well as their classification must be determined during post-processing. In addition, for landmarks that have other than a straight profile - e.g., to follow the curve or contour of a driving surface - additional angle or orientation information needs to be output by the DNN and post-processed to determine the angle or orientation of the landmark feature. These extensive post-processing tasks increase the latency of the system as well as the computational requirements of the system, thereby making these conventional DNNs less useful for real-time deployment in vehicles. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] U.S. Nonprovisional Patent Application No. 16 / 514,230 Summary of the Invention [Means for solving the problem]

[0005] An embodiment of the present disclosure relates to landmark detection using curve fitting for autonomous driving applications. Systems and methods are disclosed that use one or more deep neural networks (DNNs) to regress on the control points of a curve and use the control points to perform a curve fitting operation, e.g., Bézier curve fitting, to identify landmark locations and geometries within an environment. In that way, the output of the DNNs can indicate two-dimensional (2D) image space and / or three-dimensional (3D) world space control point locations, and post-processing techniques requiring less computation and time, e.g., clustering and temporal smoothing, can be performed to accurately and in real-time determine landmark locations and poses. As a result, in contrast to conventional systems, the control points can be regressed similar to object detection techniques, and the control points can then be used to reconstruct curves corresponding to the landmarks, e.g., lane lines, road lines, crosswalks, poles, text, etc., for each landmark within the field of view of each of one or more sensors of the vehicle.

[0006] To train the DNN for accurate predictions, ground truth data corresponding to the control point locations of landmark features can be generated from polyline and / or polygon annotations corresponding to the landmark features in the training data set. For example, the training data set can include previously annotated polylines or polygons, and these existing annotations - in addition to or in place of new or additional training data sets - can be analyzed using curve fitting techniques to determine control points for use in generating the ground truth data. In some embodiments, to increase the accuracy of the control point determination, points from the ground truth polyline can be resampled to produce more evenly spaced points along the polyline so that the determined control points more accurately reflect the curve of the landmark feature. The ground truth data used to train the DNN to regress on the control points may include a binary mask indicating pixels within a threshold distance to the curve, a distance regression channel indicating distance to the control points—in the x, y, and / or z directions (for 3D control point regression), and / or a classification channel to determine classification or semantic information corresponding to each control point or fitted curve. Once trained and deployed, the DNN can accurately and precisely compute outputs indicating its corresponding control point location and semantic information, and these outputs can be decoded and post-processed to determine the curve corresponding to the landmark feature.

[0007] The present systems and methods for landmark detection using curve fitting for autonomous driving applications are described in detail below with reference to the accompanying drawings. [Brief description of the drawings]

[0008] [Figure 1] 1 is a data flow diagram illustrating an example process for training a deep neural network (DNN) for landmark detection, according to some embodiments of the present disclosure. [Figure 2A]FIG. 2 illustrates a visualization of annotations associated with training images, according to some embodiments of the present disclosure. [Figure 2B] FIG. 1 illustrates a curve fitting operation according to some embodiments of the present disclosure. [Figure 2C] FIG. 13 illustrates a visualization of annotations converted to curves using a curve fitting operation, according to some embodiments of the present disclosure. [Figure 3A] FIG. 1 illustrates a visualization of training data for use in generating ground truth data, according to some embodiments of the present disclosure. [Figure 3B] FIG. 1 illustrates a visualization of a binary mask, according to some embodiments of the present disclosure. [Figure 3C] FIG. 2 illustrates a visualization of a bit classification mask, according to some embodiments of the present disclosure. [Figure 3D] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3E] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3F] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3G] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3H] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3I] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3J] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 3K] FIG. 1 illustrates a visualization of a distance regression mask, according to some embodiments of the present disclosure. [Figure 4A] FIG. 13 illustrates a visualization of ground truth data generated using reduced length and various weightings, according to some embodiments of the present disclosure. [Figure 4B]FIG. 13 illustrates a visualization of ground truth data generated using reduced length and various weightings, according to some embodiments of the present disclosure. [Figure 4C] FIG. 13 illustrates a visualization of ground truth data generated using reduced length and various weightings, according to some embodiments of the present disclosure. [Diagram 5] 1 is a flow diagram illustrating a method for training a DNN for landmark detection according to some embodiments of the present disclosure. [Figure 6] 1 is a data flow diagram illustrating an example process of using DNNs for landmark detection, according to some embodiments of the present disclosure. [Figure 7] 1 is a data flow diagram illustrating an example process of decoding and post-processing the output of a DNN for landmark detection, according to some embodiments of the present disclosure. [Figure 8A] FIG. 1 illustrates a visualization of predicted curves before and after post-processing, according to some embodiments of the present disclosure. [Figure 8B] FIG. 1 illustrates a visualization of predicted curves before and after post-processing, according to some embodiments of the present disclosure. [Figure 9A] FIG. 13 illustrates a visualization of example landmark types detected using a DNN, in accordance with some embodiments of the present disclosure. [Figure 9B] FIG. 13 illustrates a visualization of example landmark types detected using a DNN, in accordance with some embodiments of the present disclosure. [Figure 9C] FIG. 13 illustrates a visualization of example landmark types detected using a DNN, in accordance with some embodiments of the present disclosure. [Figure 9D] FIG. 13 illustrates a visualization of example landmark types detected using a DNN, in accordance with some embodiments of the present disclosure. [Figure 9E] FIG. 13 illustrates a visualization of example landmark types detected using a DNN, in accordance with some embodiments of the present disclosure. [Figure 10]1 is a flow diagram illustrating a method for landmark detection using DNNs, according to some embodiments of the present disclosure. [Figure 11A] 1 is an illustration of an example autonomous vehicle, according to some embodiments of the present disclosure. [Figure 11B] 11B is an illustration of camera positions and fields of view for the example autonomous vehicle of FIG. 11A, in accordance with some embodiments of the present disclosure. [Figure 11C] FIG. 11B is a block diagram of an example system architecture of the example autonomous vehicle of FIG. 11A in accordance with some embodiments of the present disclosure. [Figure 11D] FIG. 11B is a system diagram of communication between a cloud-based server and the example autonomous vehicle of FIG. 11A, according to some embodiments of the present disclosure. [Figure 12] FIG. 1 is a block diagram of an example computing device suitable for use in implementing some embodiments of the present disclosure. [Figure 13] FIG. 1 is a block diagram of an exemplary data center suitable for use in implementing some embodiments of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Systems and methods relating to landmark detection using curve fitting for autonomous driving applications are disclosed. The disclosure may be described with respect to an exemplary autonomous vehicle 1100 (an example of which is described herein with respect to FIGS. 11A-11D and alternatively referred to herein as “vehicle 1100” or “autonomous vehicle 1100”), but this is not intended to be limiting. For example, the systems and methods described herein may be used by a non-autonomous vehicle, a semi-autonomous vehicle (e.g., in an adaptive driver assistance system (ADAS)), a robot, a warehouse vehicle, an off-road vehicle, an airship, a boat, a passenger vehicle, a car, a truck, a bus, a first responder vehicle, a shuttle, an electric or moped, a motorcycle, a fire engine, a police vehicle, an ambulance, a construction vehicle, a submarine, a drone, another type of vehicle (e.g., unmanned and / or carrying one or more passengers), a piloted or non-piloted robot, and / or a piloted or non-piloted robotic platform. Additionally, although the present disclosure may be described with respect to autonomous driving, this is not intended to be limiting. For example, the systems and methods described herein may be used in other technical fields, such as robotics, aeronautical systems, marine systems, and / or perception, world model management, path planning, obstacle avoidance, and / or other processes.

[0010] Referring to FIG. 1, FIG. 1 is a data flow diagram illustrating an example process for training a deep neural network (DNN) for landmark detection, according to some embodiments of the present disclosure. It should be understood that this and other configurations described herein are described merely as examples. Other configurations and elements (e.g., machines, interfaces, functions, sequences, groupings of functions, etc.) may be used in addition to or instead of those illustrated, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities that may be implemented as separate or distributed components or in conjunction with other components, and in any suitable combination and location. Various functions described herein as being performed by an entity may be implemented by hardware, firmware, and / or software. For example, various functions may be implemented by a processor executing instructions stored in a memory. In some embodiments, the training of the deep neural network (DNN) 104 by the process 100 may be implemented, at least in part, using similar components, features, and / or functionality as described herein with respect to the vehicle 1100 of FIGS. 11A-11D, the example computing device 1200 of FIG. 12, and / or the example data center 1300 of FIG. 13.

[0011] The process 100 may include generating and / or receiving sensor data 102 from one or more sensors. The sensor data 102 may be received from one or more sensors of a vehicle (e.g., vehicle 1100 of FIGS. 11A-11D described herein), as one non-limiting example. The sensor data 102 may be used by the vehicle 1100 in the process 100 to train one or more DNNs 104 to detect landmarks, e.g., landmark locations, poses, semantic classes, etc. During training, the sensor data 102 may be generated using one or more data collection vehicles that generate sensor data for training the DNN, e.g., the DNN 104, and / or may be pre-generated and included in a training data set. The sensor data 102 used during training may additionally or alternatively be generated using simulated sensor data (e.g., sensor data generated using one or more virtual sensors of a virtual vehicle in a virtual environment) and / or augmented sensor data (e.g., sensor data generated using one or more data collection vehicles and modified with virtual data, respectively). After training and deployment in the vehicle 1100, the sensor data 102 may be generated by one or more sensors of the vehicle 1100 and processed by the DNN 104 to compute various outputs 106, as described herein.

[0012] 11A-11C , the sensor data 102 may include, but is not limited to, sensor data 102 from any of the sensors of the vehicle 1100, including, for example, a RADAR sensor 1160, an ultrasonic sensor 1162, a LIDAR sensor 1164, a stereo camera 1168, a wide view camera 1170 (e.g., a fisheye camera), an infrared camera 1172, a surround camera 1174 (e.g., a 360-degree camera), a long-range and / or mid-range camera 1178, and / or other sensor types. As another example, the sensor data 102 may include virtual (e.g., simulated or augmented) sensor data generated from any number of sensors of a virtual vehicle or other virtual object in a virtual (e.g., test) environment. In one such instance, the virtual sensors may correspond to virtual vehicles or other virtual objects in a simulated environment (e.g., used to test, train, and / or validate DNN performance) and the virtual sensor data may represent sensor data captured by the virtual sensors in the simulated or virtual environment. As such, by using virtual sensor data, the DNNs 104 described herein may be tested, trained, and / or validated using simulated or augmented data in a simulated environment, which may enable testing of more extreme scenarios outside of real-world environments where such testing may be less safe.

[0013] In some examples, the sensor data 102 may include image data representing an image, image data representing a video (e.g., a snapshot of a video), and / or sensor data representing a representation of a sensor's field of perception (e.g., a depth map for a LIDAR sensor, a value graph for an ultrasonic sensor, etc.). When the sensor data 102 includes image data, any type of image data format may be used, including, but not limited to, compressed images, such as Joint Photographic Experts Group (JPEG) or Luminance / Chrominance (YUV) formats, compressed images, such as frames resulting from compressed video formats, such as H.264 / Advanced Video Coding (AVC) or H.265 / High Efficiency Video Coding (HEVC), raw images, such as those derived from Red Clear Blue (RCCB), Red Clear (RCCC), or other types of image sensors, and / or other formats. Additionally, in some instances, the sensor data 102 may be used in the process 100 without preprocessing (e.g., in a raw or captured format), while in other instances, the sensor data 102 may undergo preprocessing (e.g., noise balancing, demosaicing, scaling, cropping, enhancement, white balancing, tone curve adjustment, etc., e.g., using a sensor data preprocessor (not shown)). As used herein, sensor data 102 may refer to raw sensor data, preprocessed sensor data, or a combination thereof.

[0014] The sensor data 102 used for training may include original images (e.g., as captured by one or more image sensors), downsampled images, upsampled images, cropped or region of interest (ROI) images, otherwise augmented images, and / or combinations thereof. The DNN 104 may be trained using the images (and / or other sensor data 102) and corresponding ground truth data 122, such as a ground truth (GT) binary mask 124, a GT distance regression mask 126, and / or a GT bit classification mask 128. The ground truth data 122 may be generated by the annotation data 114, such as annotations, labels, masks, and / or the like, a curve fitter 116, and / or an encoder 120 using control points 118. For example, in some implementations, annotation data 114 may represent polylines corresponding to any landmark type, e.g., road markings, lane markings, road boundaries, intersection lines, pedestrian paths, bike lane markings, text, poles, trees, light posts, signs, etc. For example, with reference to FIG. 2A, visualization 200A may include polylines 202A-202F, which may correspond to lane markings and / or road boundaries. Polylines 202A-202F may include corresponding semantic or classification information. For example, when annotating polyline 202, polyline 202 may be associated with a particular semantic class. For example, polyline 202A may correspond to a road boundary, or, at a finer granularity, a solid white road boundary. Similarly, polyline 202C may correspond to a dashed lane boundary, and polyline 202D may correspond to a solid road boundary.

[0015] In some embodiments, a polyline may be determined or extracted from a polygon. For example, the annotations or labels represented by annotation data 114 may correspond to a polygon that substantially defines the shape of a landmark—e.g., a polygon fits each lane, a polygon fits a pole, a polygon fits text, etc. The center line, left border, and / or right border of the polygon may be used to generate the polyline. For example, for a lane, the polygon may be used to determine the left lane boundary line (e.g., the left edge of the polygon), the right lane boundary line (e.g., the right edge of the polygon), and / or the center line or rail of the lane (e.g., a line that runs through the center of the polygon). Thus, if a polygon is used, a polyline—e.g., polyline 202 of FIG. 2A—may be generated or extracted from the polygon. In any embodiment, the annotation may include a polygon, a polyline, another annotation or label type, or a combination thereof.

[0016] As used herein, annotation data 114 may represent semantic or classification information corresponding to polygons, polylines, and / or other annotation or label types. Semantic classes may include any level of granularity. For example, semantic classes may include lane lines, road lines, intersection lines, poles, signs, text, etc. At finer levels of granularity, lane line semantic classes may include, for example, solid or dashed lines, yellow or white lines, single or double lines, or combinations thereof. Thus, by way of non-limiting example, lane line semantic classes may include white single solid line, yellow single solid line, white single dashed line, yellow single dashed line, white double solid line, yellow double solid line, white double dashed line, yellow double dashed line, white solid dashed line, yellow solid dashed line, white dashed solid line, yellow dashed solid line, triple solid line, triple dashed line, solid dashed solid line, dashed dashed solid line, dashed dashed solid line, etc. For road boundaries, the semantic classes may include, by way of non-limiting example, obstacles, curves, seams, lines, etc. For intersection lines, the semantic classes may include, by way of non-limiting example, intersection lines, crossing lines, stop lines, pedestrian lines, etc. For poles, the semantic classes may include, by way of non-limiting example, road signs, poles, lights, utility poles, trees, etc., or for signs, traffic signs, construction signs, advertising signs, etc. Thus, any number and granularity of semantic classes may be associated with annotations or labels and used to generate ground truth data 122 for training the DNN 104.

[0017] The annotations, in some instances, may be generated within a drawing program (e.g., an annotation program), a computer-aided design (CAD) program, a labeling program, another type of program suitable for generating annotations, and / or may be hand-drawn. In any instance, the annotation data 114 may be synthetically produced (e.g., generated from a computer model or rendering), physically produced (e.g., designed and manufactured from real-world data), machine automated (e.g., using feature analysis and learning to extract features from data and then generate labels), human-annotated (e.g., a labeler, or annotation expert, defines the location of the labels), and / or a combination thereof (e.g., a human identifies the center or origin and dimensions of an area, and a machine generates polygons and / or labels for the intersection area).

