Leverage obstacle and lane detection to determine lane assignments for objects in the environment

Abstract

The present invention discloses the use of obstacle and lane detection to determine lane assignments for objects in an environment. In various examples, object fences corresponding to objects detected by an ego vehicle can be used to determine the overlap of the object fence with lanes on a driving surface. A lane mask corresponding to the lanes on the driving surface can be generated, and the object fence can be compared to the lanes of the lane mask to determine the overlap. In the case where the object fence is located in more than one lane, a boundary scoring method can be used to determine the overlap rate of the boundary fence and thus the object with each lane. The overlap with one or more lanes for each object can be used to determine the lane assignment of the object, and the ego vehicle can use the lane assignment to determine a path or trajectory along the driving surface.

Description

Background Art

[0001] Autonomous driving systems and advanced driver assistance systems (ADAS) can utilize sensors such as cameras to perform various tasks, such as lane keeping, lane changing, lane assignment, camera calibration, and positioning. For example, in order for autonomous and ADAS systems to operate independently and effectively, a real-time or near real-time understanding of the vehicle's surroundings can be generated. This understanding can include information about the location of objects and / or obstacles in the environment relative to various demarcations (e.g., lanes, road boundaries, intersections, etc.). The vehicle can use information about the surrounding environment when making decisions, such as which path or trajectory to follow, whether to accelerate or decelerate, whether it is currently safe to change lanes or other maneuvers, etc.

[0002] In conventional systems, lane (or other location) assignments for objects (e.g., vehicles, pedestrians, etc.) may be interpolated from real-time images processed using various computer vision or machine learning techniques (e.g., from visual indicators identified via image processing). For example, an image of a captured driving surface including a lane and a vehicle may be analyzed to determine the correlation between the vehicle and the lane. However, because the entire vehicle-lane interaction may be used, the actual lane of travel of the vehicle may not be accurately identified in conventional systems. For example, in image space, and depending on the camera's viewpoint, a vehicle may appear to extend into an adjacent lane (e.g., when the camera's viewpoint is from the left side of the vehicle, the top of the vehicle may appear to extend into the vehicle's left adjacent lane), even if the vehicle is actually in a single lane. This may not provide the most accurate results for use by autonomous systems and / or ADAS systems when attempting to understand the surrounding environment, thereby reducing the overall effectiveness and reliability of this environmental information. Summary of the invention

[0003] Embodiments of the present disclosure relate to determining lane assignments for objects in an environment using obstacle and lane detection. Systems and methods are disclosed for effectively and accurately assigning objects to lanes or other defined portions of an environment using object detection, free space detection, object fence detection, and / or lane detection. Compared to conventional methods as described above, the current system provides a technique for detecting and assigning objects to lanes in real time or near real time using outputs from multiple sensors (e.g., cameras, RADAR sensors, LIDAR sensors, etc.) from a vehicle. Embodiments of the present disclosure may combine the outputs of various computer vision or machine learning algorithms for path detection, lane detection, and / or object or obstacle detection to assign objects to lanes around a vehicle. For example, the current system may learn to use sensor data from any number of sensors to detect pixels and corresponding lanes occupied by one or more objects in an image space. An object fence or region corresponding to a portion of an object actually occupying a lane may be used to determine the pixel occupancy of an object. For example, the footprint of an object may be determined so that only the occupied space is used to determine the space occupied by the object. By more accurately identifying the object fence or occupied space of the object, the process of determining the lane assignment can be relatively less computationally expensive because a reduced number of pixels (e.g., pixels of the object fence rather than pixels of the entire vehicle) can be analyzed to determine the lane assignment. In return, the reduction in computational expense can reduce the burden of determining the lane assignment of the object, thereby freeing up processing power for other advanced driver assistance systems (ADAS) or autonomous tasks while reducing run time. BRIEF DESCRIPTION OF THE DRAWINGS

[0004] The present system and method for determining lane assignments for objects in an environment using obstacle and lane detection is described in detail below with reference to the accompanying drawings, wherein:

[0005] Figure 1 is an example data flow diagram illustrating an example process for determining lane assignments for objects in an environment according to some embodiments of the present disclosure;

[0006] Figure 2A is an illustration of an example boundary shape prediction of a vehicle detected in an image generated by a vehicle sensor according to some embodiments of the present disclosure;

[0007] Figure 2B is an illustration of an example cropped boundary shape of a vehicle detected in an image according to some embodiments of the present disclosure;

[0008] Figure 2C is an illustration of an example drivable free space prediction in an environment according to some embodiments of the present disclosure;

[0009] Figure 2D is an illustration of an object fence for a vehicle generated using drivable free space prediction and boundary shape prediction according to some embodiments of the present disclosure;

[0010] Figure 3 is a flow chart illustrating an example process for determining lane assignments for objects in an environment according to some embodiments of the present disclosure;

[0011] Figure 4 is an illustration of an object fence for a vehicle generated by a fence generation algorithm according to some embodiments of the present disclosure;

[0012] Figure 5 is a flow chart illustrating an example process for determining lane assignments for objects in an environment according to some embodiments of the present disclosure;

[0013] Fig. 6A is a graphical representation of poly-points defining lanes generated by a lane detection algorithm according to some embodiments of the present disclosure;

[0014] Figure 6B is an illustration of a polyline connecting inflection points that define a lane generated by a lane detection algorithm according to some embodiments of the present disclosure;

[0015] Figure 6C is an illustration of performing lane triangulation on a detected lane to generate a lane mask according to some embodiments of the present disclosure;

[0016] Fig.6D is an illustration of a lane mask generated using lane data according to some embodiments of the present disclosure;

[0017] Fig. 6E is an illustration of performing lane extension on a lane mask to extend a lane according to some embodiments of the present disclosure;

[0018] Fig. 7A is an illustration of generating vertices for an object fence to determine overlap of the object fence with a lane of a lane mask according to some embodiments of the present disclosure;

[0019] Figure 7B is an illustration of generating vertices for an object fence to determine overlap of an object fence with a lane of a lane mask when the object fence crosses a lane boundary according to some embodiments of the present disclosure;

[0020] Figure 8is a flow chart illustrating an example process for determining a pixel overlap ratio between an object in an environment and two lanes according to some embodiments of the present disclosure.

[0021] Fig. 9A is an illustration of an example autonomous vehicle according to some embodiments of the present disclosure;

[0022] Fig. 9B According to some embodiments of the present disclosure, Fig. 9A Examples of autonomous vehicle camera positions and fields of view;

[0023] Fig. 9C According to some embodiments of the present disclosure, Fig. 9A a block diagram of an example system architecture for an example autonomous vehicle;

[0024] Fig.9D According to some embodiments of the present disclosure, a method for performing a multi-threaded communication between one or more cloud-based servers and a Fig. 9A An example system diagram of communicating between autonomous vehicles; and

[0025] Fig.10 is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0026] Systems and methods are disclosed for determining lane assignments for objects in an environment using obstacle and lane detection. The systems and methods described herein may be used in augmented reality, virtual reality, robotics, security and surveillance, medical imaging, autonomous machine applications, and / or any other technology space where object detection and assignment may be implemented. Although the examples may be described with respect to an example autonomous vehicle 900 (referred to herein alternatively as “vehicle 900” or “autonomous vehicle 900”), the examples may be described with respect to Figure 9A-9D The present disclosure may be described herein using the method of detecting an object or features such as a vehicle, a semi-autonomous vehicle (e.g., in one or more adaptive driver assistance systems (ADAS)), a robot, a warehouse vehicle, an off-road vehicle, a flying vessel, a ship, a shuttle, an emergency vehicle, a motorcycle, an electric or motorized bicycle, an airplane, a construction vehicle, an underwater vehicle, an unmanned aerial vehicle, and / or other vehicle types. In addition, although the detection described herein primarily relates to objects such as vehicles and / or pedestrians, the present disclosure is not intended to be limited to these detections. For non-limiting examples, the processes described herein may be used to detect any object or feature, such as, but not limited to, road signs, airplanes, boats, shuttles, bicycles, motorcycles, traffic lights, luggage, signs, utility poles, animals, trees, characters, letters, objects, products, obstacles, digital watermarks, and / or anatomical objects.