[0018] The process 100 may include performing a curve fitting operation using the curve fitter 116. For example, a polyline represented by the annotation data 114 may be subjected to a curve fitting operation to determine control points that define the points of the curve. Thus, the ground truth data 122 may be used to train the DNN 104 to predict, regress, or otherwise calculate the locations of the control points—in 2D image space or 3D world space. The curve fitting operation may include a 2D Bézier curve fitting algorithm (e.g., for image space control point estimation), a 3D Bézier curve fitting algorithm (e.g., for world space control point estimation), a cubic Bézier curve fitting algorithm (e.g., including four control points for each curve), a higher order Bézier curve fitting algorithm (e.g., five or more control points), or a split Bézier curve fitting algorithm (e.g., a single curve is split into two or more curves, each including multiple control points), and / or another curve fitting algorithm type. In some embodiments, the curve fitting algorithm may be used to determine more than two control points to increase the accuracy and precision of the curve fit. For example, with only two points, the result of the curve fit is similar to that of linear interpolation, which would produce a less accurate or precise result for a curve that is not linear. Thus, the more control points used, the more accurate and precise the curve fit may be, but the more control points, the more processing and calculations the operation requires, thereby increasing the run time of the system. Thus, through experimentation, and by way of non-limiting example, using three to five control points may provide accurate and precise results with acceptable computational requirements that allow for real-time deployment. However, any number of control points may be used without departing from the scope of the present disclosure. In addition, while the figures herein include four control points per curve, this is not intended to be limiting and is for illustrative purposes only. Furthermore, while Bézier curve fitting is primarily described herein, this is not intended to be limiting and other curve fitting algorithms may be used without departing from the scope of the present disclosure.However, Bézier curve fitting includes several advantages including the use of three or more control points to generate a more accurate curve fitting result, two of the control points being located at the beginning and end of each curve that serves the vehicle 1100 in performing planning, control, or other decision types.

[0019] In some embodiments, prior to the curve fitting operation, points along the polyline from the annotation data may be resampled. For example, the polyline may contain fewer points if the curve is straighter and more points if there is a greater curvature, so there may be more points at some locations than others. A polyline without resampling can produce accurate results, but the accuracy can be enhanced by resampling the points along the polyline to include a more evenly sampled distribution of points. As an example, the polyline 202F in FIG. 2A may include point 206 prior to resampling, while the polyline 202C may include point 206 after resampling. As shown, point 206 along polyline 202C is more evenly distributed than points along polyline 202F. Resampling can reduce the likelihood that the curve fitting operation will generate control points that favor regions of the polyline with more points. As a result, the control points determined using the curve fitter 116, and thus the ground truth data 122 generated using the control points, can be more precise for curve fitting.

[0020] The polyline - in some embodiments, after resampling - may then undergo a curve fitting operation to determine control point locations. Referring to FIG. 2B, as an example, when Bézier curve fitting is used, the Bézier curve may be defined as Q(t), where t may include parameters such that 0 < t < 1. The first matrix, T, in the matrix representation of Equation (1) below is t 0 t 1 t 2 and t 3In equation (1), the second matrix, M, can be called the Bézier matrix, and is made up of coefficients when expanding a cubic polynomial. The four control points (in this example, a cubic Bézier curve fitting algorithm) are P 0 , P 1 , P 2 , and P 3 Since there are two dimensions in this example, the control points can be expressed as a third matrix, P, in equation (1), where x 0 and y 0 is the control point P 0 (e.g., pixel coordinates in 2D image space), etc. Although expressed as 2D pixel or image space coordinates, this is not intended to be limiting, and in some embodiments, the coordinates may correspond to 3D world space coordinates, such that the DNN 104 is trained to directly calculate the 3D coordinates of the control points of a curve.

[0021]

number

[0022] A Bezier curve fitting algorithm may alternatively optimize P and T. An initial T may be calculated by the length of each line segment (or polyline). If there is only one line segment, the line segment may be divided evenly into three line segments. If there are only two line segments, the longer one may be divided evenly into two line segments. In an embodiment, any line segment may be divided into pieces such that each line segment has the same relative length. An initial T may be calculated after the line segments are determined, and then the resampled polyline may be used for optimization. When T is fixed, P may be solved, for example, by the least squares method according to the following equation (2): P=(M T T T TM) -1 M T T T Q (2)

[0023] When P is fixed, T can be solved by the Newton-Raphson method. For example, the problem to be solved can be the root of equation (3) below: f(t)=(Q(t)-Q)Q'(t) (3)

[0024] The concept is illustrated, as an example, in FIG. 2B, where the foot of f(t), f′(t), is calculated according to equation (4) below: f'(t)=Q'(t)Q'(t)+(Q(t)-Q)Q''(t) (4) The update step is then represented by the following equation (5): t=tf(t) / f'(t) (5)

[0025] At each iteration, an error, e, may be calculated according to equation (6) below: e=||P-TMQ|| (6)

[0026] Convergence may be determined if the error, e, stops decreasing. Divergence may be determined if the error, e, exceeds a maximum error. If the maximum number of steps or iterations is reached, the results of the last iteration may be used.

[0027] As a result of the curve fitting operation by the curve fitter 116, one or more curves may be generated, each defined by control points 118. As an example of the result of the curve fitting operation, FIG. 2C includes a visualization 200B of curves 204 (e.g., curves 204A-204F) generated from the polyline 202 of FIG. 2A after curve fitting. As such, each curve (of a particular semantic label) includes a comparable number of control points 118—e.g., four in this example, if a cubic Bézier curve fit was performed. As shown, even though they are adjacent in world space, curve 204C corresponds to a lane boundary line and curve 204E corresponds to a road boundary line, so that two curves 204 may be generated, each with their own control points 118. As a result, as described herein, the DNN 104 can be trained to compute a bit classification mask 112 corresponding to a particular semantic class, and / or to regress (or otherwise compute) a binary mask 108 and a distance regression mask 110 corresponding to control points 118 of the curve 204 of a particular landmark type (e.g., a particular road marking type with respect to FIG. 2C ).

[0028] The control points 118 and / or semantic classification information may then be used by the encoder 120 to generate ground truth data 122 for training the DNN 104 using the training engine 130. Thus, for each instance of sensor data 102 in the training data set, ground truth data 122 may be generated. The ground truth data 122 may include one or more GT binary masks 124, one or more GT distance regression masks 126, and / or one or more GT bit classification masks 128. In some embodiments, for example, if variable lengths and / or widths are used to encode curves in the ground truth data 122—as described in more detail herein—weight channels may also be generated for the ground truth data 122. Ultimately, the ground truth data 122 may be compared to the output 106 of the DNN 104 by a training engine 130 using one or more loss functions, such as, but not limited to, those described herein. For example, a binary mask 108 computed by the DNN 104 for an instance of sensor data 102 may be compared to a GT binary mask 124 corresponding to the same instance of sensor data 102, a distance regression mask 110 computed by the DNN 104 for an instance of sensor data 102 may be compared to a GT distance regression mask 126 corresponding to the same instance of sensor data 102, etc. The training engine 130 may use the loss function to determine updates to the parameters (e.g., weights, biases, etc.) of the DNN 104, and this process may be repeated until the DNN 104 converges to an acceptable level of accuracy, such as until the trained parameters of the DNN 104 converge to optimal, desired, and / or acceptable values.

[0029] As an example, referring to FIG. 3A, a visualization 300 may correspond to an instance of sensor data 102 (e.g., an image) and corresponding curves 302A-302D (e.g., each curve 302 may include four control points 118), each including control points 118. Each curve 302 may have been generated by the curve fitter 116 from a polyline represented by the annotation data 114. The curves 302 may each be associated with semantic classification information. For example, the curves 302A and 302D may correspond to road boundaries, and the curves 302B and 302C may correspond to lane boundaries (e.g., lane outside lines). When shown with respect to only curve 302D (for purposes of clarity), each curve 302 may include a first control point 118A, a second control point 118B, a third control point 118C, and a fourth control point 118D. The first control point 118A may correspond to the last control point, and the fourth control point 118D may correspond to the first control point 118 of each curve.

[0030] The encoder 120 can encode the GT binary mask 124 using the curve 302. For example, with reference to FIG. 3B, a GT binary mask 124A corresponding to the curve 302 from the visualization 300 is shown. The GT binary mask 124A can be generated by encoding a first value (e.g., 1) to each pixel within a defined radius of any curve 302, while encoding a second value (e.g., 0) to each pixel outside the defined radius of any curve 302. The radius can be selected to be any value. For example, in one non-limiting example, if the input and output resolutions of the DNN 104 are different (e.g., due to down-sampling via a convolution operation), the radius can include 16 to 48 pixels (e.g., if the resolution difference is 8x) at the input resolution of the DNN 104 (e.g., the resolution of the image from the visualization 300) and 2 to 6 pixels at the output resolution of the DNN 104. However, a different radius may be set depending on the implementation, and / or the difference in input to output resolution may be set depending on the implementation. For example, pixels encoded as being within the radius may be used during decoding as candidate pixels for regression (or other computation) at control point 118, so the radius may be determined such that enough candidate pixels are selected for accurate and precise prediction, without including a large number of pixels that would increase processing requirements and run time.

[0031] The encoder 120 can encode the GT bit classification mask 128 using the control points 118 of the curve 302. For example, a number of GT bit classification masks 128 or channels can be determined based on the number of different semantic classes that the DNN 104 is trained to predict. For example, if there are n semantic classes, there can be an upper bound (maximum) log2(n) channel. Thus, if there is only a single semantic class, there may be no bit classification channel and thus no GT bit classification mask 128. However, if there are two bit classification channels, there may be a single bit classification channel and thus a single GT bit classification mask 128. In one such example, a first value (e.g., 0) may indicate that the pixel corresponds to a first semantic class, and a second value (e.g., 1) may indicate that the pixel corresponds to a second semantic class. As an example, with respect to FIG. 3C, if there are two semantic classes (e.g., assuming that the DNN 104 is simply trained to predict for the two classes in one such example), the GT bit classification mask 128A may correspond to the visualization 300. For example, the curves 302C and 302B corresponding to the lane boundary lines may be encoded with a first value (e.g., 1, corresponding to white) and the curves 302A and 302D corresponding to the road boundary lines may be encoded with a second value (e.g., 0, corresponding to black). In some embodiments, each of the pixels of the GT distance regression mask 126 may be encoded (or considered for encoding) with a bit classification value, while in other examples, only pixels identified as being within a radius of the curve 302 from the GT binary mask 124 may be encoded with a bit classification value. If a radius is used for each mask type, the DNN 104 may not train on background pixels, but may be trained to make predictions on pixels that are most likely to contribute to the accuracy of the prediction.

[0032] The encoder 120 can encode the GT distance regression mask 126 using the control points 118 of the curves 302. For example, for each control point 118, an x-axis GT distance regression mask 126 and a y-axis GT distance regression mask 126 can be encoded. In an embodiment where the control points 118 are in 3D world space coordinates, each control point 118 can include an additional z-axis GT distance regression mask 126. For example, each curve 302 can include a first control point 118 (e.g., the last control point), and the pixels of the x-axis GT distance regression mask 126 can be encoded with the distance in the x-direction (e.g., pixel distance) to the nearest first control point 118A of any curve 302, and the pixels of the y-axis GT distance regression mask 126 can be encoded with the distance in the y-direction (e.g., pixel distance) to the nearest first control point 118A of any curve 302. This process may be repeated for each of the control points 118B-118D (if there are four control points 118, however, more or fewer iterations may be included if there are more or fewer control points 118). The distance values ​​may have a scaling factor applied thereto to generate updated distance values ​​that are within a smaller range of values ​​(e.g., closer to 0-1 than 0-90, or some other value). In some implementations, each of the pixels in the GT distance regression mask 126 may be encoded, while in other instances, only pixels identified from the GT binary mask 124 as being within the radius of the curve 302 may be encoded with a distance value. In instances where only pixels within the radius are used, accuracy may not be compromised, while processing requirements and run time may be reduced.

[0033] As an example, with respect to FIGS. 3D-3K, the control points 118A of each curve 302A-302D may be encoded into an x-axis GT distance regression mask 126A-X (FIG. 3D) and a y-axis GT distance regression mask 126A-Y (FIG. 3E), and the control points 118B of each curve 302A-302D may be encoded into an x-axis GT distance regression mask 126B-X (FIG. 3F) and a y-axis GT distance regression mask 126B-Y (FIG. 3G). The control points 118C of each curve 302A-302D may be encoded in an x-axis GT distance regression mask 126C-X (FIG. 3H) and a y-axis GT distance regression mask 126C-Y (FIG. 3I), and the control points 118D of each curve 302A-302D may be encoded in an x-axis GT distance regression mask 126D-X (FIG. 3J) and a y-axis GT distance regression mask 126D-Y (FIG. 3K).

[0034] In some embodiments, the distance may correspond to an output resolution of the DNN 104, while in other embodiments, the distance may correspond to an input resolution of the DNN 104. Thus, although the x-axis or y-axis distance may be calculated for pixels at the output resolution, the values ​​may correspond to distances (e.g., pixel distances) at the input resolution (e.g., of an input image to the DNN 104). Thus, during decoding, a pixel at the output resolution may have a known or estimated corresponding pixel at the input resolution, and the known or estimated corresponding pixel may then have a value from the output resolution attributed thereto. This process may be similar to FIGS. 2A-2B and the accompanying description in U.S. Non-Provisional Patent Application No. 16 / 514,230, filed July 17, 2019, which is incorporated herein by reference in its entirety.

[0035] In some embodiments, for example, when two or more of the GT binary mask 124, the GT distance regression mask 126, or the GT bit classification mask 128 are generated such that only pixels within a radius of the curve 302 are encoded, the radius may be different for the different mask types. For example, the GT binary mask 124 may have a first radius and the GT distance regression mask 126 may have a second radius that is different from the first radius. Similarly, the GT bit classification mask 128 may have a different radius than the GT binary mask 124 and / or the GT distance regression mask 126. In such embodiments, separate GT weight channels or masks may be generated to train the DNN 104 to output weight channels to aid in decoding the masks, even though their radii (e.g., pixel radii from the curve) may be different. For example, with reference to Figures 4A-4C, the GT binary mask 124B (Figure 4A) may include a first radius, the GT distance regression mask 126E (Figure 4B) may include a second, larger radius, and the weight channel or mask 400 (Figure 4C) may be used to encode this information so that the calculated weight channels of the DNN 104 being deployed may be used by the decoder 120 (Figure 6) to more accurately regress (or otherwise calculate) the positions of the control points 118.

[0036] In some embodiments, to make the predictions of the DNN 104 more focused, the ground truth data 122 may be encoded with shorter curves. For example, each of the GT binary mask 124, the GT distance regression mask 126, and / or the GT bit classification mask 128 may be encoded with shorter distances—as shown by the difference between the masks in FIGS. 3B-3K and 4A-4C. For example, in some embodiments, the encoded values ​​of the curves in each mask may be cropped at the top and / or bottom by some percentage or amount. In one non-limiting embodiment, values ​​corresponding to the top 25% of the curve and values ​​corresponding to the bottom or lower 25% of the curve may be cropped such that if the entire curve were encoded, the masks resulting from the ground truth data 122 would be 50% of their potential or initial length. These values ​​are merely for illustrative purposes, and one or more key performance indicators (KPIs) may be used to determine appropriate or ideal values ​​for any particular embodiment. This process can be useful because it accounts for the receptive field of the DNN 104. For example, if a complete curve or polyline is used, then the pixels or points (e.g., in 3D world space) at each end of the curve or polyline can have the accuracy of the regression channel output.

[0037] As such, after the ground truth data 122 is encoded, the DNN 104 may undergo training using the training engine 130. The sensor data 102 may be applied to the DNN 104, which may compute outputs 106. The outputs 106 may then be compared to the corresponding ground truth data 122 using one or more loss functions, and the outputs of the loss functions may be used by the training engine 130 to update the parameters of the DNN 104. This process may be repeated until an acceptable convergence is achieved.

[0038] In some implementations, the loss function may vary depending on the ground truth data 122 type. For example, a first loss function may be used to train the DNN 104 on the GT binary mask 124 and the binary mask 108, a second loss function may be used to train the DNN 104 on the GT distance regression mask 126 and the distance regression mask 110, and a third loss function may be used to train the DNN 104 on the GT bit classification mask 128 and the bit classification mask 112. As one non-limiting example, the binary mask channel may be trained using binary cross entropy loss, the distance regression channel may be trained using masked LI loss, and the bit classification channel may be trained using masked binary cross entropy loss.

[0039] Referring now to FIG. 5, each block of the method 500 described herein includes computational processes that may be performed using any combination of hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in a memory. The method 500 may also be implemented as computer usable instructions stored on a computer storage medium. The method 500 may be provided by a standalone application, a service, or a hosted service (standalone or in combination with another hosted service), or a plug-in to another product, to name a few. Additionally, the method 500 is described with respect to the process 100 of FIG. 1, by way of example. However, the method 500 may additionally or alternatively be performed by any one process and within any one system, or any combination of processes and systems, including but not limited to those described herein.

[0040] 5 is a flow diagram illustrating a method 500 for training a DNN 104 for landmark detection, according to some embodiments of the present disclosure. The method 500 includes, at block 502, receiving annotation data representing polylines associated with an image and classification information associated with the polylines. For example, annotation data 114 associated with an instance of sensor data 102 may be received and / or generated, and the annotation data may represent one or more polylines corresponding to landmarks and / or their associated classification information.

[0041] The method 500 includes, at block B504, performing a curve fitting algorithm to determine a plurality of control points associated with the curve based at least in part on the polyline. For example, the curve fitter 116 may use the annotation data 114 to perform a curve fitting operation to determine the control points corresponding to the curve.

[0042] The method 500 includes encoding a binary mask indicative of pixels of the image that are within a radius to the curve and encoding a plurality of distance regression masks based at least in part on the plurality of control points associated with the curve at block B506. For example, using the control points 118 and / or the curve information, the encoder 120 may encode a GT binary mask 124 and / or a GT distance regression mask 126, where each distance regression mask may indicate a distance to a control point 118 of the plurality of control points 118.

[0043] The method 500 includes encoding one or more bit classification masks based at least in part on the classification information at block B 508. For example, the encoder 120 may encode the GT bit classification mask 128 using the classification information and / or the control point or curve information.

[0044] The method 500 includes training a neural network using the binary mask, the distance regression masks, and the one or more bit classification masks as ground truth data at block B 510. For example, the training engine 130 may use the ground truth data 122 to train the DNN 104 to compute the output 106.

[0045] DNN for Landmark Detection Reference is now made to FIG. 6, which is a data flow diagram illustrating an example process 600 of using DNNs for landmark detection, according to some embodiments of the present disclosure. It should be understood that this and other arrangements described herein are described merely as examples. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, groupings of functions, etc.) may be used in addition to or instead of those illustrated, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional entities that may be implemented as separate or distributed components or in conjunction with other components, and in any suitable combination and location. Various functions described herein as being performed by an entity may be implemented by hardware, firmware, and / or software. For example, various functions may be implemented by a processor executing instructions stored in a memory. In some embodiments, deployment of the DNN 104 according to process 600 may be implemented, at least in part, using similar components, features, and / or functionality as described herein with respect to the vehicle 1100 of FIGS. 11A-11D, the exemplary computing device 1200 of FIG. 12, and / or the exemplary data center 1300 of FIG. 13.

[0046] The process 600 may include generating and / or receiving sensor data 102 generated by one or more sensors of the vehicle 1100. The sensor data 102 may be similar to that described herein, for example, with respect to the process 100 of FIG. 1. The sensor data 102 may be applied to a DNN 104, which may be trained according to the process 100 of FIG. 1, and the DNN 104 may calculate an output 106. In an embodiment, the DNN 104 may include a convolutional neural network (CNN). Although examples are described herein with respect to the use of neural networks, and specifically convolutional neural networks as the DNN 104, this is not intended to be limiting. For example, and without limitation, the DNN 104 described herein may include any type of machine learning model, such as machine learning models using linear regression, logistic regression, decision trees, support vector machines (SVMs), naive Bayes, k-nearest neighbors (Knn), K-means clustering, random forests, dimensionality reduction algorithms, gradient boosting algorithms, neural networks (e.g., autoencoders, convolutional, recurrent, perceptrons, long / short term memory (LSTM), Hopfield, Boltzmann, deep belief, deconvolution, generative adversarial, liquid state machines, etc.), and / or other types of machine learning models.

[0047] As an example, such as when the DNN 104 includes a CNN, the DNN 104 may include any number of layers. One or more of the layers may include an input layer. The input layer may hold values ​​associated with the sensor data 102 (e.g., before or after post-processing). For example, when the sensor data 102 is an image, the input layer may hold values ​​representing the raw pixel values ​​of the image as a volume (e.g., width, height, and color channels (e.g., RGB), e.g., 32x32x3).

[0048] One or more layers may include a convolutional layer. A convolutional layer may compute the output of neurons connected to a local region in the input layer, where each neuron computes the dot product between their weights and the small region to which they are connected in the input capacitance. The result of a convolutional layer may be another capacitance, with one of the dimensions based on the number of filters applied (e.g., width, height, and number of filters, e.g., 32x32x12, where 12 is the number of filters).

[0049] One or more layers may include a deconvolution layer (or a transposed convolution layer), for example, the result of the deconvolution layer may be another capacity having a higher dimensionality than the input dimensionality of the data received at the deconvolution layer.

[0050] One or more of the layers may include a rectified linear unit (ReLU) layer. The ReLU layer may, for example, apply an element-wise activation function, e.g., max(0,x), that is thresholded at zero. The resulting capacity of the ReLU layer may be the same as the capacity of the input of the ReLU layer.

[0051] One or more of the layers may include a pooling layer, which may perform a down-sampling operation along spatial dimensions (e.g., height and width) that may result in a capacity smaller than the pooling layer's input (e.g., 16x16x12 from a 32x32x12 input capacity).

[0052] One or more of the layers may include one or more fully connected layers. Each neuron in a fully connected layer may be connected to a respective neuron in the previous capacity. The fully connected layers may compute class scores, and the resulting capacity may be 1x1x the number of classes. In some instances, the CNN may include a fully connected layer, such that the output of one or more of the layers of the CNN may be provided as input to a fully connected layer of the CNN. In some instances, one or more convolutional streams may be implemented by the DNN 104, and some or all of the convolutional streams may include respective fully connected layers.

[0053] In some non-limiting examples, the DNN 104 may include a series of convolution and max-pooling layers to facilitate image feature extraction, followed by multi-scale dilated convolution and up-sampling layers to facilitate global context feature extraction.

[0054] Although input layers, convolutional layers, pooling layers, ReLU layers, and fully connected layers are discussed herein with respect to the DNN 104, this is not intended to be limiting. For example, additional or alternative layers may be used in the DNN 104, such as normalization layers, SoftMax layers, and / or other layer types.

[0055] In embodiments where the DNN 104 includes a CNN, different orders and numbers of layers of the CNN may be used depending on the embodiment, in other words, the order and number of layers of the DNN 104 is not limited to any one architecture.

[0056] In addition, some of the layers may include parameters (e.g., weights and / or biases), e.g., convolutional layers and fully connected layers, while other layers, e.g., ReLU layers and pooling layers, may not. In some instances, the parameters may be learned by the DNN 104 during training. Furthermore, some of the layers may include additional hyperparameters (e.g., learning rate, stride, epoch, etc.), e.g., convolutional layers, fully connected layers, and pooling layers, while other layers, e.g., ReLU layers, may not. The parameters and hyperparameters are not limited and may vary depending on the embodiment.

[0057] The output 106 may be decoded using the decoder 602 to determine the control points 118. For example, each pixel classified as positive (or within the radius of the curve) from the binary mask 108 may be considered a candidate pixel (e.g., for 2D prediction) or point (e.g., for 3D prediction). The corresponding values ​​of the same pixel in the distance regression mask 110 and the bit classification mask 112 may be used as properties of the curve candidate (e.g., Bézier curve candidate). For example, for pixels of the binary mask that are positive pixels (e.g., white pixels, or pixels identified as within the radius of the curve), the x-distance value from one or more (e.g., all) of the x-axis distance regression masks 110 and the y-distance value from one or more (e.g., all) of the y-axis distance regression masks 110 may be used to determine the candidate control point location. This process may be repeated for each of the candidate pixels. Ultimately, as shown in FIG. 8A, there may be multiple overlapping or similar control points and curves for each landmark shown in the image. For example, for a single landmark—e.g., a lane marking—there may be multiple sets of control points 118. The curve fitter 116 may perform a curve fitting operation using each set of control points 118 to generate a curve (e.g., a Bézier curve).

[0058] Due to the potential for multiple curves being generated for the same landmark, various post-processing operations may be performed using the post-processor 604 to determine a final curve for use by the vehicle 1100, for example, by the autonomous driving software stack (or drive stack) 608. For example, clustering may be performed by the post-processor 604 using any clustering algorithm type. In one non-limiting example, a density-based spatial clustering of applications with noise (DBSCAN) algorithm may be performed on the curve candidates, while in another non-limiting example, a probabilistic voting algorithm may be used. In any instance, a similarity function may be used within the clustering algorithm. In some instances, the similarity function may include an integral loss similarity function. For example, the following equation (7) may represent one instance of an integral loss similarity function:

[0059]

number

[0060] In equation (7), P 1 and P 2 are the points on the two (Bézier) curves being compared, t is the parameter of the curve going from 0 to 1, and a 1 and a 2 are the cubic coefficients of the first and second curves, respectively, and b 1 and b 2 are the quadratic coefficients of the first and second curves, respectively, and c 1 and c 2 are the linear coefficients of the first and second curves, respectively, and d 1 and d 2 are the constant term coefficients of the first and second curves, respectively. Therefore, P 1 (t)=a 1 * t 3 +b1 * t 2 +c 1 * t+d 1 (0<=t<=1) and P 2 (t)=a 2 * t 3 +b 2* t 2 +c 2 * t+d2 (0<=t<=1) may correspond to the formula of points on two Bezier curves, respectively. After clustering, there should be a single curve for each landmark feature. However, in some instances, there may still be more than one curve for one or more of the landmark features. In such instances, another post-processing technique of duplicate removal may be performed by the post-processor 604. For example, the curves may be compared to other curves that are close to each other (e.g., within some threshold distance of each other), and those curves with lower confidence may be removed. As such, based on the predictions and / or clustering of the DNN 104, each curve may have a confidence associated with it, and the curve with the lowest confidence may be removed.

[0061] As an example, with reference to FIG. 8A, the decoded output 106 of the DNN 104 may be represented in visualization 800A. As shown, multiple candidate curves 802 may be generated for each landmark type, e.g., lane boundaries or road boundaries in the illustration of FIG. 8A. Accordingly, post-processing may be performed to generate final curves 804, including one final curve 804 for each landmark feature. For example, candidate curve 802A may undergo post-processing to generate final curve 804A, candidate curve 802B may undergo post-processing to generate final curve 804B, etc. These final curves 804 may be associated with landmarks 606 such that the position, orientation, pose, and / or other information regarding the landmarks may be known.

[0062] Referring to FIG. 7, FIG. 7 is a data flow diagram illustrating an example process 700 of decoding and post-processing the output of the DNN 104 for landmark detection, according to some embodiments of the present disclosure. For example, the process 700 may include an end-to-end (E2E) process for landmark detection. Various classes of landmarks 606 may be used together to determine final information, such as lane detection, intersection detection, road marking detection, etc. As such, to determine solid / dashed lines (or lane boundary lines), the output 106 may be decoded to generate candidate curves, the candidate curves may be clustered (e.g., to populate the lane marking output) to generate a final curve, and if the final curve is close to another final curve, duplicate removal may be performed to result in solid / dashed line (or lane boundary line) detection. A similar process may be performed for road boundaries, and a combination of lane boundaries and road boundary lines may be used to identify each lane of the driving surface, such as to generate a lane graph. As another example, a similar process may be performed for intersections or cross-junctions (e.g., crosswalks, pedestrian walkways, etc.), poles or signs, and / or text or other shapes identified using DNN 104. These outputs, e.g., intersection detection using intersection lines, intersection detection using poles or road signs, and road marking detection, may be used to determine intersection location, structure, and pose, and / or road marking information. As such, even if separate classes of information are computed by DNN 104, one or more classes of detected curves may be used to determine a final output, e.g., lane graph, intersection structure, etc., that may be used by vehicle 1100, e.g., by drive stack 608.

[0063] Referring again to FIG. 1 , the landmarks 606 may include any type of landmark, such as, but not limited to, road markings, lane lines, road boundaries, intersection lines, pedestrian paths, bike lane lines, text, poles, trees, light posts, or signs. As described herein, the various classes may include any number of subclasses that the DNN 104 is trained to predict. As an example of the landmarks 606, FIG. 9A includes a visualization 900A of a curve fit to a lane line, FIG. 9B includes a visualization 900B of a curve fit to a road boundary, FIG. 9C includes a visualization 900C of a curve fit to an intersection line, and FIG. 9D includes a visualization of a curve fit to a pole and a sign.

[0064] Referring now to FIG. 10, each block of the method 1000 described herein includes computational processes that may be performed using any combination of hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in a memory. The method 1000 may also be implemented as computer usable instructions stored on a computer storage medium. The method 1000 may be provided by a standalone application, a service, or a hosted service (standalone or in combination with another hosted service), or a plug-in to another product, to name a few. Additionally, the method 1000 is described with respect to the process 600 of FIG. 6 and the vehicle 1100 of FIGS. 11A-11D, as examples. However, the method 1000 may additionally or alternatively be performed by any one process and in any one system, or any combination of processes and systems, including but not limited to those described herein.

[0065] 10 is a flow diagram illustrating a method 1000 of landmark detection using a DNN, according to some embodiments of the present disclosure. The method 1000 includes, at block B1002, applying sensor data representing an image of a sensor's perception field of a machine to a neural network. For example, an instance of sensor data 102 may be applied to the DNN 104.

[0066] The method 1000 includes, at block B1004, calculating first data indicative of pixels of the image within a radius of at least one of the one or more curves identified in the image and second data indicative of distances to control points of the one or more curves based at least in part on the sensor data using a neural network. For example, the DNN 104 may calculate a binary mask 108 and a distance regression mask 110 based at least in part on processing the instances of the sensor data 102.

[0067] The method 1000 includes decoding the first data and the second data to determine positions of the control points of the one or more curves at block B 1006. For example, the binary mask 108 and the distance regression mask 110 may be decoded using the decoder 602 to determine positions of the control points 118, e.g., in 2D image space or 3D world space.

[0068] The method 1000 includes generating one or more curves based at least in part on the positions of the control points using a curve fitting algorithm at block B1008. For example, the curve fitter 116 may use the control points 118 to determine the shape of a curve that corresponds to a landmark in the environment.

[0069] The method 1000 includes performing, by a machine, one or more actions based at least in part on the one or more curves at block B1010. For example, the vehicle 1000 may use the curves (and / or the corresponding semantic information) to perform one or more actions, such as world model management, mapping, planning, control, etc.

[0070] Exemplary Autonomous Vehicle 11A is a diagram of an example autonomous vehicle 1100 according to some embodiments of the disclosure. Autonomous vehicle 1100 (alternatively referred to herein as “vehicle 1100”) may include, but is not limited to, passenger vehicles, such as cars, trucks, buses, first responder vehicles, shuttles, electric or mopeds, motorcycles, fire engines, police vehicles, ambulances, boats, construction vehicles, submarines, drones, and / or other types of vehicles (e.g., unmanned and / or carrying one or more passengers). Autonomous vehicles are generally described in terms of automation levels as defined by the National Highway Traffic Safety Administration (NHTSA), a division of the U.S. Department of Transportation, and the Society of Automotive Engineers (SAE) "Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicle" (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 1100 may be capable of functioning according to one or more of levels 3 through 5 of autonomous driving. For example, vehicle 1100 may be capable of conditional automation (Level 3), highly automated (Level 4), and / or fully automated (Level 5), depending on the embodiment.

[0071] The vehicle 1100 may include components such as a vehicle chassis, body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other components. The vehicle 1100 may include a propulsion system 1150, such as an internal combustion engine, a hybrid power plant, a fully electric engine, and / or another propulsion system type. The propulsion system 1150 may be connected to a drive train of the vehicle 1100, which may include a transmission, to enable propulsion of the vehicle 1100. The propulsion system 1150 may be controlled in response to receiving a signal from a throttle / accelerator 1152.