[0027] As described above, compared to conventional systems, the current system provides an accurate and computationally efficient method to detect objects and assign them to lanes in real time or near real time using outputs from multiple sensors of the vehicle (e.g., cameras, LIDAR sensors, RADAR sensors, SONAR sensors, etc.). For example, before lane assignment, an object fence or occupied space can be determined. In some embodiments, a combination of inputs can be used to determine the object fence. For example, a machine learning model (e.g., a convolutional neural network (CNN) trained to detect objects) and / or a computer vision object detection algorithm can be used to generate an object detection input. The object detection input can include a boundary shape corresponding to the object, such as a box. In some examples, in a non-limiting example, at least a portion of the boundary shape can be cropped, such as to reduce the boundary shape to the bottom fifth of the shape. Another input can include a drivable free space input, which can be similarly generated using a machine learning model and / or a computer vision algorithm. The drivable free space input can be used to crop (or further crop) or cut out (carve out) a portion of the boundary shape corresponding to the drivable free space. As a result, the remaining portion of the bounding shape (e.g., the bounding shape with the cut-out drivable free space) can be referred to as an object fence or footprint within a two-dimensional (2D) image space. This process can be performed for each object in the field of view of the ego-vehicle or the sensory field of the ego-vehicle's sensor. Using this method, false positive lane assignments can be reduced because the top or upper portion of the object may be less likely to actually intersect the lane in world space (although it may appear to intersect the lane in image space, which is a drawback of conventional methods).

[0028] In some examples, the object fence can be generated using a fence detection algorithm that is specifically programmed (or trained in examples where the algorithm is a machine learning model) to detect objects and generate fences for the objects. For example, sensor data from any number of sensors (e.g., cameras, RADAR sensors, LIDAR sensors, SONAR sensors, etc.) can be used as input to the fence detection algorithm, and the fence detection algorithm can generate an object fence in a world space (e.g., a three-dimensional world space). The object fence can be converted from a three-dimensional (3D) world space to a 2D image space using intrinsic and / or extrinsic parameters of the sensor that generated the sensor data.

[0029] In some embodiments, in addition to the object fence, sensor data (e.g., LIDAR data, SONAR data, image data, RADAR data, etc.) can be used to determine the future position of an object in the environment. For example, once the object fence is determined, the sensor data can be used to determine the path or trajectory of the object in the environment to determine one or more future positions (e.g., 0.5 seconds in the future, 1 second in the future, etc.), and the object fence can be applied at the future position. In this way, using the current position of the object and the speed and / or acceleration data corresponding to the object, one or more algorithms and / or machine learning models can be used to determine the future position of the object, thereby determining the position of the corresponding object fence of the object at the future position. In some examples, a scaling factor can be used as a change in the future position of the object relative to the vehicle position (e.g., as the object moves farther, the object fence can be reduced from the current size, as the object moves closer due to deceleration, for example, the size of the object fence may increase, and so on). This information can then be used to inform the vehicle of the lane or other parts of the environment where the object may be located in the future to help with trajectory or path planning, obstacle avoidance, and / or other operations of the vehicle.

[0030] The object fence can then be used in conjunction with the lane data to determine the overlap of the object fence with the lane in the environment (e.g., the driving surface) to determine one or more lanes associated with each object. The lane data can be received in the form of a lane map and / or a map (e.g., a navigation-level map, a high-definition (HD) map, etc.). The lane data can be used to generate a lane mask representing pixels corresponding to each lane in a 2D image space. In some examples, the lane map, the map, and / or a combination thereof can be used to determine the number and location of lanes in the environment. For example, a lane mask algorithm can be used that receives lane data as input and can generate a lane mask. In an embodiment, a lane can be defined by a polyline (e.g., including points along a line that demarcates a lane). Thus, in order to generate a mask from a polyline, a triangulation method can be used in some embodiments. The triangulation method can connect points along a polyline (e.g., for each lane, connect every other point in a crisscross manner from one polyline to another polyline). As a non-limiting example, the triangulation method can include monotone polygon triangulation. The results of the triangulation may include polygons in the 2D image space, each polygon corresponding to a lane, such that each pixel within the polygon for a given lane is determined to correspond to that lane.

[0031] In some examples, the full or available extension of a lane may not be determined during lane mask generation. In such examples, a lane extension algorithm (e.g., curve fitting, lane extrapolation, etc.) may be used to extend the lane. By extending the lane, false positives may be reduced. In other examples, there may be no lane detection data for the system to use. In such examples, virtual lanes may be generated. For example, in some embodiments, at least three lanes may be generated, including an ego lane (e.g., the lane of the ego vehicle) and adjacent lanes on either side. By generating virtual lanes, objects may be assigned to the virtual lanes, even if not completely accurate, to enable the ego vehicle to better understand the location of objects relative to the path of the ego vehicle (e.g., the ego lane).

[0032] Once the object fence and lane mask are determined, the object can be associated with the lane by determining the overlap between the pixels of the object fence and the lane mask in the 2D image space. The object can be assigned to one or more lanes in the lane mask based on the overlap of the pixels in the object fence with the lane pixels in the lane mask. In some examples, each object can be assigned to one or more lanes based on a simple pixel count. In other examples, the object fence can be represented by vertices (or pixels corresponding thereto) along the perimeter of the object fence. In the case where the object is located in a single lane, the pixel distance between each group of two surrounding pixels can be calculated, accumulated and normalized to create a per-lane intersection ratio, which is 1 / 1 for a single lane. In the case where the object is located in more than one lane, a set of points that intersect the lane boundary can be determined. Once two pixels are determined, a new vertex can be generated at the intersection between the two pixels. The pixel can then be used to determine the distance between the new vertex and each other surrounding pixel or vertex on either side of the intersection. A first sum of distances between the new vertex for the first lane and the first set of surrounding pixels may be calculated, and a second sum of distances between the new vertex for the second lane and the second set of surrounding pixels may be calculated. In some embodiments, these sums may then be normalized. Ultimately, the first sum and the second sum may be used to determine a per-lane intersection ratio. Knowing the intersection ratio and how it changes from frame to frame may provide an indication of the trajectory of the object (e.g., switching lanes, turning, keeping lanes, etc.).

[0033] In some embodiments, temporal smoothing may be used to leverage previous lane assignments of previous frames of an object. For example, because lane assignments may be more accurate as an object is closer to the vehicle, previous lane assignments may be leveraged to provide a more accurate prediction of the object's lane assignment at greater distances as the vehicle moves farther away, where lanes tend to appear to merge in 2D image space.

[0034] Ultimately, data representing lane assignments for objects may then be compiled and sent to a perception layer, a world model management layer, a planning layer, a control layer, and / or another layer of the autonomous driving software stack to assist autonomous vehicles and / or vehicles with ADAS capabilities in safely and efficiently navigating driving surfaces.

[0035] refer to Figure 1 , Figure 1 is an example data flow diagram illustrating an example process 100 for determining lane assignments for objects in an environment according to some embodiments of the present disclosure. Figure 1 The type of detection described is lane detection, but this is for example purposes only and is not intended to be limiting. For example, process 100 may be used to detect paths, tracks, rows, lines, and / or other portions of an environment and assign objects thereto without departing from the scope of the present disclosure.

[0036] Process 100 may include generating and / or receiving sensor data 102 from one or more sensors of vehicle 900. Sensor data 102 may be used by vehicle 900 and within process 100 to assign objects to lanes or other defined areas of an environment in real time or near real time. Sensor data 102 may include, but is not limited to, sensor data 102 from any sensor of vehicle 900 (and / or other vehicles or objects, such as robotic devices, VR systems, AR systems, etc., in some examples). For example, and with reference to Figure 9A-9C , the sensor data 102 may include data generated by, but not limited to, one or more global navigation satellite system (GNSS) sensors 958 (e.g., one or more global positioning system sensors), one or more RADAR sensors 960, one or more ultrasonic sensors 962, one or more LIDAR sensors 964, one or more inertial measurement unit (IMU) sensors 966 (e.g., one or more accelerometers, one or more gyroscopes, one or more magnetic compasses, one or more magnetometers, etc.), one or more microphones 996, one or more stereo cameras 968, one or more wide-angle cameras 970 (e.g., fisheye cameras), one or more infrared cameras 972, one or more surround cameras 974 (e.g., 360-degree cameras), one or more long-range and / or mid-range cameras 998, one or more speed sensors 944 (e.g., for measuring the speed of the vehicle 900), and / or other sensor types.