[0072] A steering system 1154, which may include a steering wheel, may be used to steer the vehicle 1100 (e.g., along a desired course or route) when the propulsion system 1150 is operating (e.g., when the vehicle is moving). The steering system 1154 may receive signals from a steering actuator 1156. The steering wheel may be an option for fully automated (Level 5) functionality.

[0073] The brake sensor system 1146 may be used to operate the vehicle brakes in response to receiving a signal from a brake actuator 1148 and / or a brake sensor.

[0074] The controller 1136, which may include one or more system on chip (SoC) 1104 (FIG. 11C) and / or a GPU, can provide signals (e.g., representations of commands) to one or more components and / or systems of the vehicle 1100. For example, the controller can send signals to operate vehicle brakes via one or more brake actuators 1148, operate a steering system 1154 via one or more steering actuators 1156, and operate a propulsion system 1150 via one or more throttle / accelerators 1152. The controller 1136 may include one or more on-board (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operational commands (e.g., signals representing commands) to enable autonomous driving and / or assist a driver in driving the vehicle 1100. The controllers 1136 may include a first controller 1136 for autonomous driving functions, a second controller 1136 for functional safety functions, a third controller 1136 for artificial intelligence functions (e.g., computer vision), a fourth controller 1136 for infotainment functions, a fifth controller 1136 for redundancy in emergency situations, and / or other controllers. In some instances, a single controller 1136 may handle two or more of the aforementioned functions, and two or more controllers 1136 may handle a single function, and / or any combination thereof.

[0075] The controller 1136 may provide signals to control one or more components and / or systems of the vehicle 1100 in response to sensor data (e.g., sensor inputs) received from one or more sensors. The sensor data may be received from, for example, and without limitation, global navigation satellite system sensors 1158 (e.g., global positioning system sensors), RADAR sensors 1160, ultrasonic sensors 1162, LIDAR sensors 1164, inertial measurement unit (IMU) sensors 1166 (e.g., accelerometers, gyroscopes, magnetic compasses, magnetometers, etc.), microphones 1196, stereo cameras 1168, wide-view cameras 1170 (e.g., fisheye cameras), infrared cameras 1172, surround cameras 1174 (e.g., 360 degree cameras), long-range and / or medium-range cameras 1198, speed sensors 1144 (e.g., for measuring the speed of the vehicle 1100), vibration sensors 1142, steering sensors 1140, brake sensors (e.g., as part of a brake sensor system 1146), and / or other sensor types.

[0076] One or more of the controllers 1136 may receive input (e.g., represented by input data) from the instrument cluster 1132 of the vehicle 1100 and provide output (e.g., represented by output data, display data, etc.) via a human-machine interface (HMI) display 1134, an audible annunciator, a loudspeaker, and / or other components of the vehicle 1100. The output may include information such as vehicle velocity, speed, time, map data (e.g., HD map 1122 of FIG. 11C ), position data (e.g., the position of the vehicle 1100 on a map, etc.), direction, the positions of other vehicles (e.g., occupancy grid), information regarding objects and the status of objects as perceived by the controller 1136, etc. For example, the HMI display 1134 may display information regarding the presence of one or more objects (e.g., road signs, warning signs, changing traffic signals, etc.) and / or a driving maneuver that the vehicle has performed, is performing, or will perform (e.g., changing lanes now, taking exit 34B in 2 miles, etc.).

[0077] Vehicle 1100 further includes a network interface 1124 that can communicate over one or more networks using one or more wireless antennas 1126 and / or a modem. For example, network interface 1124 can be capable of communication over LTE, WCDMA, UMTS, GSM, CDMA2000, etc. Wireless antenna 1126 can also enable communication between objects (e.g., vehicles, mobile devices, etc.) in the environment using local area networks such as Bluetooth, Bluetooth LE, Z-Wave, ZigBee, etc., and / or low power wide-area networks (LPWANs) such as LoRaWAN, SigFox, etc.

[0078] 11B is an illustration of camera positions and fields of view of the example autonomous vehicle 1100 of FIG. 11A, according to some embodiments of the present disclosure. The cameras and their respective fields of view are one example 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 in different positions on the vehicle 1100.

[0079] The camera type of the camera may include, but is not limited to, a digital camera that may be adapted for use with components and / or systems of the vehicle 1100. The camera may be capable of operating at automotive safety integrity level (ASIL) B and / or at another ASIL. The camera type may be capable of any image capture rate, such as 60 frames per second (fps), 120 fps, 240 fps, etc., depending on the embodiment. The camera may be capable of using a rolling shutter, a global shutter, another type of shutter, or a combination thereof. In some instances, the color filter array may include a red clear clear clear (RCCC) color filter array, a red clear clear blue (RCCB) color filter array, a red blue green clear (RBGC) color filter array, a Foveon X3 color filter array, a Bayer sensor (RGGB) color filter array, a monochrome sensor color filter array, and / or another type of color filter array. In some embodiments, clear pixel cameras, such as cameras with RCCC, RCCB, and / or RBGC color filter arrays, may be used in an effort to increase light sensitivity.

[0080] In some instances, one or more of the cameras may 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 mono camera may be installed to provide functions including lane departure warning, traffic sign assist, and intelligent headlamp control. One or more of the cameras (e.g., all of the cameras) may simultaneously record and provide image data (e.g., video).

[0081] One or more of the cameras may be mounted in a mounting part, such as a custom designed (3D printed) part, to filter out stray light and reflections from within the vehicle (e.g., reflections from the dashboard reflected in the windshield mirror) that may interfere with the camera's image data capture capabilities. With reference to a side mirror mounting part, the side mirror part may be custom 3D printed such that the camera mounting plate fits the shape of the side mirror. In some instances, the camera may be integrated into the side mirror. For side view cameras, the cameras may also be integrated into four posts at each corner of the cabin.

[0082] A camera (e.g., a forward-facing camera) with a field of view that includes a portion of the environment in front of the vehicle 1100 may be used for surround view to help identify the forward path and obstacles and, with the aid of one or more controllers 1136 and / or control SoCs, provide information essential to generating an occupancy grid and / or determining a preferred vehicle path. Forward-facing cameras may be used to perform many of the same ADAS functions as LIDAR, including emergency braking, pedestrian detection, and collision avoidance. Forward-facing cameras may also be used for ADAS functions and systems, including other functions such as lane departure warning ("LDW"), autonomous cruise control ("ACC"), and / or traffic sign recognition.

[0083] A variety of cameras may be used in a forward-facing configuration, including, for example, a monocular camera platform that includes a complementary metal oxide semiconductor (CMOS) color imager. Another example may be a wide-view camera 1170 that may be used to understand objects coming into view from the periphery (e.g., pedestrians, crossing traffic, or bicycles). Although only one wide-view camera is shown in FIG. 11B, there may be any number of wide-view cameras 1170 in the vehicle 1100. Additionally, a long-range camera 1198 (e.g., a long-view stereo camera pair) may be used for depth-based object detection, especially for objects for which a neural network has not yet been trained. The long-range camera 1198 may also be used for object detection and classification, as well as basic object tracking.

[0084] One or more stereo cameras 1168 may also be included in the forward-facing configuration. The stereo camera 1168 may include an integrated control unit with an extensible processing unit that may provide programmable logic (FPGA) and a multi-core microprocessor with CAN or Ethernet interfaces integrated on a single chip. Such a unit may be used to generate a 3D map of the vehicle's environment, including distance estimates for all points in the image. An alternative stereo camera 1168 may include a compact stereo vision sensor that may include two camera lenses (one left and one right) and an image processing chip that can measure distances from the vehicle to objects of interest and use the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning functions. Other types of stereo cameras 1168 may be used in addition to or instead of those described herein.

[0085] A camera having a field of view that includes a portion of the environment to the side of the vehicle 1100 (e.g., a side-view camera) may be used for surround view, providing information used to create and update the occupancy grid and to generate side impact collision warnings. For example, surround cameras 1174 (e.g., four surround cameras 1174 as shown in FIG. 11B) may be positioned on the vehicle 1100. The surround cameras 1174 may include a wide-view camera 1170, a fish-eye camera, a 360-degree camera, and / or the like. For example, four fish-eye cameras may be positioned at the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle may use three surround cameras 1174 (e.g., left, right, and rear) and may utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround view camera.

[0086] A camera having a field of view that includes the portion of the environment behind the vehicle 1100 (e.g., a rear-view camera) may be used for parking assistance, surround view, rear collision warning, and creating and updating an occupancy grid. As described herein, a wide variety of cameras may be used, including, but not limited to, cameras that are also suitable as forward-facing cameras (e.g., long-range and / or mid-range camera 1198, stereo camera 1168, infrared camera 1172, etc.).

[0087] FIG. 11C is a block diagram of an example system architecture of the example autonomous vehicle 1100 of FIG. 11A, according to some embodiments of the present disclosure. It should be understood that this and other arrangements described herein are merely illustrative. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, groupings of functions, etc.) may be used in addition to or instead of those shown, and some elements may be omitted altogether. Furthermore, many of the elements described herein are functional entities that may be implemented as separate or distributed components or in conjunction with other components, and in any suitable combination and location. Various functions described herein as being performed by entities may be implemented by hardware, firmware, and / or software. For example, various functions may be implemented by a processor executing instructions stored in a memory.

[0088] Each of the components, features, and systems of the vehicle 1100 of FIG. 11C are illustrated as being connected via a bus 1102. The bus 1102 may include a controller area network (CAN) data interface (alternatively referred to as a “CAN bus”). The CAN may be a network within the vehicle 1100 that is used to help control various features and functions of the vehicle 1100, such as 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 may be read to find steering angle, ground speed, engine revolutions per minute (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0089] The bus 1102 is described herein as being a CAN bus, but this is not intended to be limiting. For example, FlexRay and / or Ethernet may be used in addition to or as an alternative to a CAN bus. Additionally, a single line is used to represent the bus 1102, but this is not intended to be limiting. There may be any number of buses 1102, which may include, for example, 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 instances, two or more buses 1102 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 1102 may be used for collision avoidance functions and a second bus 1102 may be used for operational control. In any instance, each bus 1102 may communicate with any of the components of the vehicle 1100, and two or more buses 1102 may communicate with the same component. In some examples, each SoC 1104, each controller 1136, and / or each computer in the vehicle may have access to the same input data (e.g., input from sensors in the vehicle 1100) and may be connected to a common bus, such as a CAN bus.

[0090] 11A. The controller 1136 may be used for a variety of functions. The controller 1136 may be coupled to any of a variety of other components and systems of the vehicle 1100 and may be used for control of the vehicle 1100, artificial intelligence of the vehicle 1100, infotainment for the vehicle 1100, and / or the like.

[0091] The vehicle 1100 may include a system on a chip (SoC) 1104. The SoC 1104 may include a CPU 1106, a GPU 1108, a processor 1110, a cache 1112, an accelerator 1114, a data store 1116, and / or other components and features not shown. The SoC 1104 may be used to control the vehicle 1100 in a variety of platforms and systems. For example, the SoC 1104 may be coupled in a system (e.g., the system of the vehicle 1100) having an HD map 1122 that can obtain map refreshes and / or updates via a network interface 1124 from one or more servers (e.g., server 1178 of FIG. 11D ).

[0092] The CPU 1106 may include a CPU cluster or CPU complex (alternatively referred to as a "CCPLEX"). The CPU 1106 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU 1106 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 1106 may include four dual-core clusters, each with its own dedicated L2 cache (e.g., a 2M B L2 cache). The CPU 1106 (e.g., a CCPLEX) may be configured to support concurrent cluster operation allowing any combination of the CPU 1106 clusters to be active at any given time.

[0093] The CPU 1106 may implement power management capabilities including one or more of the following features: individual hardware blocks may be automatically clock gated when idle to conserve dynamic power, each core clock may be gated when the core is not actively executing instructions by execution of WFI / WFE instructions, each core may be independently power gated, each core cluster may be independently clock gated when all cores are clock gated or power gated, and / or each core cluster may be independently power gated when all cores are power gated. The CPU 1106 may further implement an enhanced algorithm for managing power states where allowed power states and expected wake-up times are specified and the hardware / microcode determines the best power state for entering the cores, clusters, and CCPLEXes. The processing cores may support simplified power state entry sequences in software where work is offloaded to the microcode.

[0094] The GPU 1108 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The GPU 1108 may be programmable and efficient for parallel workloads. In some instances, the GPU 1108 may use an enhanced tensor instruction set. The GPU 1108 may include one or more streaming microprocessors, where each streaming microprocessor may include an L1 cache (e.g., an L1 cache having at least 96 KB storage capacity) and two or more of the streaming microprocessors may share a cache (e.g., an L2 cache having 512 KB storage capacity). In some embodiments, the GPU 1108 may include at least eight streaming microprocessors. The GPU 1108 may use a computational application programming interface (API). In addition, the GPU 1108 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0095] The GPU 1108 may be power optimized for best performance in automotive and embedded use cases. For example, the GPU 1108 may be fabricated on FinFETs (Fin field-effect transistors). However, this is not intended to be limiting and the GPU 1108 may be fabricated using other semiconductor fabrication processes. Each streaming microprocessor may incorporate several mixed-precision processing cores partitioned into multiple blocks. For example, and without limitation, 64 PF32 cores and 32 PF64 cores may be partitioned into four processing blocks. In such an example, each processing block may be assigned 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two mixed-precision NVIDIA tensor cores for deep learning matrix operations, an L0 instruction cache, a warp scheduler, a dispatch unit, and / or a 64KB register file. In addition, the streaming microprocessor may include independent parallel integer and floating-point data paths to provide efficient execution of workloads having a mix of computational and addressing operations. Streaming microprocessors may include independent thread scheduling capabilities to allow finer grain synchronization and coordination among concurrent threads. Streaming microprocessors may include a combined L1 data cache and shared memory unit to improve performance while simplifying programming.

[0096] The GPU 1108 may include high bandwidth memory (HBM) and / or a 16GB HBM2 memory subsystem to provide about 900GB / sec peak memory bandwidth in some instances. In some instances, synchronous graphics random-access memory (SGRAM), such as graphics double data rate type five synchronous random-access memory (GDDR5), may be used in addition to or in place of the HBM memory.

[0097] The GPU 1108 may include unified memory technology including access counters to enable more accurate movement of memory pages to the processors that access them most frequently, thereby improving the efficiency of storage ranges shared between processors. In some instances, address translation service (ATS) support may be used to enable the GPU 1108 to directly access the CPU 1106 page tables. In such instances, when the GPU 1108 memory management unit (MMU) experiences a miss, an address translation request may be sent to the CPU 1106. In response, the CPU 1106 may consult its page tables for a virtual-to-real mapping of the address and send the translation back to the GPU 1108. As such, the unified memory technology may enable a single unified virtual address space for both the CPU 1106 and GPU 1108 memories, thereby simplifying GPU 1108 programming and porting of applications to the GPU 1108.

[0098] In addition, the GPU 1108 may include access counters that can record the frequency of accesses of the GPU 1108 to the memory of other processors. The access counters can help ensure that memory pages are moved to the physical memory of the processor that is accessing the page most frequently.

[0099] The SoC 1104 may include any number of caches 1112, including those described herein. For example, the cache 1112 may include an L3 cache available to both the CPU 1106 and the GPU 1108 (e.g., connected to both the CPU 1106 and the GPU 1108). The cache 1112 may include a write-back cache that can record line state, such as by using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). The L3 cache may include 4 MB or more, depending on the implementation, although smaller cache sizes may be used.

[0100] The SoC 1104 may include an arithmetic logic unit (ALU) that may be utilized in performing processing for any of the various tasks or operations (e.g., processing DNNs) of the vehicle 1100. In addition, the SoC 1104 may include a floating point unit (FPU) (or other math co-processor or numeric co-processor type) for performing mathematical operations within the system. For example, the SoC 104 may include one or more FPUs integrated as execution units within the CPU 1106 and / or GPU 1108.