[0037] In some examples, sensor data 102 may include sensor data generated by one or more forward-facing sensors (e.g., cameras (e.g., one or more center- or near-center-mounted cameras) and / or other sensor types (e.g., a RADAR sensor mounted on a bumper)). This sensor data 102 may be useful for computer vision and / or perception when navigating (e.g., within a lane, through a lane change, through a turn, through an intersection, etc.) because the forward-facing camera may include a field of view that includes both the current lane of travel of the vehicle 900 and the boundaries of one or more adjacent lanes of travel and / or driving surfaces of the vehicle 900 (e.g., Fig. 9B In some examples, side view sensors and / or rear view sensors may be employed in addition to forward-facing sensors, such as where lane assignments are to include objects to the side of vehicle 900 and / or to the rear of vehicle 900. In some examples, more than one camera or other sensor (e.g., LIDAR sensor 964, RADAR sensor 960, etc.) may be used to combine multiple fields of view (e.g., Fig. 9B The field of view of the remote camera 898, the forward stereo camera 868 and / or the forward wide-angle camera 870), and / or the sensing field (e.g., of the LIDAR sensor 964, the RADAR sensor 960, etc.).

[0038] The sensor data 102 may include image data representing one or more images, image data representing a video (e.g., a snapshot of a video), and / or sensor data representing a representation of a sensing field of a sensor (e.g., a depth map for a LIDAR sensor, a value map for an ultrasonic sensor, etc.). Where the sensor data 102 includes image data, any type of image data format may be used, such as, but not limited to, compressed images in Joint Photographic Experts Group (JPEG) or Luma / Chro (YUV) formats, compressed images derived from frames of a compressed video format such as H.264 / Advanced Video Coding (AVC) or H.265 / High Efficiency Video Coding (HEVC), such as raw images derived from red-transparent-blue (RCCB), red-transparent-blue (RCCC), or other types of imaging sensors, and / or other formats. Additionally, in some examples, sensor data 102 may be used within process 100 without any preprocessing (e.g., in a raw or captured format), while in other examples, sensor data 102 may be preprocessed (e.g., noise balancing, demosaicing, scaling, cropping, enhancement, white balancing, tone curve adjustment, etc., such as using a sensor data preprocessor (not shown)). As used herein, sensor data 102 may refer to unprocessed sensor data, preprocessed sensor data, or a combination thereof.

[0039] The sensor data 102 may be used to generate an object fence 110 for one or more objects in the environment captured by the sensors of the vehicle 900. The object fence 110 may be generated using any of a variety of different methods, such as, but not limited to, those described herein. In some examples, the object fence 110 may be generated using a combination of free space detection 106 and object detection 104, while in other examples, the object fence 110 may be generated using a fence generation algorithm 108. In either example, the object fence 110 may be used as an input to lane detection and allocation within the process 100.

[0040] In embodiments where object detection 104 and free space detection 106 are used to generate object fence 110, sensor data 102 may be used as input to object detection 104 and free space detection 106. For example, object detection 104 may use sensor data 102 (e.g., image data from one or more cameras, LIDAR data from one or more LIDAR sensors, RADAR data from one or more RADAR sensors, etc.) to detect objects (e.g., vehicles, pedestrians, bicycles, debris, etc.) and generate bounding shapes (e.g., bounding boxes, circles, polygons, etc.) for the detected objects. Object detection 104 may use a machine learning model (e.g., a deep neural network (DNN) such as a convolutional neural network (CNN)), a computer vision algorithm, and / or another type of object detection algorithm to detect objects and generate bounding shapes for the objects. In the case where a machine learning model is used for object detection 104, the machine learning model may include, but is not limited to, one or more machine learning models using linear regression, logistic regression, decision trees, support vector machines (SVM), naive Bayes, k-nearest neighbors (Knn), K-means clustering, random forests, dimensionality reduction algorithms, gradient boosting algorithms, neural networks (e.g., autoencoders, convolutions, loops, perceptrons, long / short term memory (LSTM), Hopfield, Boltzmann, deep belief, deconvolution, convolution, generative adversarial, liquid machines, etc.) and / or other types of machine learning models.

[0041] The bounding shapes output using object detection 104 may represent portions of one or more images (represented by sensor data 102) that correspond to objects (e.g., on the driving surface of vehicle 900). Figure 2A 200, a bounding shape 204 may be generated for a vehicle 202 using object detection 104 to identify a portion of an image corresponding to the vehicle 202. In a non-limiting example, the output of object detection 104 may include points (e.g., pixels) in one or more images where an object is determined to be located. In such an example, the points may include each point or each pixel within the bounding shape. In addition to the object mask generated within the bounding shape, the bounding shape may also define each pixel corresponding to the object, and this process of object detection 104 may be performed for any number of objects in each frame.

[0042] In some embodiments, at least a portion of the bounding shape of each detected object may be cropped in order to determine the object fence 110. For example, because at least some of the objects within the bounding shape may not correspond to the space occupied by the object on the driving surface (e.g., in 2D image space, the upper portion of the vehicle 202 may extend beyond the roadway), at least a portion of the bounding shape may be cropped. Figure 2B As shown in visualization 220 of the vehicle 202, the boundary shape 204 can be cropped to generate a cropped boundary shape 222 corresponding to the vehicle 202. In some examples, a predetermined portion, a predetermined amount, or a predetermined percentage of the boundary shape can be cropped, such as, but not limited to, the upper half, the upper three quarters, the upper four fifths, the upper two quarters, and the bottom quarter of the boundary shape, etc. By cropping at least a portion of the boundary shape, false negatives can be reduced because the resulting cropped boundary shape more accurately corresponds to the footprint of the vehicle 202 on the driving surface. In addition to the accuracy advantage relative to the object footprint, cropping a portion of the boundary shape can also reduce the computational cost for determining the object fence 110 (e.g., by reducing the effective area for applying free space detection 106) and / or the computational cost of using the object fence 110 for lane assignment (e.g., by reducing the number of pixels in the cropped portion of the boundary shape that can be used in the calculation).

[0043] As described herein, the output of free space detection 106 can be used in conjunction with the bounding shapes and / or clipped bounding shapes generated during object detection 104. For example, sensor data 102 can be used by one or more computer vision algorithms, machine learning models, and / or other detection algorithms for drivable free space to determine one or more fields of view and / or portions of the environment within one or more sensing fields of sensors of vehicle 900. Machine learning models that can be used for free space detection 106 may include, but are not limited to, those described herein (e.g., with respect to object detection 104 above). As a non-limiting example, drivable free space can be determined using the methods described in U.S. Non-Provisional Application No. 16 / 355,328, filed on March 15, 2019, the entire contents of which are incorporated herein by reference.

[0044] The output of free space detection 106 may represent a portion of the environment (represented within an image captured by one or more cameras of vehicle 900) that corresponds to drivable free space and / or non-drivable space on a driving surface. Drivable free space may represent portions of the environment that may be traversed or driven by vehicle 900 (e.g., roads may be drivable, while sidewalks or buildings may be non-drivable) as well as non-drivable space (e.g., buildings, parks, etc.). Visualization 230 of FIG2C may represent at least a portion of free space detection 106 applied to a clipped boundary shape 222. For example, a portion of drivable free space 232 determined during free space detection 106 may be clipped out of clipped boundary shape 222 to generate object fence 110 (e.g., Figure 2D Thus, Figure 2C As shown, after the drivable free space 232 is cropped out, the object fence 110 may represent the non-drivable space 234 within the cropped boundary shape 222. In the example using the complete boundary shape 204, the drivable free space within the complete boundary shape 204 may be cropped out. In some examples, the free space detection 106 may include points (e.g., pixels) in the image where the drivable free space and / or the non-drivable space are detected, and these points may be used to perform cropping of the boundary shape and / or generation of the object fence 110. In such an example, the points corresponding to the drivable free space may be removed from the points corresponding to the boundary shape (or a cropped version thereof) to generate the object fence 110. The cropping process may be performed for each object detected in one or more input images representing the field of view or sensing field of the sensor of the vehicle 900.

[0045] In some embodiments, in addition to or in lieu of using object detection 104 and free space detection 106, a fence generation algorithm 108 may be used to generate object fences 110. Fence generation algorithm 108 may include a machine learning model, a computer vision algorithm, and / or an object detection algorithm. Where a machine learning model is used, the machine learning model may include, but is not limited to, the machine learning models described herein (e.g., above with respect to object detection 104). Fence generation algorithm 108 may be trained and / or programmed to use sensor data 102 as input to detect objects and generate object fences 110 for each detected object (e.g., as described in the example of FIG. 1 ). Figure 2D). For example, sensor data 102 from any number of sensors (e.g., cameras, RADAR sensors 960, LIDAR sensors 964, ultrasonic sensors 962, etc.) may be used as input to the fence generation algorithm 108. In some embodiments, the fence generation algorithm 108 may generate an object fence 110 in a 2D or 3D world space, and the intrinsic and / or extrinsic parameters of the corresponding sensor that generated the sensor data 102 may be used to transform the object fence 110 from the world space to the image space. For example, in order to accurately determine the relationship between world space coordinates and image coordinates, a 3D to 2D projection may be used. For example, one or more intrinsic (e.g., focal length f, optical center (u0, v0), pixel aspect ratio α, skew s, etc.) and / or extrinsic (e.g., 3D rotation R, translation t, etc.) camera parameters may be used to calibrate the sensor. In addition, in some non-limiting examples, one or more constraints may also be imposed, such as requiring that the 3D point always be on the ground plane of the driving surface. In some examples, one or more parameters of one or more sensors may be dynamic (eg, due to vibration, motion, orientation, etc.), and the results may dynamically update the 3D to 2D projection.

[0046] In some embodiments, one or more object fences 110 may be generated as points (e.g., pixels and / or vertices) corresponding to a polygon of the one or more object fences 110. Thus, by connecting the vertices, a polygon corresponding to the one or more object fences 110 may be generated, and pixels within the one or more object fences 110 may be determined to correspond to an object for lane assignment. One or more object fences 110 may be generated (using any of the methods described herein) prior to lane assignment 120 to improve the accuracy and reliability of lane assignment predictions by more closely defining the shape or footprint of an object on the driving surface. For example, once the object fences 110 are determined, the object fences 110 may be used in conjunction with the lane mask 116 to make an overlap determination 118 for lane assignment 120.

[0047] In some embodiments, sensor data 102 (e.g., LIDAR data, SONAR data, image data, RADAR data, etc.) and object fence 110 may be used to determine future locations and / or assignments of objects in the environment. For example, once object fence 110 is determined, sensor data 102 representing velocity, speed, acceleration, yaw rate, etc. may be used to determine a future path or trajectory of an object in the environment to determine one or more future locations (e.g., 0.5 seconds in the future, 1 second in the future, etc.). For example, a linear dynamic model (e.g., future location = current location + (velocity × time)), a constant acceleration model, a constant turn rate model, a machine learning model, another algorithm type, and / or a combination thereof may use sensor data (or data represented thereby) to determine one or more future locations of an object (e.g., in world space). Once one or more future locations are known, the future locations and object fence 110 information may be used to generate object fence 110 (e.g., in image space). In some examples, the scaling factor may be used as a measure of future changes in the position of the object relative to the position of the vehicle 900 (e.g., as the object moves farther away, the object fence 110 may decrease from its current size, as the object moves closer due to deceleration, for example, the size of the object fence 110 may increase, and so on. This information may then be used to inform the vehicle 900 of lanes or other portions of the environment where the object may be located at future times to aid in trajectory or path planning, obstacle avoidance, and / or other operation of the vehicle. The determination of lane assignments 120 may be similar for future object fence locations as for current object fence locations, and the combination of the vehicle's current and future object fences 110 may provide a more detailed understanding of the area or volume within the environment that the object is expected to occupy over time (e.g., the next second, two seconds, etc.).

[0048] Lane detection 112 may use sensor data 102 and / or lane data 126 to determine the location of one or more lanes, divisions, segments, and / or boundaries of a driving surface. For example, lane detection 112 may use computer vision algorithms, machine learning models, and / or object detection algorithms to determine the location of lanes (or lane lines), divisions, segments, boundaries, and / or other areas of a driving surface. Where a machine learning model is used, the machine learning model may include, but is not limited to, the machine learning models described herein (e.g., above with respect to object detection 104). As non-limiting examples, each of the methods described in U.S. non-provisional application No. 16 / 286,329, filed on February 15, 2019, and U.S. non-provisional application No. 16 / 514,404, filed on July 17, 2019, are incorporated herein by reference in their entirety.

[0049] In some examples, lane detection 112 may use lane data 226, where lane data 226 may include one or more lane graphs and / or maps (e.g., navigation-grade maps, high-definition (HD) maps). For example, lane data 226 may represent multiple lanes on a driving surface and / or exit ramps, merges, splits, and / or other features in the environment. In this way, lane detection 112 may be generated using sensor data 102, lane data 226, or a combination thereof. For example, sensor data 102 may be used in conjunction with a lane detection algorithm or machine learning model to identify one or more lanes, and lane data 226 (e.g., from a map application) may be used to identify one or more other lanes. The output of the lane detection may include lines corresponding to lane lines, boundary lines, merge lines, split lines, and / or other lines of the driving surface (e.g., polylines represented by multiple vertices).

[0050] In a non-limiting example, the output of lane detection 112 may include one or more lane boundaries in the image (e.g., Fig. 6A The fold points 602A-602H and 604A-604F) of the visualization 600 are determined as points (eg, pixels) located therein. Figure 6B The polylines 622A-622G and 624A-624E of the visualization 620 can be generated by connecting adjacent polylines of various lanes and / or boundaries, where the connection between the polylines can define the extent of the lane lines or the boundaries of the image. In this way, the polylines and / or polylines can define or outline detected lanes, ramps, ramps, boundaries, and / or other features in the environment represented by the sensor data 102 (e.g., image) and / or map data 226 (e.g., lane map and / or map).

[0051] To determine lane mask 116, in some embodiments, lane triangulation 114 may be performed. For example, using the output of lane detection 112 (e.g., the fold points corresponding to the fold lines), lane triangulation 114 may be used to generate lane mask 116. In such an example, lane triangulation 114 may connect the fold points between adjacent fold lines in a pattern (e.g., a cross pattern). In a non-limiting example, as Figure 6C As shown, each vertex can be connected to the vertex of the adjacent polyline corresponding to the same line in a cross-wise manner (e.g., the vertex from the left lane line can be connected to the vertex from the right lane line). The first vertex of the first polyline (e.g., the bottom-most vertex) can be connected to the first (or second) vertex of the second polyline, and the first (or second) vertex of the second polyline can be connected to the second vertex of the first polyline, and so on, until each vertex is connected to form a triangle, as shown in FIG. Figure 6C640. In some non-limiting examples, lane triangulation 114 may use monotone polygon triangulation to generate lane mask 116. Polylines may be assumed to be monotone and non-curved in at least one axis (e.g., direction). Vectors may be determined between end polyline points and start polyline points in respective lanes. The vectors may then be used to generate triangles (e.g., Fig.6D The lane mask 116 may be rendered using triangles 642A-642X) to determine pixels within the image corresponding to each lane, on-ramp, off-ramp, or other defined portion of the driving surface.

[0052] In some embodiments, a lane regression algorithm may be used for lane triangulation 114 to generate lane mask 116. The lane regression algorithm may regress a line through a polyline to generate triangles. For example, any known regression algorithm may be used to generate triangles. Lane mask 116 may be generated using the triangles generated for each lane by lane triangulation 116. The combination of rendered triangles may form polygons in image space, where each polygon may correspond to a lane.

[0053] As a result of lane triangulation 114 and / or another mask generation method, a lane mask 116 may be generated to represent pixels in image space that correspond to lanes or other portions of the driving surface. In some examples, each pixel in lane mask 116 may be indexed as a color corresponding to a corresponding lane (e.g., relative to the lane color). Fig.6D , pixels of the left lane 662 may be a first color or pattern, pixels of the middle or ego lane 664 may be a second color or pattern, and pixels of the third lane 666 may be a third color or pattern).

[0054] In an embodiment, the scale factor may be used to scale down the lane mask 116. For example, the lane mask may be scaled down by a predetermined scale factor (e.g., half, quarter, etc.) to improve performance or reduce runtime. Even after scaling, the lane mask 116 may retain the fidelity of one or more source images, one or more maps, and / or one or more graphics used to generate the lane mask 116 because, in an embodiment, a lookup may be performed based on the scale so that each pixel in the sensor data 102 may be scaled down by the scale factor, thereby retaining information of higher spatial resolution even at a lower scaled down resolution. In such an example, the sensor data 102 may be scaled down prior to the lane triangulation 114, and the sensor data 102 (e.g., an image) may be scaled down using a rasterizer. The scaled down sensor data 102 may then be used for the lane triangulation 114. In such an embodiment, the object fence 110 (e.g., pixels of the object fence 110) may also be scaled down by the same scale factor.