[0101] The SoC 1104 may include one or more accelerators 1114 (e.g., hardware accelerators, software accelerators, or a combination thereof). For example, the SoC 1104 may include a hardware acceleration cluster, which may include optimized hardware accelerators and / or large on-chip memory. The large on-chip memory (e.g., 4MB of SRAM) may enable the hardware acceleration cluster to accelerate neural networks and other operations. The hardware acceleration cluster may be used to complement the GPU 1108 and to offload some of the tasks of the GPU 1108 (e.g., to free up more cycles for the GPU 1108 to perform other tasks). As an example, the accelerator 1114 may be used for target workloads that are stable enough to be suitable for acceleration (e.g., perception, convolutional neural networks (CNNs), etc.). As used herein, the term "CNN" may include all types of CNNs, including region-based or regional convolutional neural networks (RCNNs) and Faster RCNNs (e.g., as used for object detection).

[0102] The accelerator 1114 (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 and optimized to perform image processing functions (e.g., CNN, RCNN, etc.). The DLA may further be optimized for a specific set of neural network types and floating point operations, as well as inference. The design of the DLA may provide more performance per millimeter than a general-purpose GPU, greatly exceeding the performance of a CPU. The TPU may perform several functions, including, for example, single instance convolution functions, supporting INT8, INT16, and FP16 data types for both features and weights, as well as post-processor functions.

[0103] The DLA can quickly and efficiently run neural networks, in particular CNNs, on processed or unprocessed data for any of a variety of functions, including but not limited to: CNNs for object identification and detection using data from camera sensors, CNNs for distance estimation using data from camera sensors, CNNs for emergency vehicle detection and identification using data from microphones, CNNs for face recognition and vehicle owner identification using data from camera sensors, and / or CNNs for security and / or safety related events.

[0104] The DLA can perform any function of the GPU 1108, and by using an inference accelerator, for example, a designer can target either the DLA or the GPU 1108 for any function. For example, a designer can focus on processing CNNs and floating point operations on the DLA and offload other functions to the GPU 1108 and / or other accelerators 1114.

[0105] The accelerator 1114 (e.g., 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 may provide a balance between performance and flexibility. For example, each PVA may include, but is not limited to, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors, for example.

[0106] The RISC cores may interact with an image sensor (e.g., an image sensor of any of the cameras described herein), an image signal processor, and / or the like. Each RISC core may include any amount of memory. The RISC cores may use any of a number of protocols, depending on the embodiment. In some instances, the RISC cores may run a real-time operating system (RTOS). The RISC cores may be implemented using one or more integrated circuit devices, application specific integrated circuits (ASICs), and / or memory devices. For example, the RISC cores may include an instruction cache and / or tightly coupled RAM.

[0107] The DMA may allow components of the PVA to access system memory independent of the CPU 1106. The DMA may support any number of features used to provide optimizations to the PVA, including but not limited to supporting multi-dimensional addressing and / or circular addressing. In some instances, the DMA may support up to six or more dimensions of addressing, which may include block width, block height, block depth, horizontal block stepping, vertical block stepping, and / or depth stepping.

[0108] The vector processor may be a programmable processor that may be designed to efficiently and flexibly execute computer vision algorithm programming and provide signal processing capabilities. In some instances, the PVA may include a PVA core and two vector processing subsystem partitions. The PVA core may include a processor subsystem, a DMA engine (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem may act as the primary processing engine of the PVA and may include a vector processing unit (VPU), an instruction cache, and / or a vector memory (e.g., VMEM). The VPU core may include a digital signal processor, such as, for example, a single instruction, multiple data (SIMD), very long instruction word (VLIW) digital signal processor. The combination of SIMD and VLIW may increase throughput and speed.

[0109] Each vector processor may include an instruction cache and may be coupled to dedicated memory. As a result, in some instances, each vector processor may be configured to execute independently of other vector processors. In other instances, the vector processors included in a particular PVA may be configured to employ data parallelism. For example, in some embodiments, multiple vector processors included in a single PVA may execute the same computer vision algorithm, but on different regions of an image. In other instances, the vector processors included in a particular PVA may execute different computer vision algorithms simultaneously on the same image, or even execute different algorithms on sequential images or portions of an image. In particular, any number of PVAs may be included in a hardware accelerated cluster, and any number of vector processors may be included in each PVA. In addition, the PVAs may include additional error correcting code (ECC) memory to increase overall system security.

[0110] The accelerator 1114 (e.g., a hardware acceleration cluster) may include a computer vision network-on-chip and SRAM to provide high bandwidth, low latency SRAM for the accelerator 1114. In some instances, the on-chip memory may include, for example and without limitation, at least 4 MB of SRAM consisting of eight field configurable memory blocks that may be accessible by both the PVA and DLA. Each pair of memory blocks may include an advanced peripheral bus (APB) interface, configuration circuitry, controllers, and multiplexers. Any type of memory may be used. The PVA and DLA may access the memory through a backbone that provides the PVA and DLA with high speed access to the memory. The backbone may include a computer vision network-on-chip that interconnects the PVA and DLA to the memory (e.g., using the APB).

[0111] The computer vision network-on-chip may include an interface that determines, prior to transmission of any control signals / addresses / data, that both the PVA and DLA provide ready and valid signals. Such an interface may provide separate phases and separate channels for transmitting control signals / addresses / data, as well as burst type communication for continuous data transfer. This type of interface may follow ISO26262 or IEC61508 standards, although other standards and protocols may be used.

[0112] In some examples, the SoC 1104 may include a real-time ray tracing hardware accelerator, such as described in U.S. Patent Application No. 16 / 101,232, filed August 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the location and scale of objects (e.g., within a world model) to generate real-time visualization simulations for RADAR signal interpretation, for acoustic propagation synthesis and / or analysis, for simulation of SONAR systems, for general wave propagation simulation, for comparison against LIDAR data for localization and / or other functions, and / or for other uses. In some examples, one or more tree traversal units (TTUs) may be used to perform one or more ray tracing related operations.

[0113] The accelerator 1114 (e.g., a hardware accelerator cluster) has a variety of applications for autonomous driving. The PVA may be a programmable vision accelerator that can be used for critical processing stages in ADAS and autonomous vehicles. The capabilities of the PVA are well suited to the domain of algorithms that require predictable processing at low power and low latency. In other words, the PVA works well with semi-dense or dense regular computations on small data sets that require predictable execution times with low latency and low power. Hence, in the context of a platform for autonomous vehicles, the PVA is designed to run classical computer vision algorithms, since the PVA is efficient in object detection and operating on integer computations.

[0114] For example, according to one embodiment of the present technology, the PVA is used to perform computer stereo vision. A semi-global matching based algorithm may be used in some instances, but this is not intended to be limiting. Many applications for Level 3-5 autonomous driving require motion estimation / stereo matching on the fly (e.g., structure from motion, pedestrian recognition, lane detection, etc.). The PVA can perform computer stereo vision functions with input from two monocular cameras.

[0115] In some instances, PVA may be used to perform dense optical flow by processing raw RADAR data (e.g., using a 4D Fast Fourier Transform) to provide processed RADAR. In other instances, PVA is used for time of flight depth processing, for example, by processing raw time of flight data to provide processed time of flight data.

[0116] DLA can be used to implement any type of network to enhance control and driving safety, including, for example, a neural network that outputs a measure of the confidence of each object detection. Such confidence value can be interpreted as a probability or as providing a relative "weight" of each detection compared to other detections. This confidence value allows the system to make further decisions regarding which detections should be considered true positive detections rather than false positive detections. For example, the system can set a confidence threshold and consider only detections above the threshold as true positive detections. In an automatic emergency braking (AEB) system, a false positive detection would cause the vehicle to automatically perform emergency braking, which is clearly undesirable. Therefore, only the most confident detections should be considered as triggers for AEB. DLA can implement a neural network that regresses the confidence value. The neural network may receive as its inputs at least some subset of parameters, such as bounding box dimensions, ground plane estimates obtained (e.g., from another subsystem), inertial measurement unit (IMU) sensor 1166 outputs that correlate with the vehicle 1100 orientation, distance, and 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LIDAR sensor 1164 or RADAR sensor 1160), and so forth.

[0117] The SoC 1104 may include a data store 1116 (e.g., a memory). The data store 1116 may be an on-chip memory of the SoC 1104 and may store the neural network to be executed on the GPU and / or DLA. In some instances, the data store 1116 may have a capacity large enough to store multiple instances of the neural network for redundancy and safety. The data store 1112 may comprise an L2 or L3 cache 1112. References to the data store 1116 may include references to memory associated with the PVA, DLA, and / or other accelerators 1114, as described herein.

[0118] The SoC 1104 may include one or more processors 1110 (e.g., embedded processors). The processors 1110 may include a boot and power management processor, which may be a dedicated processor and subsystem for handling boot power and management capabilities and associated security enforcement. The boot and power management processor may be part of the SoC 1104 boot sequence and may provide run-time power management services. The boot power and management processor may provide clock and voltage programming, assist with system low power state transitions, management of the SoC 1104 thermal and temperature sensors, and / or management of the SoC 1104 power states. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and the SoC 1104 may use the ring oscillator to detect the temperature of the CPU 1106, the GPU 1108, and / or the accelerator 1114. If the temperature is determined to exceed a threshold, the boot and power management processor may enter a temperature fault routine, place the SoC 1104 in a lower power state and / or place the vehicle 1100 in a Chauffeur safe shutdown mode (e.g., bring the vehicle 1100 to a safe shutdown).

[0119] The processor 1110 may further include a set of embedded processors that may perform the functions of an audio processing engine. The audio processing engine may be an audio subsystem that allows full hardware support of multi-channel audio through multiple interfaces and a wide and flexible range of audio I / O interfaces. In some instances, the audio processing engine is a dedicated processor core that includes a digital signal processor with dedicated RAM.

[0120] The processor 1110 may further include an always-on processor engine that can provide the necessary hardware features to support low power sensor management and wake use cases. The always-on processor engine may include a processor core, tightly coupled RAM, support peripherals (e.g., timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0121] The processor 1110 may further include a safety cluster engine that includes a processor subsystem dedicated to handling safety management of automotive applications. The safety cluster engine may include two or more processor cores, tightly coupled RAM, supporting peripherals (e.g., timers, interrupt controllers, etc.), and / or routing logic. In a safety mode, the two or more cores may operate in lockstep mode and function as a single core with comparison logic to detect any differences between their operations.

[0122] The processor 1110 may further include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0123] The processor 1110 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.

[0124] The processor 1110 may include a video image compositor, which may be a processing block (e.g., implemented in a microprocessor) that implements video post-processing functions required by the video playback application to produce a final image for the player window. The video image compositor may perform lens distortion correction on the wide-view camera 1170, on the surround camera 1174, and / or on the in-cabin surveillance camera sensor. The in-cabin surveillance camera sensor is preferably monitored by a neural network running on another instance of the advanced SoC, configured to identify in-cabin events and respond appropriately. The in-cabin system may perform lip reading to activate cellular service and make phone calls, dictate emails, change the vehicle's destination, activate or change the vehicle's infotainment system and settings, or provide voice-activated web surfing. Certain features are only available to the driver when operating in autonomous mode and are disabled otherwise.

[0125] The video image synthesizer may include enhanced temporal noise reduction for both spatial and temporal noise reduction. For example, when motion occurs in the video, the noise reduction reduces the weight of information provided by adjacent frames and appropriately weights the spatial information. When an image or part of an image does not contain motion, the temporal noise reduction performed by the video image synthesizer may use information from previous images to reduce noise in the current image.

[0126] The video image composer may also be configured to perform stereo rectification on the input stereo lens frames. The video image composer may further be used for user interface compositing when the operating system desktop is in use and the GPU 1108 is not required to continuously render new surfaces. Even when the GPU 1108 is powered on and actively performing 3D rendering, the video image composer may be used to offload the GPU 1108 for improved performance and responsiveness.

[0127] The SoC1104 may further include a mobile industry processor interface (MIPI) camera serial interface for receiving video and input from a camera, a high-speed interface, and / or a video input block that may be used for camera and related pixel input functions. The SoC1104 may further include an input / output controller that may be controlled by software and that may be used to receive I / O signals that are not committed to a specific role.

[0128] The SoC 1104 may further include a wide range of peripheral interfaces to enable peripherals, audio codecs, power management, and / or communication with other devices. The SoC 1104 may be used to process data from cameras (e.g., connected via gigabit multimedia serial links and Ethernet), sensors (e.g., LIDAR sensors 1164, RADAR sensors 1160, etc., which may be connected via Ethernet), data from the bus 1102 (e.g., vehicle 1100 speed, steering wheel position, etc.), and data from the GNSS sensor 1158 (e.g., connected via Ethernet or CAN bus). The SoC 1104 may further include a dedicated high-performance mass storage controller, which may include its own DMA engine and may be used to offload the CPU 1106 from routine data management tasks.

[0129] The SoC 1104 may be an end-to-end platform with a flexible architecture that spans levels 3-5 of automation, thereby providing a comprehensive functional safety architecture that leverages and efficiently uses computer vision and ADAS techniques for diversity and redundancy, and provides a platform for a flexible, reliable driving software stack along with deep learning tools. The SoC 1104 may be faster, more reliable, and more energy- and space-efficient than traditional systems. For example, when the accelerator 1114 is combined with the CPU 1106, the GPU 1108, and the data store 1116, the accelerator 1114 may provide a fast and efficient platform for levels 3-5 of autonomous vehicles.

[0130] Thus, the present technology provides capabilities and functionality that cannot be achieved by conventional systems. For example, computer vision algorithms may be executed on a CPU, which may be configured using a high-level programming language, such as the C programming language, to execute a wide variety of processing algorithms across a wide variety 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 in-vehicle ADAS applications and for practical Level 3-5 autonomous vehicles.

[0131] In contrast to conventional systems, by providing a CPU complex, a GPU complex, and a hardware acceleration cluster, the techniques described herein allow multiple neural networks to be run simultaneously and / or sequentially and the results to be combined to enable Level 3-5 autonomous driving capabilities. For example, a CNN running on the DLA or dGPU (e.g., GPU1120) can include text and word recognition that allows the supercomputer to read and understand traffic signs, including signs for which the neural network was not specifically trained. The DLA can further include neural networks that can identify, interpret, and provide a semantic understanding of the signs and pass the semantic understanding to a route planning module running on the CPU complex.

[0132] As another example, multiple neural networks may be run simultaneously, as required for level 3, 4, or 5 driving. For example, a warning sign consisting of "Caution: Flashing lights indicate icy conditions" along with a lightning flash may be interpreted by several neural networks, independently or collectively. The sign itself may be identified as a traffic sign by a first deployed neural network (e.g., a trained neural network), and the text "Flashing lights indicate icy conditions" may be interpreted by a second deployed neural network that informs the vehicle's route planning software (preferably running on the CPU complex) that icy conditions exist when a flashing light is detected. The flashing light may be identified by running a third deployed neural network through multiple frames, informing the vehicle's route planning software of the presence (or absence) of the flashing light. All three neural networks may be run simultaneously, such as within the DLA and / or on the GPU 1108.

[0133] In some instances, a CNN for facial recognition and vehicle owner identification can use data from the camera sensors to identify the presence of a legitimate driver and / or owner of the vehicle 1100. An always-on sensor processing engine can be used to unlock the vehicle and turn on the lights when the owner approaches the driver's door, and in security mode, to disable vehicle operation when the owner leaves the vehicle. In this manner, the SoC 1104 provides security against theft and / or carjacking.

[0134] In another example, a CNN for emergency vehicle detection and identification can detect and identify emergency vehicle sirens using data from microphone 1196. In contrast to conventional systems that use general classifiers to detect sirens and manually extract features, SoC 1104 uses CNNs for classification of environmental and urban sounds, as well as classification of visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative terminal velocity of emergency vehicles (e.g., by using the Doppler effect). The CNN can also be trained to identify emergency vehicles specific to the local area in which the vehicle is operating, as identified by the GNSS sensor 1158. Thus, for example, when operating in Europe, the CNN will attempt to detect European sirens, and when in the United States, the CNN will attempt to identify only North American sirens. After an emergency vehicle is detected, a control program can be used to perform emergency vehicle safety routines, such as slowing the vehicle down, stopping at the side of the road, parking the vehicle, and / or idling the vehicle, with the assistance of ultrasonic sensor 1162, until the emergency vehicle has passed.

[0135] The vehicle may include a CPU 1118 (e.g., a separate CPU, or a dCPU) that may be coupled to the SoC 1104 via a high-speed interconnect (e.g., PCIe). The CPU 1118 may include, for example, an X86 processor. The CPU 1118 may be used to perform any of a variety of functions, including, for example, reconciling potentially inconsistent results between the ADAS sensors and the SoC 1104, and / or monitoring the status and health of the controller 1136 and / or the infotainment SoC 1130.

[0136] The vehicle 1100 may include a GPU 1120 (e.g., a discrete GPU, or a dGPU) that may be coupled to the SoC 1104 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU 1120 may provide additional artificial intelligence functionality, such as by running redundant and / or different neural networks, and may be used to train and / or update neural networks based on input (e.g., sensor data) from sensors of the vehicle 1100.

[0137] The vehicle 1100 may further include a network interface 1124, which may include one or more wireless antennas 1126 (e.g., one or more wireless antennas for different communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 1124 may be used to enable wireless connections with the cloud over the Internet (e.g., with the server 1178 and / or other network devices), with other vehicles, and / or with computing devices (e.g., passenger client devices). To communicate with other vehicles, a direct link may be established between the two vehicles and / or an indirect link may be established (e.g., through a network and via the Internet). A direct link may be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link may provide the vehicle 1100 information about vehicles in close proximity to the vehicle 1100 (e.g., vehicles in front of, beside, and / or behind the vehicle 1100). This function may be part of a collaborative adaptive cruise control function of the vehicle 1100.

[0138] The network interface 1124 may include a SoC that provides modulation and demodulation functions and enables the controller 1136 to communicate over a wireless network. The network interface 1124 may include a radio frequency front end for upconversion from baseband to radio frequency and downconversion from radio frequency to baseband. The frequency conversion may be performed through well known processes and / or may be performed using a superheterodyne process. In some instances, the radio frequency front end functions may be provided by a separate chip. The network interface may include wireless functions for communicating via LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-Wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0139] The vehicle 1100 may further include a data store 1128, which may include off-chip (e.g., off-SoC 1104) storage. The data store 1128 may include one or more memory elements, including RAM, SRAM, DRAM, VRAM, flash, hard disk, and / or other components and / or devices capable of storing at least one bit of data.

[0140] The vehicle 1100 may further include a GNSS sensor 1158. The GNSS sensor 1158 (e.g., a GPS, an aided GPS sensor, a differential GPS (DGPS) sensor, etc.) aids in mapping, perception, occupancy grid generation, and / or route planning functions. Any number of GNSS sensors 1158 may be used, including, but not limited to, a GPS using a USB connector with an Ethernet to serial (RS-232) bridge, for example.

[0141] The vehicle 1100 may further include a RADAR sensor 1160. The RADAR sensor 1160 may be used by the vehicle 1100 for long-range vehicle detection, even in darkness and / or severe weather conditions. The RADAR functional safety level may be ASIL B. In some instances, the RADAR sensor 1160 may use the CAN and / or bus 1102 for control and to access object tracking data (e.g., to transmit data generated by the RADAR sensor 1160), with access to Ethernet to access raw data. A wide variety of RADAR sensor types may be used. For example, and without limitation, the RADAR sensor 1160 may be suitable for front, rear, and side RADAR use. In some instances, a pulsed Doppler RADAR sensor is used.

[0142] The RADAR sensor 1160 may include different configurations, such as long range with a narrow field of view, short range with a wide field of view, and short range side coverage. In some instances, the long range RADAR may be used for adaptive cruise control functions. Long range RADAR systems may provide a wide field of view achieved by two or more independent scans, such as within a 250m range. The RADAR sensor 1160 may help distinguish between static and moving objects and may be used by ADAS systems for emergency brake assist and forward collision warning. Long range RADAR sensors may include monostatic multimodal RADAR with multiple (e.g., six or more) fixed RADAR antennas and high speed CAN and FlexRay interfaces. In one instance with six antennas, the center four antennas may create a focused beam pattern designed to record the surroundings of the vehicle 1100 at high speeds with minimal interference from traffic in adjacent lanes. The other two antennas may widen the field of view, allowing for quick detection of vehicles entering or leaving the lane of the vehicle 1100.

[0143] As an example, a medium-range RADAR system may include a range of up to 1160m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 1150 degrees (rear). A short-range RADAR system may include, but is not limited to, a RADAR sensor designed to be mounted on either end of the rear bumper. When mounted on either end of the rear bumper, such a RADAR sensor system can create two beams that constantly monitor the blind spots behind and adjacent to the vehicle.

[0144] Short-range RADAR systems may be used in ADAS systems for blind spot detection and / or lane change assist.

[0145] The vehicle 1100 may further include ultrasonic sensors 1162. The ultrasonic sensors 1162, which may be located on the front, rear, and / or sides of the vehicle 1100, may be used for parking assist and / or for creating and updating an occupancy grid. A wide variety of ultrasonic sensors 1162 may be used, with different ultrasonic sensors 1162 being used for different ranges of detection (e.g., 2.5m, 4m). The ultrasonic sensors 1162 may operate at a functional safety level of ASIL B.

[0146] The vehicle 1100 may include a LIDAR sensor 1164. The LIDAR sensor 1164 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The LIDAR sensor 1164 may be functional safety level ASIL B. In some instances, the vehicle 1100 may include multiple (e.g., 2, 4, 6, etc.) LIDAR sensors 1164 that may use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0147] In some instances, the LIDAR sensor 1164 may have the ability to provide a list of objects and their distances in a 360-degree field of view. A commercially available LIDAR sensor 1164 may have an advertised range of about 1100 m, for example, with an accuracy of 2 cm to 3 cm, and with support for an 1100 Mbps Ethernet connection. In some instances, one or more non-protruding LIDAR sensors 1164 may be used. In such instances, the LIDAR sensor 1164 may be implemented as a small device that may be integrated into the front, rear, sides, and / or corners of the vehicle 1100. In such instances, the LIDAR sensor 1164 may have a range of 200 m even for low-reflecting objects, and provide up to 120 degrees horizontal and 35 degrees vertical field of view. A front-mounted LIDAR sensor 1164 may be configured for a horizontal field of view between 45 degrees and 135 degrees.

[0148] In some instances, LIDAR technology such as 3D flash LIDAR may also be used. 3D flash LIDAR uses a laser flash as a transmitter to illuminate the surroundings of the vehicle up to about 200 m. The flash LIDAR unit includes a receptor that records the laser pulse transit time and the reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LIDAR may allow a highly accurate and distortion-free image of the surroundings to be generated with every laser flash. In some instances, four flash LIDAR sensors may be deployed, one on each side of the vehicle 1100. Available 3D flash LIDAR systems include solid-state 3D steering array LIDAR cameras with no moving parts other than the blower (e.g., non-scanning LIDAR devices). Flash LIDAR devices may use 5 nanosecond Class I (eye-safe) laser pulses per frame and may capture reflected laser light in the form of a 3D range point cloud and coregistered intensity data. By using flash LIDAR, and because flash LIDAR is a solid-state device with no moving parts, the LIDAR sensor 1164 may be less susceptible to motion blur, vibration, and / or shock.

[0149] The vehicle may further include an IMU sensor 1166. In some instances, the IMU sensor 1166 may be positioned at the center of the rear axle of the vehicle 1100. The IMU sensor 1166 may include, for example, but not limited to, an accelerometer, a magnetometer, a gyroscope, a magnetic compass, and / or other sensor types. In some instances, such as in a 6-axis application, the IMU sensor 1166 may include an accelerometer and a gyroscope, while in a 9-axis application, the IMU sensor 1166 may include an accelerometer, a gyroscope, and a magnetometer.

[0150] In some embodiments, the IMU sensor 1166 may be implemented as a miniature, high-performance GPS-Aided Inertial Navigation System (GPS / INS) that combines micro-electro-mechanical system (MEMS) inertial sensors, highly sensitive GPS receivers, and advanced Kalman filtering algorithms to provide estimates of position, velocity, and attitude. As such, in some instances, the IMU sensor 1166 may enable the vehicle 1100 to estimate heading without requiring input from a magnetic sensor by directly observing and correlating changes in velocity from the GPS to the IMU sensor 1166. In some instances, the IMU sensor 1166 and the GNSS sensor 1158 may be combined in a single integrated unit.

[0151] The vehicle may include microphones 1196 placed in and / or around the vehicle 1100. The microphones 1196 may be used for emergency vehicle detection and identification, among other things.

[0152] The vehicle may further include any number of camera types, including stereo cameras 1168, wide view cameras 1170, infrared cameras 1172, surround cameras 1174, long and / or mid-range cameras 1198, and / or other camera types. The cameras may be used to capture image data around the entire exterior of the vehicle 1100. The type of cameras used depends on the embodiment and requirements of the vehicle 1100, and any combination of camera types may be used to achieve the required coverage around the vehicle 1100. Additionally, the number of cameras may vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. The cameras may support, by way of example only, Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet, without limitation. Each camera is described in further detail herein with respect to FIGS. 11A and 11B.

[0153] The vehicle 1100 may further include a vibration sensor 1142. The vibration sensor 1142 may measure vibrations of vehicle components, such as an axle. For example, a change in vibration may indicate a change in the surface of the road. In another example, when two or more vibration sensors 1142 are used, the difference in vibration may be used to determine friction or slippage of the road surface (e.g., when the difference in vibration is between a powered axle and a free-running axle).

[0154] The vehicle 1100 may include an ADAS system 1138. In some instances, the ADAS system 1138 may include a SoC. The ADAS system 1138 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward crash warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keep assist (LKA), blind spot warning (BSW), rear cross-traffic warning (RCTW), collision warning system (CWS), lane centering (LC), and / or other features and functions.

[0155] The ACC system may use a RADAR sensor 1160, a LIDAR sensor 1164, and / or a camera. The ACC system may include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to the vehicle directly ahead of the vehicle 1100 and automatically adjusts the vehicle speed to maintain a safe distance from the vehicle ahead. Lateral ACC performs distance keeping and advises the vehicle 1100 to change lanes when necessary. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0156] CACC uses information from other vehicles, which may be received from other vehicles via a wireless link via the network interface 1124 and / or the wireless antenna 1126, or indirectly via a network connection (e.g., via the Internet). A direct link may be provided by a vehicle-to-vehicle (V2V) communication link, while an indirect link may be an infrastructure-to-vehicle (I2V) communication link. In general, a V2V communication concept provides information about the immediately preceding vehicle (e.g., the vehicle directly ahead of the vehicle 1100, in the same lane as the vehicle 1100), while an I2V communication concept provides information about the traffic further ahead. A CACC system may include either or both I2V and V2V information sources. Given the information of the vehicles ahead of the vehicle 1100, CACC may be more reliable and may have the potential to make traffic flow smoother and reduce congestion on the roads.

[0157] The FCW system is designed to warn the driver of hazards so that the driver can take corrective action. The FCW system uses a forward-facing camera and / or RADAR sensor 1160 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, electrically coupled to driver feedback such as a display, speaker, and / or vibration components. The FCW system can provide warnings in the form of audio, visual alerts, vibrations, and / or quick brake pulses.

[0158] An AEB system can detect an imminent forward collision with another vehicle or other object and automatically apply the brakes if the driver does not take corrective action within specified time or distance parameters. The AEB system can use a forward-facing camera and / or RADAR sensor 1160 coupled to a dedicated processor, DSP, FPGA, and / or ASIC. When the AEB system detects a hazard, the AEB system typically first alerts the driver to take corrective action to avoid the collision, and if the driver does not take corrective action, the AEB system can automatically apply the brakes as part of an effort to prevent, or at least mitigate, the effects of the predicted collision. The AEB system can include techniques such as Dynamic Brake Support and / or Collision Imminent Braking.

[0159] The LDW system provides visual, audible, and / or tactile warnings, such as vibrations in the steering wheel or seat, to alert the driver when the vehicle 1100 crosses a lane marking. The LDW system does not activate when the driver indicates an intentional lane departure by activating a turn signal. The LDW system may use a forward-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC electrically coupled to driver feedback, such as a display, speaker, and / or vibration components.

[0160] The LKA system is a modification of the LDW system, which provides steering input or braking to correct the vehicle 1100 if it begins to drift out of its lane.

[0161] The BSW system detects and warns the vehicle driver in the vehicle's blind spot. The BSW system can provide visual, audible, and / or tactile warnings to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses a turn signal. The BSW system can use rear-facing cameras and / or RADAR sensors 1160 coupled to dedicated processors, DSPs, FPGAs, and / or ASICs electrically coupled to driver feedback, e.g., displays, speakers, and / or vibration components.

[0162] The RCTW system can provide visual, audible, and / or tactile notifications when an object is detected outside the range of the rear camera when the vehicle 1100 is backing up. Some RCTW systems include AEB to ensure vehicle brakes are applied to avoid a collision. The RCTW system can use one or more rear-facing RADAR sensors 1160 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, electrically coupled to driver feedback, e.g., a display, speaker, and / or vibration components.

[0163] Because conventional ADAS systems alert the driver and allow the driver to determine whether a safety condition truly exists and act accordingly, conventional ADAS systems can be prone to producing false positive results that are usually not catastrophic but can be annoying and distracting to the driver. However, in an autonomous vehicle 1100, the vehicle 1100 itself must decide whether to listen to the results from the primary computer or the secondary computer (e.g., the first controller 1136 or the second controller 1136) when the results are conflicting. For example, in some embodiments, the ADAS system 1138 may be a backup and / or secondary computer to provide perception information to a backup computer rationality module. The backup computer rationality monitor can run redundant and diverse software on hardware components to detect failures in perception and dynamic driving tasks. Output from the ADAS system 1138 may be provided to a supervisory MCU. When the outputs from the primary and secondary computers conflict, the supervisory MCU must decide how to reconcile the conflict to ensure safe operation.

[0164] In some instances, the primary computer may be configured to provide a reliability score to the overseer MCU indicating the reliability of the primary computer in a selected outcome. If the reliability score exceeds a threshold, the overseer MCU may follow the instructions of the primary computer regardless of whether the secondary computers provide conflicting or inconsistent results. If the reliability score does not meet the threshold, and if the primary and secondary computers show different (e.g., conflicting) results, the overseer MCU may arbitrate between the computers to determine the appropriate outcome.

[0165] The supervisory MCU may be configured to execute a neural network that is trained and configured to determine conditions under which the secondary computer provides a false alarm based on the outputs from the primary and secondary computers. Thus, the neural network in the supervisory MCU may learn when the output of the secondary computer may be trusted and when it may not be trusted. For example, when the secondary computer is a RADAR-based FCW system, the neural network in the supervisory MCU may learn when the FCW identifies a metal object that is not actually a hazard, such as a sewer grate or manhole cover, that triggers an alarm. Similarly, when the secondary computer is a camera-based LDW system, the neural network in the supervisory MCU may learn to ignore the LDW when a bicyclist or pedestrian is present and lane departure is, in fact, the safest maneuver. In embodiments that include a neural network executing on the supervisory MCU, the supervisory MCU may include at least one of a DLA or a GPU suitable for executing the neural network with associated memory. In a preferred embodiment, the director MCU may comprise and / or be included as a component of the SoC 1104.

[0166] In other instances, the ADAS system 1138 may include a secondary computer that performs ADAS functions using classical rules of computer vision. As such, the secondary computer may use classical computer vision rules (if-then) and the presence of a neural network in the supervisory MCU may improve reliability, safety and performance. For example, the diverse implementations and intentional non-identity may make the overall system more fault-tolerant, especially to failures caused by software (or software-hardware interface) functions. For example, if there is a software bug or error in the software running on the primary computer and non-identical software code running on the secondary computer provides the same overall result, the supervisory MCU may have greater confidence that the overall result is correct and that a bug in the software or hardware on the primary computer has not caused a critical error.