[0055] In some examples, lane extension 122 and / or virtual lane generation 124 may be used to extend detected lanes and / or generate lanes in lane mask 116, e.g., as a result of lane detection 112 and / or lane triangulation 114, where a lane of full or usable extent is not generated for lane mask 116. A predetermined length of a lane may be used to determine whether lane extension 122 may be performed to extend one or more lanes. In such examples, a lane extension algorithm or machine learning model (e.g., those described herein) may be used to extend one or more lanes. Lane extensions (e.g., Fig. 6E 682, 684, 686, and 688 in visualization 680 of FIG. 4 ). Such lane extensions may reduce false positives in lane assignments for detected objects by generating more depth for lane mask 116. Additionally, in some examples, the lane extension algorithm may be configured to limit lane extensions when the extension intersects another lane boundary.

[0056] Virtual lane generation 124 may be performed when one or more lanes are not detected (e.g., below a threshold number of lanes). Thus, virtual lane generation 124 may be used when there is limited lane data 126 and / or sensor data 102 for lane detection 112. In an embodiment, a predetermined number of lanes may be used to determine whether a virtual lane should be generated. For example, where the threshold is three lanes, when three lanes are not detected, a virtual lane may be generated to fill the gap. In such an example, the three lanes may include at least the ego lane of the vehicle 900, and adjacent lanes on either side of the ego lane. Similar to lane extension 122, an algorithm or machine learning model may be used to determine the number of virtual lanes to be generated and the locations of the virtual lanes to be generated. By generating virtual lanes, objects may be assigned to virtual lanes, even if not as accurate as actual detection, to provide the vehicle 900 with a better understanding of the location of the object relative to the path of the ego vehicle.

[0057] Once one or more object fences 110 and lane masks 116 are generated, an overlap determination 118 may be performed to determine which object fences or portions thereof are within each lane or other defined portion of the environment. Thus, the overlap determination 118 may use the lane masks 116 and the object fences 110 to determine overlap between pixels and / or boundaries of the object fences 110 and the lane masks 116 (e.g., in 2D image space).

[0058] In some examples, objects may be assigned to one or more lanes in the lane mask 116 during lane assignment 120 based on an overlap of pixels in the object fence 110 with lane pixels in the lane mask 116. Pixel counts 132 may be used to determine overlap between pixels of the object fence 110 and the lane mask 116 so that each object with a corresponding object fence 110 may be assigned to one or more lanes based on the pixel counts. Pixels throughout the object fence 110 may be matched to lane pixels for each lane in the lane mask 116, and objects may be assigned to one or more lanes based on the pixel counts of the object fence 110 pixels in each lane. Then, pixel indices (e.g., colors or patterns, such as those associated with lane masks 116) associated with the lane masks 116 may be used to determine overlap between pixels of the object fence 110 and the lane mask 116. Fig.6D ) can be used to determine what lane a particular pixel overlaps with. In some examples, a number of overlapping pixels greater than a threshold may be required (e.g., greater than twenty pixels, greater than forty pixels, greater than ten percent of the pixels of the object fence 110, etc.) in order to associate an object with a lane. By using a threshold, false positives can be reduced because an object fence 110 that slightly overlaps a lane can be considered unimportant (which can explain some inaccuracies in the object fence 110).

[0059] In some embodiments, in addition to or in lieu of pixel counting 132, boundary scoring 134 may be performed to determine overlap between pixels of object fence 110 and lane mask 116. In such an example, object fence 110 may be represented by vertices (e.g., pixels) along the perimeter of object fence 110 (e.g., vertices 702A, 702B, 702C, ..., 702N). Pixel distances between vertices of object fence 110 may be determined to determine overlap. For example, in the case where the object is located in a single lane, the pixel distances between each set of two peripheral pixels may be calculated, accumulated, and normalized to create a per-lane intersection ratio. For a single lane, the ratio is 1 / 1.

[0060] In an example where an object may be in more than one lane, a set of points that intersect the lane boundary may be determined. For example, it may be determined that vertex 702B and vertex 702C are separated along a lane line. Thus, intersection 722A (or a new vertex) may be determined by finding a point between vertices 702B and 702C on the lane line. Once the intersection or new vertex can be used to determine the distance between the new vertex and the surrounding pixels of each other or the vertices of the object fence 110 on either side of the intersection (e.g., along a straight line, along the pixel distance of the boundary of the object fence 100, etc.). A first sum of the distances between the new vertex in the first lane and the first group of surrounding pixels corresponding to the object fence 110 may be calculated, and a second sum of the distances between the new vertex in the second lane and the second group of surrounding pixels corresponding to the object fence 110 may be calculated. In some embodiments, these sums may then be normalized. Ultimately, a per-lane intersection ratio may be determined based on the first sum of the distances and the second sum of the distances. Knowing the intersection ratio and how the ratio changes from frame to frame may provide an indication of the trajectory of the object (e.g., switching lanes, turning, keeping lanes, etc.). Additionally, during lane assignment 120 , intersection ratios may be used to assign objects to multiple lanes.

[0061] Overlap determination 118 can be used to determine lane assignment 120 for each object having a respective object fence 110. In some examples, pixel count 132 can provide a pixel count of object fence 110 pixels that overlap with lane mask 116 pixels, and the overlap can be used to determine lane assignment. As another example, a per-lane intersection ratio of the object fence 110 (determined by boundary score 134) can be used to assign objects to two or more lanes. Lane assignment 120 can represent the current lane of travel of the corresponding object. Vehicle 900 can use lane assignment 120 when making decisions, such as what path or trajectory the vehicle 900 should or can follow, whether to brake, accelerate, or change lanes, and / or other types of decisions.

[0062] In some embodiments, temporal smoothing 142 may be performed to utilize previous lane assignments 120 of one or more previous frames and / or one or more images of the object to estimate the position and / or lane assignment of the object in one or more subsequent frames and / or one or more images. For example, because lane assignments may be more accurate when the object is closer to one or more sensors of the vehicle 900, as the object moves farther away, previous lane assignments 120 may be utilized to provide a more accurate prediction of the object's position, as well as a more accurate prediction of lane assignments at greater distances (e.g., where lanes appear to merge in image space). In such an example, one or more previous predictions may be weighted relative to the current prediction of the lane assignment 120 (e.g., current prediction 90% / one or more previous predictions 10%, current prediction 70% / immediately preceding prediction 20% / two prior predictions 10%, etc.). Thus, temporal smoothing may utilize previous predictions to improve the accuracy of the lane assignments 120 within the process 100.

[0063] Reference now Figure 2A-2D , Figure 2A-2D A process of generating an object fence for a vehicle in an image using object detection 104 and free space detection 106 is shown in accordance with some embodiments of the present disclosure. Figure 2A 204 for a vehicle 202 (as represented in visualization 200) may be generated using an object detection algorithm and / or a machine learning model. In some examples, bounding shape 204 may be a bounding box around vehicle 202. Bounding shape 204 may be represented as points in input image 200 where vehicle 202 may be located (e.g., corresponding to each pixel of bounding shape 204, vertices along the perimeter of bounding shape 204, or a combination thereof).

[0064] Figure 2B A clipped bounding shape 222 of vehicle 202 detected in an image is shown, according to some embodiments of the present disclosure. For example, at least a portion of bounding shape 204 may be clipped to generate clipped bounding shape 222 as described herein.

[0065] Figure 2C Defining an object fence using drivable free space determination according to some embodiments of the present disclosure is shown. For example, drivable free space 232 (as described herein) can be clipped to boundary shape 204 and / or clipped boundary shape 222 to generate object fence 110. Figure 2COnly drivable free space 323 and non-drivable space 234 are shown within the clipped boundary shape 222, but this is not intended to be limiting. In some examples, some or all of each image may be defined as drivable free space or non-drivable space, and the portion of this output corresponding to the boundary shape 204 and / or the clipped boundary shape 222 may be used to generate the object fence 110.

[0066] Figure 2D An object fence for a vehicle generated by removing drivable free space from an object boundary shape is shown in accordance with some embodiments of the present disclosure. Figure 2D As shown, drivable free space 232 may be removed from clipped bounding shape 222 to generate object fence 110 .

[0067] View now Figure 3 , each block of the method 300 described herein includes a computing process that can be performed using any combination of hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in a memory. The method can also be embodied as computer usable instructions stored on a computer storage medium. The method can be provided by a stand-alone application, a service or a hosted service (standalone or in combination with another hosted service), or a plug-in for another product, to name a few. In addition, by way of example, for determining Figure 1 The process 100 of assigning lanes to objects in an environment of the present invention describes the method 300. However, the method may additionally or alternatively be performed by any one system or any combination of systems, including but not limited to the systems described herein.