[0167] In some instances, the output of the ADAS system 1138 may be fed to a perception block of the primary computer and / or a dynamic driving task block of the primary computer. For example, if the ADAS system 1138 indicates a forward collision warning due to an object directly ahead, the perception block may use this information when identifying the object. In other instances, the secondary computer may have its own neural network that is trained as described herein, thus reducing the risk of false positives.

[0168] The vehicle 1100 may further include an infotainment SoC 1130 (e.g., an in-vehicle infotainment system (IVI)). Although shown and described as an SoC, the infotainment system may not be an SoC and may include two or more separate components. The infotainment SoC 1130 may include a combination of hardware and software that may be used to provide audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming, etc.), telephony (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.), and / or information services (e.g., navigation system, rear park assist, wireless data system, vehicle related information such as fuel level, total distance traveled, brake fuel level, oil level, door open / close, air filter information, etc.) to the vehicle 1100. For example, the infotainment SoC 1130 may be a radio, a disc player, a navigation system, a video player, USB and Bluetooth connectivity, a car computer, in-car entertainment, Wi-Fi, steering wheel audio controls, hands-free voice control, a heads-up display (HUD), an HMI display 1134, a telematics device, a control panel (e.g., for controlling and / or interacting with various components, features, and / or systems), and / or other components. The infotainment SoC 1130 may further be used to provide information (e.g., visual and / or audible) to a user of the vehicle, such as information from an ADAS system 1138, autonomous driving information such as planned vehicle maneuvers, trajectory, surrounding environment information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0169] The infotainment SoC 1130 may include GPU functionality. The infotainment SoC 1130 may communicate with other devices, systems, and / or components of the vehicle 1100 via the bus 1102 (e.g., CAN bus, Ethernet, etc.). In some instances, the infotainment SoC 1130 may be coupled to a supervisory MCU such that the infotainment system's GPU can perform some self-drive functions in the event that the primary controller 1136 (e.g., the vehicle's 1100 primary and / or backup computer) fails. In such instances, the infotainment SoC 1130 may place the vehicle 1100 in a Chauffeur safe shutdown mode, as described herein.

[0170] The vehicle 1100 may further include an instrument cluster 1132 (e.g., a digital dash, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 1132 may include a controller and / or a supercomputer (e.g., a separate controller or a supercomputer). The instrument cluster 1132 may include a set of instruments such as a speedometer, a fuel level, an oil pressure, a tachometer, an odometer, a turn signal, a gear shift position indicator, a seat belt warning light, a parking brake warning light, an engine malfunction light, an airbag (SRS) system information, lighting controls, safety system controls, navigation information, etc. In some instances, information may be displayed and / or shared between the infotainment SoC 1130 and the instrument cluster 1132. In other words, the instrument cluster 1132 may be included as part of the infotainment SoC 1130, or vice versa.

[0171] 11D is a system diagram of communication between a cloud-based server and an example autonomous vehicle 1100 of FIG. 11A in accordance with some embodiments of the present disclosure. The system 1176 may include a server 1178, a network 1190, and a vehicle including the vehicle 1100. The server 1178 may include multiple GPUs 1184(A)-1184(H) (collectively referred to herein as GPUs 1184), PCIe switches 1182(A)-1182(H) (collectively referred to herein as PCIe switches 1182), and / or CPUs 1180(A)-1180(B) (collectively referred to herein as CPUs 1180). The GPUs 1184, CPUs 1180, and PCIe switches may be interconnected with a high-speed interconnect, such as, but not limited to, an NVLink interface 1188 developed by NVIDIA and / or a PCIe connection 1186. In some instances, the GPUs 1184 are connected via an NVLink and / or NVSwitch SoC, and the GPUs 1184 and PCIe switch 1182 are connected via a PCIe interconnect. Although eight GPUs 1184, two CPUs 1180, and two PCIe switches are illustrated, this is not intended to be limiting. Depending on the embodiment, each server 1178 may include any number of GPUs 1184, CPUs 1180, and / or PCIe switches. For example, the servers 1178 may each include 8, 16, 32, and / or more GPUs 1184.

[0172] The server 1178 can receive image data from the vehicle over the network 1190, the image data representing images showing unexpected or changed road conditions, such as recently started road construction. The server 1178 can transmit the neural network 1192, the updated neural network 1192, and / or the map information 1194, including information about traffic and road conditions, to the vehicle over the network 1190. The updates to the map information 1194 can include updates to the HD map 1122, such as information about construction sites, potholes, detours, flooding, and / or other obstacles. In some instances, the neural network 1192, the updated neural network 1192, and / or the map information 1194 may have arisen from new training and / or experience represented in data received from any number of vehicles in the environment and / or based on training performed at a data center (e.g., using the server 1178 and / or other servers).

[0173] The server 1178 may be used to train a machine learning model (e.g., a neural network) based on the training data. The training data may be generated by the vehicle and / or generated in a simulation (e.g., using a game engine). In some instances, the training data is tagged (e.g., if the neural network benefits from supervised learning) and / or undergoes other pre-processing, while in other instances, the training data is not tagged and / or pre-processed (e.g., if the neural network does not require supervised learning). The training may be performed according to any one or more classes of machine learning techniques, including, but not limited to, the following classes: supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, federated learning, transfer learning, feature learning (including principal component and cluster analysis), multi-linear subspace learning, manifold learning, representation learning (including preliminary dictionary learning), rule-based machine learning, anomaly detection, and variations or combinations thereof. After the machine-learned model has been traced, the machine-learned model may be used by the vehicle (e.g., transmitted to the vehicle via network 1190) and / or the machine-learned model may be used by server 1178 to remotely monitor the vehicle.

[0174] In some instances, the server 1178 can receive data from the vehicles and apply the data to state-of-the-art real-time neural networks for real-time intelligent inference. The server 1178 can include deep learning supercomputers and / or dedicated AI computers powered by GPUs 1184, such as the DGX and DGX Station machines developed by NVIDIA. However, in some instances, the server 1178 can include a deep learning infrastructure that uses only CPU-powered data centers.

[0175] The deep learning infrastructure of the server 1178 may be capable of rapid real-time inference and may use that capability to evaluate and validate the health of the processor, software, and / or associated hardware within the vehicle 1100. For example, the deep learning infrastructure may receive periodic updates from the vehicle 1100, such as a sequence of images and / or objects that the vehicle 1100 has located within the sequence of images (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural networks to identify objects and compare them to objects identified by the vehicle 1100, and if the results do not match and the infrastructure concludes that the AI ​​within the vehicle 1100 is not functioning properly, the server 1178 may send a signal to the vehicle 1100 to infer control, notify the passenger, and command the failsafe computer of the vehicle 1100 to complete a safe parking maneuver.

[0176] For inference, the server 1178 may include a GPU 1184 and one or more programmable inference accelerators (e.g., NVIDIA's TensorRT). The combination of a GPU-powered server and inference acceleration can enable real-time responsiveness. In other instances, such as when less performance is required, servers powered by CPUs, FPGAs, and other processors may be used for inference.

[0177] Exemplary Computing Device 12 is a block diagram of an example computing device 1200 suitable for use in implementing some embodiments of the present disclosure. The computing device 1200 may include an interconnect system 1202 that indirectly or directly couples the following devices: memory 1204, one or more central processing units (CPUs) 1206, one or more graphic processing units (GPUs) 1208, a communication interface 1210, input / output (I / O) ports 1212, input / output components 1214, a power supply 1216, one or more presentation components 1218 (e.g., displays), and one or more logic units 1220. In at least one embodiment, the computing device 1200 may include one or more virtual machines (VMs) and / or any of its components may include virtual components (e.g., virtual hardware components). As non-limiting examples, one or more of GPUs 1208 may include one or more vGPUs, one or more of CPUs 1206 may include one or more vCPUs, and / or one or more of logical units 1220 may include one or more virtual logical units. As such, computing device 1200 may include discrete components (e.g., an entire GPU dedicated to computing device 1200), virtual components (e.g., a portion of a GPU dedicated to computing device 1200), or a combination thereof.

[0178] Although the various blocks in FIG. 12 are shown connected via interconnect system 1202 with lines, this is not intended to be limiting and is merely for clarity. For example, in some embodiments, presentation component 1218, such as a display device, may be considered an I / O component 1214 (e.g., where the display is a touch screen). As another example, CPU 1206 and / or GPU 1208 may include memory (e.g., memory 1204 may represent a storage device in addition to the memory of GPU 1208, CPU 1206, and / or other components). In other words, the computing devices of FIG. 12 are merely exemplary. Categories such as "workstation," "server," "laptop," "desktop," "tablet," "client device," "mobile device," "handheld device," "gaming machine," "electronic control unit (ECU)," "virtual reality system," and / or other device or system types are all intended to be within the scope of the computing devices of FIG. 12 and therefore are not to be differentiated.

[0179] The interconnect system 1202 may represent one or more links or buses, such as an address bus, a data bus, a control bus, or a combination thereof. The interconnect system 1202 may include one or more bus or link types, such as an industry standard architecture (ISA) bus, an extended industry standard architecture (EISA) bus, a video electronics standards association (VESA) bus, a peripheral component interconnect (PCI) bus, a peripheral component interconnect express (PCIe) bus, and / or another type of bus or link. In some embodiments, there are direct connections between components. As an example, the CPU 1206 may be directly connected to the memory 1204. Additionally, the CPU 1206 may be directly connected to the GPU 1208. When there is a direct or point-to-point connection between components, the interconnect system 1202 may include a PCIe link to implement the connection. In these instances, a PCI bus need not be included in the computing device 1200 .

[0180] Memory 1204 may include any of a variety of computer-readable media. Computer-readable media may be any available media that can be accessed by computing device 1200. Computer-readable media may include both volatile and nonvolatile media, and removable and non-removable media. By way of example, and not limitation, computer-readable media may comprise computer storage media and communication media.

[0181] Computer storage media may include both volatile and nonvolatile media and / or removable and non-removable media implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules, and / or other data types. For example, memory 1204 may store computer readable instructions (e.g., representing 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 memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that may be used to store the desired information and that may be accessed by computing device 1200. As used herein, computer storage media does not include the signals themselves.

[0182] Computer storage media may embody computer-readable instructions, data structures, program modules, and / or other data types in a modulated data signal such as a carrier wave or other transport mechanism and include any information delivery media. The term "modulated data signal" may refer to a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, computer storage media may include wired media, such as a wired network or direct-wired connection, and wireless media, such as acoustic, RF, infrared and other wireless media. Combinations of any of the above should also be included within the scope of computer-readable media.

[0183] The CPU 1206 may be configured to execute at least some of the computer readable instructions to control one or more components of the computing device 1200 to perform one or more of the methods and / or processes described herein. The CPU 1206 may include one or more (e.g., 1, 2, 4, 8, 28, 72, etc.) cores each capable of processing multiple software threads simultaneously. The CPU 1206 may include any type of processor, and may include different types of processors depending on the type of computing device 1200 implemented (e.g., a processor with fewer cores for a mobile device and a processor with more cores for a server). For example, depending on the type of computing device 1200, the processor may be an Advanced RISC Machines (ARM) processor implemented using Reduced Instruction Set Computing (RISC) or an x86 processor implemented using Complex Instruction Set Computing (CISC). The computing device 1200 may include one or more CPUs 1206 within one or more microprocessors or auxiliary coprocessors, such as computational coprocessors.

[0184] In addition to or in lieu of CPU 1206, GPU 1208 may be configured to execute at least some of the computer readable instructions to control one or more components of computing device 1200 to perform one or more of the methods and / or processes described herein. One or more of GPUs 1208 may be integrated GPUs (e.g., with one or more of CPUs 1206 and / or one or more of GPUs 1208 may be discrete GPUs. In an embodiment, one or more of GPUs 1208 may be coprocessors of one or more of CPUs 1206. GPU 1208 may be used by computing device 1200 to render graphics (e.g., 3D graphics) or perform general purpose computations. For example, GPU 1208 may be configured to perform general purpose computing on a GPU (GPGPU). A graphics processing unit (GPU) 1208 may be used for processing graphics data. The GPU 1208 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. The GPU 1208 may generate pixel data for an output image in response to rendering commands (e.g., rendering commands from the CPU 1206 received via a host interface). The GPU 1208 may include graphics memory, e.g., display memory, for storing the pixel data or any other suitable data, e.g., GPGPU data. The display memory may be included as part of the memory 1204. GPU 1208 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 may connect the GPUs via a switch (e.g., using NVSwitch). When coupled together, each GPU 1208 may generate pixel data or GPGPU data for a different portion of the output or for a different output (e.g., a first GPU for a first image and a second GPU for a second image). Each GPU may include its own memory or may share memory with the other GPUs.

[0185] In addition to or instead of the CPU 1206 and / or GPU 1208, the logic unit 1220 may be configured to execute at least some of the computer readable instructions to control one or more of the computing devices 1200 to perform one or more of the methods and / or processes described herein. In an embodiment, the CPU 1206, the GPU 1208, and / or the logic unit 1220 may execute any combination of the methods, processes, and / or portions thereof, either discretely or jointly. One or more of the logic units 1220 may be part of and / or integrated with one or more of the CPU 1206 and / or GPU 1208, and / or one or more of the logic units 1220 may be a discrete component to or otherwise external to the CPU 1206 and / or GPU 1208. In an embodiment, one or more of the logic units 1220 may be a co-processor of one or more of the CPU 1206 and / or GPU 1208.

[0186] Examples of logic unit 1220 include one or more processing cores and / or components thereof, such as a Tensor Core (TC), a Tensor Processing Unit (TPU), a Pixel Visual Core (PVC), a Vision Processing Unit (VPU), a Graphics Processing Cluster (GPC), a Texture Processing Cluster (TPC), a Streaming Multiprocessor (SM), a Tree Traversal Unit (TTU), an Artificial Intelligence Accelerator (AIA), a Deep Learning Accelerator (DLA), an Arithmetic Logic Unit (ALU), an Application Specific Integrated Circuit (ASIC), a Floating Point Unit (FPU), an Input / Output (I / O) element, a Peripheral Component Interconnect (PCI) or Peripheral Component Interconnect Express (PCIe) element, and / or the like.

[0187] Communications interface 1210 may include one or more receivers, transmitters, and / or transceivers that enable computing device 1200 to communicate with other computing devices over electronic communications networks, including wired and / or wireless communications. Communications interface 1210 may include components and functionality to enable communication over any of a number of different networks, such as a wireless network (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), a wired network (e.g., communicating over Ethernet or InfiniBand), a low power wide area network (e.g., LoRaWAN, SigFox, etc.), and / or the Internet.

[0188] The I / O ports 1212 may enable the computing device 1200 to be logically coupled to other devices, including I / O components 1214, presentation components 1218, and / or other components, some of which may be built-in (e.g., integrated) to the computing device 1200. Exemplary I / O components 1214 include a microphone, a mouse, a keyboard, a joystick, a game pad, a game controller, a satellite dish, a scanner, a printer, a wireless device, and the like. The I / O components 1214 may provide a natural user interface (NUI) that processes air gestures, voice, or other physiological input generated by a user. In some cases, the input may be sent to an appropriate network element for further processing. The NUI may implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, gesture recognition on and next to the screen, air gestures, head and eye tracking, and touch recognition in conjunction with the display of the computing device 1200 (as described in more detail below). The computing device 1200 may include a depth camera, such as a stereoscopic camera system, an infrared camera system, an RGB camera system, touch screen technology, and combinations thereof, for gesture detection and recognition. Additionally, the computing device 1200 may include an accelerometer or gyroscope (e.g., as part of an inertia measurement unit (IMU)) to enable detection of motion. In some instances, the output of the accelerometer or gyroscope may be used by the computing device 1200 to render immersive augmented or virtual reality.

[0189] The power supply 1216 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 1216 may provide power to the computing device 1200 to enable the components of the computing device 1200 to operate.

[0190] The presentation component 1218 may include a display (e.g., a monitor, a touch screen, a television screen, a heads-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The presentation component 1218 can receive data from other components (e.g., GPU 1208, CPU 1206, etc.) and output data (e.g., as images, video, sound, etc.).

[0191] Exemplary Data Center 13 illustrates an example data center 1300 that may be used in at least one embodiment of the present disclosure. The data center 1300 may include a data center infrastructure layer 1310, a framework layer 1320, a software layer 1330, and / or an application layer 1340.