[0068] Figure 3 is a flow chart illustrating a method 300 for determining lane assignments for objects in an environment according to some embodiments of the present disclosure. At block B302, the method 300 includes receiving object detection data representing a boundary shape corresponding to an object depicted in an image. For example, after object detection 112, a boundary shape corresponding to an object (e.g., a vehicle, a pedestrian, debris, etc.) depicted in an image (e.g., represented by sensor data 102) may be received. In some examples, the boundary shape may be a cropped boundary shape representing a portion of the boundary shape (e.g., Figure 2B The clipped boundary shape 222).

[0069] At block B304, the method 300 includes receiving free space data representing at least a portion of a bounding shape determined to include drivable free space. For example, after performing free space detection 106 on the sensor data 102, free space data representing at least a portion of a bounding shape determined to include drivable free space may be received.

[0070] At block B306 , method 300 includes cropping the portion of the bounding shape to generate an object fence from within the bounding shape. For example, object fence 110 may be generated by cropping the portion of the bounding shape from object detection 102 that represents free space data from free space determination 106 .

[0071] At block B308 , method 300 includes receiving lane data representing the location of a plurality of lane lines within the image. For example, lane data representing the location of a plurality of lane lines within sensor data 102 and / or lane data 126 may be received after performing lane detection 112 .

[0072] At block B310 , the method 300 includes generating a lane mask using the lane data. For example, the lane mask 116 may be generated from the lane data received from the lane detection 112 , for example using the lane triangulation 114 .

[0073] At block B312 , the method 300 includes determining an overlap between the object fence and one or more lanes defined by the lane mask. For example, the overlap between the object fence 110 and one or more lanes defined by the lane mask 116 may be determined by an overlap determination 110 using the pixel count 132 and / or the boundary score 134 .

[0074] At block B314 , method 300 includes assigning the object to at least one of the one or more lanes based at least in part on the overlap. For example, lane assignment 120 may be used to assign the object to at least one of the one or more lanes from lane mask 116 .

[0075] See also Figure 4 , Figure 4 An object fence 404 for a vehicle generated by a fence generation algorithm is shown according to some embodiments of the present disclosure. As described herein, the fence generation algorithm 108 may be used to generate an object fence 404 for the vehicle 402.

[0076] View now Figure 5, each block of the method 500 described herein includes a computing process that can be performed using any combination of hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in a memory. The method can also be embodied as computer usable instructions stored on a computer storage medium. The method can be provided by a stand-alone application, a service or a hosted service (standalone or in combination with another hosted service), or a plug-in for another product, to name a few. In addition, by way of example, for determining Figure 1 The process 100 of assigning lanes to objects in an environment of the present invention describes the method 500. However, the method may additionally or alternatively be performed by any one system or any combination of systems, including but not limited to the systems described herein.

[0077] Figure 5 1 is a flow chart showing a method 500 for determining lane assignments for objects in an environment according to some embodiments of the present disclosure. At block B502, the method 500 includes applying sensor data generated by one or more sensors to an object fence generation algorithm. For example, the sensor data 102 generated by one or more sensors may be applied to the fence generation algorithm 108.

[0078] At block B504, method 500 includes identifying, using an object fence algorithm, a world-space object fence for objects within a sensing field of at least one of the one or more sensors. For example, object fence algorithm 108 may be used to identify object fence 110 in world space for objects within a sensing field of at least one of the one or more sensors from which sensor data 102 is received.

[0079] At block B506, method 500 includes converting the world-space object fence to an image-space object fence based at least in part on sensor parameters corresponding to at least one of the one or more sensors. For example, fence generation algorithm 108 may be used to convert object fence 110 from world space to image space based at least in part on sensor parameters corresponding to at least one of the one or more sensors.

[0080] At block B508 , method 500 includes receiving lane data representing the positions of a plurality of lane lines within the image represented by the sensor data. For example, lane data representing the positions of a plurality of lane lines within the image represented by sensor data 102 may be received via lane detection 112 .

[0081] At block B510 , the method 500 includes generating a lane mask using the lane data. For example, the lane mask 116 may be generated using lane data received after the lane detection 112 (eg, using the lane triangulation 114 ).

[0082] At block B512 , method 500 includes determining an overlap between the image-space object fence and one or more lanes defined by the lane mask. For example, the overlap between the object fence 110 and one or more lanes defined by the lane mask 116 may be determined by the overlap determination 110 using the pixel count 132 and / or the boundary score 134 .

[0083] At block B514 , method 500 includes assigning the object to at least one of the one or more lanes based at least in part on the overlap. For example, lane assignment 120 may be used to assign the object to at least one of the one or more lanes in lane mask 116 .

[0084] View now Figure 6A-6E , Figure 6A-6E A process for generating a lane mask using lane data according to some embodiments of the present disclosure is shown. Fig. 6A Inflection points that define lanes generated by a lane detection algorithm are shown. The lane detection algorithm can detect lane lines in an input image from sensor data 102. Lane lines can include inflection points 602A-602H and 604A-604F that define lanes on a driving surface of vehicle 900. Inflection points can be points (e.g., pixels) along lane lines.

[0085] Figure 6B A polyline is shown that connects the breakpoints that define a lane generated by a lane detection algorithm according to some embodiments of the present disclosure. The lane detection algorithm can generate polylines 622A-622G and 624A-624E by connecting adjacent breakpoints 602A-602H and 604A-604F on the lane line of each lane, respectively. The polyline can represent a lane boundary (e.g., a line) representing the driving surface of the vehicle 900 in the image.

[0086] Figure 6C Lane triangulation performed on detected lanes to generate a lane mask according to some embodiments of the present disclosure is shown. Triangles (e.g., triangles 620A-620L, and other unnumbered triangles) may be generated by connecting the breakpoints 602A-602Z of lane boundaries with the breakpoints 604A-604Z of adjacent lane boundaries in a cross-wise manner using lane triangulation 114. Lane triangulation may be performed for each lane by generating triangles between lane boundaries to define the lane.

[0087] Fig.6D A lane mask 660 generated using lane data according to some embodiments of the present disclosure is shown. Lane mask 660 may include triangular lanes 662, 664, and 668. Each lane may be represented by an index of a different color. The pixels within each lane may then be stored in an index under the corresponding lane color. Lane mask 660 represents some or all lanes on the driving surface of vehicle 900.

[0088] Fig. 6E Lane extension is shown performed on lane mask 680 to extend a lane in accordance with some embodiments of the present disclosure. For example, if the full or available extent of a lane is not determined (as shown by lane extensions 682, 684, 686, and 688), lane extension 122 may be used to extend a detected lane in the lane mask.

[0089] View now Figure 7A-7B , Figure 7A-7B Vertices generated for an object fence to determine an overlap of the object fence with a lane in a lane mask are shown in accordance with some embodiments of the present disclosure. FIG. 7A shows vertices generated for an object fence to determine an overlap of the object fence with a lane in a lane mask when the object fence is within a lane relative to a particular frame. The object fence 706 may be represented by vertices (e.g., pixels) 702A, 702B, 702C, ..., 702N along the perimeter of the object fence 706 of the vehicle 704 in the image. The pixel distances between the vertices 702 of the object fence 706 may be determined to determine an overlap between the object fence and a lane 708 of the lane mask. For example, in the case where the object is located in a single lane, the pixel distances between each set of two surrounding pixels may be calculated, accumulated, and normalized to create a per-lane intersection ratio, which is 1 / 1 for a single lane.

[0090] Figure 7B Vertices generated for an object fence when the object fence intersects the lane boundary are shown to determine the overlap of the object fence with the lane in the lane mask according to some embodiments of the present disclosure. In the case where the vehicle 704 is in more than one lane (e.g., lanes 708 and 710), a set of points that intersect the lane boundary can be determined, intersection points 722A and 722B. Once the intersection points 722 are determined, these new vertices can be used to determine the distance between the new vertices and each other surrounding pixel or vertex of the object fence on either side of the intersection point, as described herein.

[0091] View now Figure 8, each block of the method 800 described herein includes a computing process that can be performed using any combination of hardware, firmware, and / or software. For example, various functions can be performed by a processor executing instructions stored in a memory. The method can also be embodied as computer usable instructions stored on a computer storage medium. The method can be provided by a stand-alone application, a service or a hosted service (standalone or in combination with another hosted service), or a plug-in for another product, to name a few. In addition, by way of example, for determining Figure 1 The process 100 of assigning lanes to objects in an environment of the present invention describes the method 800. However, the method may additionally or alternatively be performed by any one system or any combination of systems, including but not limited to the systems described herein.

[0092] Figure 8 8 is a flow chart showing a method 800 for determining a pixel overlap ratio between an object in an environment and two lanes according to some embodiments of the present disclosure. At block B802, the method 800 includes: determining that the object fence partially overlaps each of the first and second lanes based at least in part on the object fence and a lane mask representing the first lane and the second lane. For example, the overlap determination 118 may determine that the object fence 110 overlaps at least two lanes in the lane mask 116.

[0093] At block B804, method 800 includes identifying a pair of pixels corresponding to a vertex of an object fence, the pair of pixels being separated by a lane line that divides a first lane and a second lane. Figure 7B 708) and a second lane (eg, Figure 7B Pixel pairs (eg, lane lines (eg, lane boundaries) of object fences (eg, object fences 110, object fences 706) separated by lane lines (eg, lane boundaries) of lanes 710) Figure 7B Vertices 702A and 702B) can be identified via overlap determination 118.

[0094] At block B806, method 800 includes determining pixels corresponding to lane lines between pixel pairs. For example, the pixel corresponding to a pixel pair (e.g., Figure 7B vertices 702A and 702B) corresponding to the lane line (eg, Figure 7B intersection 722A).

[0095] At block B808, method 800 includes calculating a first sum of distances between a pixel and each of a plurality of first vertices of the object fence in the first lane and a second sum of distances between the pixel and each of a plurality of second vertices of the object fence in the second lane. For example, boundary score 134 may calculate a pixel (e.g., Figure 7B The first sum of the distances between the intersection 722A of FIG. 7B and each of the plurality of first vertices of the object fence (e.g., the object fence 110, the object fence 706) in the first lane (e.g., the lane 708 of FIG. 7B ) and the pixel and the second lane (e.g., Figure 7B A second sum of the distances between each of a plurality of second vertices of the object fence in lane 710).

[0096] At block B810, method 800 includes determining a first overlap ratio of the object fence in the first lane based at least in part on the first sum of the distances, and determining a second overlap ratio of the object fence in the second lane based at least in part on the second sum of the distances. For example, boundary score 134 may determine the first lane (e.g., Figure 7B The object fence 110 (or Figure 7B The first overlap ratio of the object fence 706) and the second lane determined based on at least the second and the distance (eg, Figure 7B A second overlap ratio of the object fence in lane 710).

[0097] At block B 812 , the method 800 includes performing at least one operation based at least in part on the first overlap ratio and the second overlap ratio. For example, the lane assignment 120 may assign the object to a lane based on the first overlap ratio and the second overlap ratio determined via the boundary score 134 .

[0098] Example Autonomous Vehicle

[0099] Fig. 9Ais an illustration of an example autonomous vehicle 900 according to some embodiments of the present disclosure. Autonomous vehicle 900 (alternatively, referred to herein as "vehicle 900") may include, but is not limited to, a passenger vehicle, such as a car, a truck, a bus, a first responder vehicle, a shuttle, an electric or motorized bicycle, a motorcycle, a fire truck, a police car, an ambulance, a boat, a construction vehicle, an underwater vehicle, a drone, and / or another type of vehicle (e.g., an unmanned vehicle and / or a vehicle that accommodates one or more passengers). Autonomous vehicles are generally described according to the automation levels defined by the National Highway Traffic Safety Administration (NHTSA), a division of the U.S. Department of Transportation, and the Society of Automotive Engineers (SAE) "Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles" (Standard No. J3016-201806 issued on June 15, 2018, Standard No. J3016-201609 issued on September 30, 2016, and previous and future versions of the standard). The vehicle 900 may be capable of implementing functions consistent with one or more of levels 3-5 of autonomous driving levels. For example, depending on the embodiment, the vehicle 900 may be capable of conditional automation (level 3), high automation (level 4), and / or full automation (level 5).

[0100] The vehicle 900 may include components such as a chassis, a body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other components of the vehicle. The vehicle 900 may include a propulsion system 950, such as an internal combustion engine, a hybrid power plant, an all-electric engine, and / or another type of propulsion system. The propulsion system 950 may be connected to a drive train of the vehicle 900, which may include a transmission, to achieve propulsion of the vehicle 900. The propulsion system 950 may be controlled in response to receiving a signal from a throttle / accelerator 952.

[0101] A steering system 954, which may include a steering wheel, may be used to steer the vehicle 900 (e.g., along a desired path or route) when the propulsion system 950 is operating (e.g., when the vehicle is in motion). The steering system 954 may receive signals from a steering actuator 956. For fully automated (Level 5) functionality, a steering wheel may be optional.

[0102] Brake sensor system 946 may be used to operate vehicle brakes in response to receiving signals from brake actuator 948 and / or brake sensors.

[0103] May include one or more system on chip (SoC) 904 ( Fig. 9C ) and / or one or more controllers 936 of one or more GPUs may provide signals (e.g., representing commands) to one or more components and / or systems of the vehicle 900. For example, the one or more controllers may send signals to operate the vehicle brakes via one or more brake actuators 948, to operate the steering system 954 via one or more steering actuators 956, and to operate the propulsion system 950 via one or more throttles / accelerators 952. The one or more controllers 936 may include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operating commands (e.g., signals representing commands) to enable autonomous driving and / or assist a human driver in driving the vehicle 900. The one or more controllers 936 may include a first controller 936 for autonomous driving functions, a second controller 936 for functional safety functions, a third controller 936 for artificial intelligence functions (e.g., computer vision), a fourth controller 936 for infotainment functions, a fifth controller 936 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 936 may handle two or more of the above functions, two or more controllers 936 may handle a single function, and / or any combination thereof.

[0104] The one or more controllers 936 may provide signals for controlling one or more components and / or systems of the vehicle 900 in response to sensor data (e.g., sensor input) received from one or more sensors. The sensor data may be received from, for example and without limitation, a global navigation satellite system sensor 958 (e.g., a global positioning system sensor), a RADAR sensor 960, an ultrasonic sensor 962, a LIDAR sensor 964, an inertial measurement unit (IMU) sensor 966 (e.g., an accelerometer, a gyroscope, a magnetic compass, a magnetometer, etc.), a microphone 996, a stereo camera 968, a wide angle camera 970 (e.g., a fisheye camera), an infrared camera 972, a surround camera 974 (e.g., a 360 degree camera), a long-range and / or mid-range camera 998, a speed sensor 944 (e.g., for measuring the velocity of the vehicle 900), a vibration sensor 942, a steering sensor 940, a brake sensor (e.g., as part of a brake sensor system 946), and / or other sensor types.

[0105] One or more of the controllers 936 may receive inputs (e.g., represented by input data) from the instrument cluster 932 of the vehicle 900 and provide outputs (e.g., represented by output data, display data, etc.) via a human machine interface (HMI) display 934, an audible annunciator, a speaker, and / or via other components of the vehicle 900. These outputs may include information such as vehicle speed, velocity, time, map data (e.g., Fig. 9C The HMI display 934 may display information such as the HD map 922 of the vehicle 900, location data (e.g., the location of the vehicle 900 on the map), directions, locations of other vehicles (e.g., occupancy grid), information about objects and object states as sensed by the controller 936, etc. For example, the HMI display 934 may display information about the presence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information about driving maneuvers that the vehicle has made, is making, or will make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0106] The vehicle 900 further includes a network interface 924 that can communicate via one or more networks using one or more wireless antennas 926 and / or a modem. For example, the network interface 924 may be capable of communicating via LTE, WCDMA, UMTS, GSM, CDMA 2000, etc. The one or more wireless antennas 926 may also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using one or more local area networks such as Bluetooth, Bluetooth LE, Z-wave, ZigBee, etc. and / or one or more low power wide area networks (LPWANs) such as LoRaWAN, SigFox, etc.

[0107] Fig. 9B For use according to some embodiments of the present disclosure Fig. 9A An example of camera positions and fields of view for an example autonomous vehicle 900. The cameras and respective fields of view are an example embodiment and are not intended to be limiting. For example, additional and / or replaceable cameras may be included, and / or the cameras may be located at different locations on the vehicle 900.