[0192] 13, data center infrastructure layer 1310 may include a resource orchestrator 1312, grouped computational resources 1314, and node computational resources ("node CRs") 1316(1)-1316(N), where "N" represents any integer, natural number. In at least one embodiment, node CRs 1316(1)-1316(N) may include, but are not limited to, any number of central processing units ("CPUs") or other processors (including accelerators, field programmable gate arrays (FPGAs), graphics processors or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memory), storage devices (e.g., solid state or disk drives), network input / output ("NW I / O") devices, network switches, virtual machines ("VMs"), power modules, and / or cooling modules, etc. In some embodiments, one or more of the nodes CR 1316(1)-1316(N) may correspond to a server having one or more of the aforementioned computing resources. Additionally, in some embodiments, the nodes CR 1316(1)-1316(N) may include one or more virtual components, such as a vGPU, a vCPU, and / or the like, and / or one or more of the nodes CR 1316(1)-1316(N) may correspond to a virtual machine (VM).

[0193] In at least one embodiment, the grouped computing resources 1314 may include separate groups of nodes CR 1316 housed in one or more racks (not shown), or multiple racks housed in data centers in various geographic locations (also not shown). The separate groups of nodes CR 1316 in the grouped computing resources 1314 may include grouped computing, network, memory, or storage resources that may be configured or assigned to support one or more workloads. In at least one embodiment, several nodes CR 1316 including CPUs, GPUs, and / or other processors may be grouped in one or more racks to provide computing resources to support one or more workloads. The one or more racks may also include any number of power modules, cooling modules, and / or network switches in any combination.

[0194] The resource orchestrator 1322 may configure or otherwise control one or more nodes CR 1316(1)-1316(N) and / or grouped computational resources 1314. In at least one embodiment, the resource orchestrator 1322 may include a software design infrastructure ("SDI") management entity of the data center 1300. The resource orchestrator 1322 may include hardware, software, or some combination thereof.

[0195] In at least one embodiment, as shown in FIG. 13 , framework layer 1320 may include a job scheduler 1332, a configuration manager 1334, a resource manager 1336, and / or a distributed file system 1338. Framework layer 1320 may include frameworks to support software 1332 in software layer 1330 and / or one or more applications 1342 in application layer 1340. Software 1332 or applications 1342 may include web-based service software or applications, such as those offered by Amazon Web Services, Google Cloud, and Microsoft Azure, respectively. Framework layer 1320 may be a type of free and open source software web application framework, such as, but not limited to, Apache Spark™ (hereinafter “Spark”), which may use distributed file system 1338 for large-scale data processing (e.g., “big data”). In at least one embodiment, the job scheduler 1332 may include a Spark driver to facilitate scheduling of workloads supported by various tiers of the data center 1300. The configuration manager 1334 may be capable of configuring different tiers, for example, the software tier 1330 and the framework tier 1320 including Spark and a distributed file system 1338 to support large scale data processing. The resource manager 1336 may be capable of managing clustered or grouped computing resources that are mapped or assigned in support of the distributed file system 1338 and the job scheduler 1332. In at least one embodiment, the clustered or grouped computing resources may include the computing resources 1314 grouped in the data center infrastructure tier 1310. The resource manager 1336 may coordinate with the resource orchestrator 1312 to manage these mapped or assigned computing resources.

[0196] In at least one embodiment, software 1332 included in software layer 1330 may include software used by at least a portion of nodes CR 1316(1)-1316(N), grouped computational resources 1314, and / or distributed file system 1338 of framework layer 1320. The one or more types of software may include, but are not limited to, Internet web page searching software, email virus scanning software, database software, and streaming video content software.

[0197] In at least one embodiment, the applications 1342 included in the application layer 1340 may include one or more types of applications used by at least a portion of the nodes CRs 1316(1)-1316(N), the grouped computational resources 1314, and / or the distributed file system 1338 of the framework layer 1320. The one or more types of applications may include, but are not limited to, any number of 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.

[0198] In at least one embodiment, any of configuration manager 1334, resource manager 1336, and resource orchestrator 1312 can implement any number and type of self-rewriting actions based on any amount and type of data obtained in any technically possible manner. The self-rewriting actions can free data center operators of data center 1300 from making potentially poor configuration decisions and possibly avoiding underutilized and / or underperforming portions of the data center.

[0199] Data center 1300 may include tools, services, software, or other resources to train one or more machine learning models or to predict or infer information using one or more machine learning models according to one or more embodiments described herein. For example, a machine learning model may be trained by calculation of weight parameters by a neural network architecture using the software and / or computing resources described above with respect to data center 1300. In at least one embodiment, a trained or deployed machine learning model corresponding to one or more neural networks may be used to infer or predict information using the resources described above with respect to data center 1300, for example, by using weight parameters calculated via one or more training techniques, including but not limited to those described herein.

[0200] In at least one embodiment, data center 1300 may use CPUs, application specific integrated circuits (ASICs), GPUs, FPGAs, and / or other hardware (or corresponding virtual computing resources) for training and / or performing inference using such resources. Additionally, one or more of such software and / or hardware resources may be configured as services, such as image recognition, speech recognition, or other artificial intelligence services, to enable a user to train or perform inference on information.

[0201] Example Network Environment A network environment suitable for use in implementing embodiments of the present disclosure may include one or more client devices, servers, network attached storage (NAS), other back-end devices, and / or other device types. The client devices, servers, and / or other device types (e.g., each device) may be implemented with one or more instances of computing device 1200 of FIG. 12, e.g., each device may include similar components, features, and / or functionality of computing device 1200. In addition, if a back-end device (e.g., server, NAS, etc.) is implemented, the back-end device may be included as part of data center 1300, examples of which are further detailed herein with respect to FIG. 13.

[0202] Components of a network environment may communicate with each other via a network, which may be wired, wireless, or both. A network may include multiple networks or a network of networks. Illustratively, a network may 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. When a network includes a wireless telecommunications network, components such as base stations, communication towers, or access points (as well as other components) may provide wireless connectivity.

[0203] Compatible network environments may include one or more peer-to-peer network environments (wherein a server may not be included in the network environment) and one or more client-server network environments (wherein a server or servers may be included in the network environment). In a peer-to-peer network environment, the functionality described herein with respect to a server may be implemented in any number of client devices.

[0204] In at least one embodiment, the network environment may include one or more cloud-based network environments, distributed computing environments, combinations thereof, and the like. The cloud-based network environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more of the servers, which may include one or more core network servers and / or edge servers. The framework layer may include a framework to support software in the software layer and / or one or more applications in the application layer. The software or applications may include web-based service software or applications, respectively. In an embodiment, one or more of the client devices may use the web-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 web application framework that may use a distributed file system for large-scale data processing (e.g., "big data"), for example.

[0205] A cloud-based network environment may provide cloud computing and / or cloud storage performing any combination of the computing and / or data storage functions (or one or more portions thereof) described herein. Any of these various functions may be distributed across multiple locations from a central or core server (e.g., one or more data centers that may be distributed across a state, territory, country, or world). When a connection to a user (e.g., a client device) is relatively close to an edge server, the core server may delegate at least a portion of the functionality to the edge server. A cloud-based network environment may be private (e.g., restricted to a single organization), public (e.g., available to multiple organizations), and / or a combination thereof (e.g., a hybrid cloud environment).

[0206] A client device may include at least some of the components, features, and functionality of the exemplary computing device 1200 described herein with respect to Figure 12. By way of illustration, and not by way of limitation, a client device may be embodied as a personal computer (PC), a laptop computer, a mobile device, a smartphone, a tablet computer, a smart watch, a wearable computer, a personal digital assistant (PDA), an MP3 player, a virtual reality headset, a global positioning system (GPS) or device, a video player, a video camera, a surveillance device or system, a vehicle, a boat, an airship, a virtual machine, a drone, a robot, a handheld communications device, a hospital device, a gaming device or system, an entertainment system, a vehicle computer system, an embedded system controller, a remote control, an appliance, a consumer electronic device, a workstation, an edge device, any combination of these depicted devices, or any other suitable device.

[0207] The present disclosure may be described in the general context of computer code or machine usable instructions, including computer executable instructions, such as program modules, being executed by a computer or other machine, such as a personal digital assistant or other handheld device. Generally, program modules, including routines, programs, objects, components, data structures, etc., refer to code that performs particular tasks or implements particular abstract data types. The present disclosure may be implemented in a variety of configurations, including handheld devices, consumer electronics, general purpose computers, more specialized computing devices, etc. The present disclosure may also be implemented in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network.

[0208] As used herein, the statement "and / or" with respect to two or more elements should be interpreted to mean only one element or a combination of elements. For example, "element A, element B, and / or element C" may include element A only, element B only, element C only, elements A and element B, elements A and element C, elements B and element C, or elements A, B, and C. In addition, "at least one of element A or element B" may include at least one of element A, at least one of element B, or at least one of element A and at least one of element B. Furthermore, "at least one of element A and element B" may include at least one of element A, at least one of element B, or at least one of element A and at least one of element B.

[0209] The subject matter of the present disclosure has been described with specificity to meet statutory requirements. However, the description itself is not intended to limit the scope of the present disclosure. Instead, the inventors intend that the claimed subject matter may be implemented in other ways, including different steps or combinations of steps similar to those described in this document, in conjunction with other current or future technologies. Furthermore, although the terms "step" and / or "block" may be used herein to connote different elements of the method used, these terms should not be construed as implying any particular order between the various steps disclosed herein unless and when the order of the individual steps is explicitly described.

Claims

1. A method of computing, using a neural network, based at least in part on sensor data representing a sensor's perceptual field, one or more positions of one or more control points of one or more curves identified in an image represented by the sensor data and classification information for the one or more curves; generating the one or more curves based at least in part on the one or more positions of the one or more control points using a curve fitting algorithm; associating said classification information with said one or more curves; and performing, by a machine, one or more operations based at least in part on the classification information and the one or more curves.

2. 2. The method of claim 1 , wherein the calculating the one or more positions of the one or more control points comprises: calculating, using the neural network, first data indicative of one or more pixels of the image within a radius of at least one of the one or more curves identified in the image; calculating second data indicative of one or more distances of the one or more curves to the one or more control points; and decoding the first data and the second data to determine the one or more positions of the one or more control points.

3. The method of claim 1 , wherein the classification information represents a one or more bit classification mask.

4. 4. The method of claim 3, wherein the number of bit classification masks of the one or more bit classification masks is equal to one less than the number of classification types for which the neural network calculates outputs.

5. 3. The method of claim 2, wherein the first data represents a binary mask, in which the pixels of the image within the radius of the at least one of the one or more curves shown in the one or more images have a first value, and one or more other pixels other than the one or more pixels have a second value different from the first value.

6. 3. The method of claim 2, wherein each curve includes a number of control points greater than three, and the second data represents a number of regression masks, the number of regression masks including a number of x-axis distance regression masks and a number of y-axis distance regression masks equal to the number of control points.

7. 2. The method of claim 1, wherein the curve fitting algorithm comprises at least one of a two-dimensional Bezier curve fitting algorithm, a three-dimensional Bezier curve fitting algorithm, a cubic Bezier curve fitting algorithm, a higher order Bezier curve fitting algorithm, or a split Bezier curve fitting algorithm.

8. 2. The method of claim 1, further comprising running a clustering algorithm on the one or more curves to generate a final set of curves, wherein the performing of the one or more operations is based at least in part on the final set of curves and the classification information.

9. The method of claim 8 , wherein the clustering algorithm comprises a Density-Based Spatial Clustering of Noisy Applications (DBSCAN) operation using integral loss as a similarity function.

10. 2. The method of claim 1, wherein the one or more curves correspond to one or more of a road marking, a lane marking, a road boundary, an intersection line, a pedestrian path, a bike lane marking, text, a pole, a tree, a light post, or a sign.

11. A system comprising one or more sensors and one or more processing units, the one or more processing units calculating, based at least in part on sensor data representative of an image corresponding to a sensor's perception field, using a neural network, positions of one or more control points of one or more curves corresponding to one or more features identified in the image and classification information of the one or more curves; generating the one or more curves based at least in part on the positions of the one or more control points using a curve fitting algorithm; and performing one or more operations based at least in part on the one or more curves and the classification information.

12. The system of claim 11, wherein the classification information corresponds to one or more of a road marking, a lane line, a road boundary, an intersection line, a pedestrian path, a bike lane line, text, a pole, a tree, a light post, or a sign.

13. 12. The system of claim 11, wherein calculating positions of one or more control points of the one or more curves comprises: calculating a binary mask indicative of pixels of the one or more images within a radius of at least one of the one or more curves that correspond to one or more features identified in the image; calculating one or more distance regression masks indicative of distances to control points of the one or more curves; and determining the one or more positions of the one or more control points based on the binary mask and the one or more distance regression masks.

14. 12. The system of claim 11, wherein calculating classification information of the one or more curves comprises using the neural network to calculate a one or more bit classification mask indicative of classification information of the one or more curves based at least in part on the sensor data.

15. 15. The system of claim 14, wherein the number of bit classification masks of the one or more bit classification masks is equal to one less than the number of classification types for which the neural network computes its output.

16. 14. The system of claim 13, wherein each curve includes a number of control points, and the one or more distance regression masks include a number of x-axis distance regression masks and a y-axis distance regression masks equal to the number of control points.

17. 12. The system of claim 11, wherein the curve fitting algorithm comprises at least one of a two-dimensional Bezier curve fitting algorithm, a three-dimensional Bezier curve fitting algorithm, a cubic Bezier curve fitting algorithm, a higher order Bezier curve fitting algorithm, or a split Bezier curve fitting algorithm.

18. 12. The system of claim 11, further comprising: the one or more processing units running a clustering algorithm on the one or more curves to generate a final set of curves, and wherein the performing of the one or more operations is based at least in part on the final set of curves and the classification information.

19. 12. The system of claim 11, wherein the one or more processing units comprise at least one of a system for performing simulation operations, a system for performing simulation operations for testing or validating an autonomous machine application, a system for performing deep learning operations, a system implemented using edge devices, a system incorporating one or more virtual machines (VMs), a system implemented at least partially in a data center, or a system implemented at least partially using cloud computing resources.

20. 11. A method comprising: receiving annotation data representing a polyline associated with an image and classification information associated with the polyline; performing a curve fitting algorithm based at least in part on the polyline to determine a plurality of control points associated with the curve; encoding a binary mask representing pixels of the image that are within a radius to the curve based at least in part on the plurality of control points associated with the curve; encoding a plurality of distance regression masks, each distance regression mask indicating one or more distances to a control point of the plurality of control points; encoding one or more bit classification masks based at least in part on the classification information; and updating one or more parameters of a neural network using the binary mask, the plurality of distance regression masks, and the one or more bit classification masks as ground truth data.

21. 21. The method of claim 20, further comprising cropping at least one of an upper portion or a lower portion of the polyline to generate a cropped polyline, and wherein the performing of the curve fitting algorithm is based at least in part on the cropped polyline.

22. 21. The method of claim 20, further comprising: receiving initial annotation data representing a polygon; and extracting the polyline as one of a left edge, a right edge, or a center rail of the polygon based at least in part on the initial annotation data to generate the annotation data.

23. 21. The method of claim 20, wherein one or more parameters of the neural network are updated using one or more of an integral loss, a binary cross entropy loss, a masked L1 loss, or a masked binary cross entropy loss.

24. 21. The method of claim 20, wherein the encoding of the plurality of distance regression masks comprises encoding an x-axis distance regression mask and a y-axis distance regression mask for each control point of the plurality of control points.

25. 21. The method of claim 20, wherein the neural network is trained to predict a number, n, of classification types, and wherein the encoding the one or more bit classification masks comprises encoding n-1 bit classification masks.

26. 21. The method of claim 20, wherein the curve fitting algorithm comprises at least one of a two-dimensional Bezier curve fitting algorithm, a three-dimensional Bezier curve fitting algorithm, a cubic Bezier curve fitting algorithm, a higher order Bezier curve fitting algorithm, or a split Bezier curve fitting algorithm.

Images