[0108] The camera type for the camera may include, but is not limited to, a digital camera that may be suitable for use with components and / or systems of the vehicle 900. The camera may operate at an automotive safety integrity level (ASIL) B and / or at another ASIL. The camera type may have any image capture rate, such as 60 frames per second (fps), 920fps, 240fps, etc., depending on the embodiment. The camera may be capable of using a rolling shutter, a global shutter, another type of shutter, or a combination thereof. In some examples, the color filter array may include a red, white, white, white (RCCC) color filter array, a red, white, white, blue (RCCB) color filter array, a red, blue, green, white (RBGC) color filter array, a Foveon X3 color filter array, a Bayer sensor (RGGB) color filter array, a monochrome sensor color filter array, and / or another type of color filter array. In some embodiments, a clear pixel camera such as a camera with an RCCC, RCCB, and / or RBGC color filter array may be used in an effort to improve light sensitivity.

[0109] In some examples, one or more of the cameras can be used to perform advanced driver assistance system (ADAS) functions (e.g., as part of a redundant or fail-safe design). For example, a multi-function monocular camera can be installed to provide functions including lane departure warning, traffic sign assistance, and intelligent headlight control. One or more of the cameras (e.g., all cameras) can simultaneously record and provide image data (e.g., video).

[0110] One or more of the cameras may be mounted in a mounting assembly such as a custom designed (3-D printed) assembly to cut off stray light and reflections from within the car (e.g., reflections from the dashboard reflected in the windshield mirror) that may interfere with the camera's image data capture capabilities. With respect to the wing mirror mounting assembly, the wing mirror assembly may be custom 3-D printed so that the camera mounting plate matches the shape of the wing mirror. In some examples, one or more cameras may be integrated into the wing mirror. For side view cameras, one or more cameras may also be integrated into the four pillars at each corner of the cab.

[0111] A camera (e.g., a front-facing camera) having a field of view that includes a portion of the environment in front of the vehicle 900 can be used for surround vision to help identify the forward path and obstacles, as well as assist in providing information critical to generating an occupancy grid and / or determining a preferred vehicle path with the help of one or more controllers 936 and / or control SoCs. The front-facing camera can be used to perform many of the same ADAS functions as LIDAR, including emergency braking, pedestrian detection, and collision avoidance. The front-facing camera can also be used for ADAS functions and systems, including lane departure warning ("LDW"), autonomous cruise control ("ACC"), and / or other functions such as traffic sign recognition.

[0112] A variety of cameras may be used in the front configuration, including, for example, a monocular camera platform including a CMOS (complementary metal oxide semiconductor) color imager. Another example may be a wide-angle camera 970, which may be used to sense objects (e.g., pedestrians, intersection traffic, or bicycles) entering the field of view from the periphery. Although Fig. 9B Only one wide-angle camera is shown in the figure, but there may be any number of wide-angle cameras 970 on the vehicle 900. In addition, long-range cameras 998 (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 been trained. Long-range cameras 998 may also be used for object detection and classification and basic object tracking.

[0113] One or more stereo cameras 968 may also be included in the front configuration. The stereo camera 968 may include an integrated control unit, which includes an expandable processing unit, which may provide a multi-core microprocessor and programmable logic (FPGA) with an integrated CAN or Ethernet interface on a single chip. Such a unit can be used to generate a 3-D map of the vehicle environment, including distance estimates for all points in the image. Alternative stereo cameras 968 may include a compact stereo vision sensor, which may include two camera lenses (one each on the left and right) and an image processing chip that can measure the distance from the vehicle to the target object and use the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning functions. In addition to those described herein or alternatively, other types of stereo cameras 968 may be used.

[0114] Cameras with a field of view of portions of the environment including the sides of the vehicle 900 (e.g., side view cameras) can be used for surround viewing, providing information used to create and update occupancy grids and generate side impact collision warnings. For example, surround cameras 974 (e.g., Fig. 9BThe four surround cameras 974 shown in FIG. 1 and FIG. 2 can be placed on the vehicle 900. The surround cameras 974 can include wide-angle cameras 970, fisheye cameras, 360-degree cameras, and / or the like. For example, the four fisheye cameras can be placed in front, back, and sides of the vehicle. In an alternative arrangement, the vehicle can use three surround cameras 974 (e.g., left, right, and back), and can utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround camera.

[0115] A camera having a field of view that includes a portion of the environment behind the vehicle 900 (e.g., a rear view camera) can be used to assist with parking, surround view, rear collision warning, and creating and updating occupancy grids. A variety of cameras can be used, including but not limited to cameras that are also suitable as front cameras as described herein (e.g., long-range and / or mid-range cameras 998, stereo cameras 968, infrared cameras 972, etc.).

[0116] Fig. 9C For use according to some embodiments of the present disclosure Fig. 9A Block diagram of an example system architecture of an example autonomous vehicle 900. It should be understood that this arrangement and other arrangements described herein are set forth merely as examples. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Further, many of the elements described herein are functional entities that may be implemented as discrete or distributed components or in combination with other components, and in any appropriate combination and location. The various functions described herein as being performed by an entity may be implemented by hardware, firmware, and / or software. For example, the various functions may be implemented by a processor executing instructions stored in a memory.

[0117] Fig. 9C Each of the components, features, and systems of the vehicle 900 is illustrated as being connected via a bus 902. The bus 902 may include a controller area network (CAN) data interface (alternatively, referred to herein as a "CAN bus"). The CAN may be a network inside the vehicle 900 that assists in controlling various features and functions of the vehicle 900, such as the actuation of brakes, acceleration, braking, steering, windshield wipers, and the like. The CAN bus may be configured to have tens or even hundreds of nodes, each with its own unique identifier (e.g., CAN ID). The CAN bus may be read to find steering wheel angle, ground speed, engine speed per minute (RPM), button position, and / or other vehicle status indicators. The CAN bus may be ASIL B compatible.

[0118] Although bus 902 is described as a CAN bus here, this is not intended to be restrictive. For example, in addition to or alternatively to the CAN bus, FlexRay and / or Ethernet can be used. In addition, although bus 902 is represented by a single line, this is not intended to be restrictive. For example, there can be any number of buses 902, which may include one or more CAN buses, one or more FlexRay buses, one or more Ethernet buses, and / or one or more other types of buses using different protocols. In some examples, two or more buses 902 can be used to perform different functions, and / or can be used for redundancy. For example, a first bus 902 can be used for a collision avoidance function, and a second bus 902 can be used for drive control. In any example, each bus 902 can communicate with any component of the vehicle 900, and two or more buses 902 can communicate with the same component. In some examples, each SoC 904, each controller 936, and / or each computer in the vehicle can have access to the same input data (e.g., input from a sensor of the vehicle 900), and can be connected to a common bus such as a CAN bus.

[0119] The vehicle 900 may include one or more controllers 936, such as those described herein. Fig. 9A Those controllers described. Controller 936 can be used for a variety of functions. Controller 936 can be coupled to any other different components and systems of vehicle 900, and can be used for control of vehicle 900, artificial intelligence of vehicle 900, infotainment for vehicle 900, and / or the like.

[0120] The vehicle 900 may include one or more system on chip (SoC) 904. The SoC 904 may include a CPU 906, a GPU 908, a processor 910, a cache 912, an accelerator 914, a data store 916, and / or other components and features not shown. The SoC 904 may be used to control the vehicle 900 in a variety of platforms and systems. For example, one or more SoCs 904 may be combined with an HD map 922 in a system (e.g., a system of the vehicle 900), and the HD map may be downloaded from one or more servers (e.g., a server) via a network interface 924. Fig.9D one or more servers 978) to obtain map refreshes and / or updates.

[0121] The CPU 906 may include a CPU cluster or CPU complex (alternatively, referred to herein as a "CCPLEX"). The CPU 906 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU 906 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 906 may include four dual-core clusters, each of which has a dedicated L2 cache (e.g., a 2MB L2 cache). The CPU 906 (e.g., CCPLEX) may be configured to support simultaneous cluster operations, such that any combination of CPU 906 clusters can be active at any given time.

[0122] CPU 906 may implement power management capabilities including one or more of the following features: Each hardware block may be automatically clock gated when idle to save dynamic power; each core clock may be gated when the core is not actively executing instructions due to the execution of WFI / WFE instructions; each core may be independently power gated; each core cluster may be independently clock gated when all cores are clock gated or power gated; and / or each core cluster may be independently power gated when all cores are power gated. CPU 906 may further implement an enhanced algorithm for managing power states, in which allowed power states and expected wake-up times are specified, and hardware / microcode determines the optimal power state to enter for the core, cluster, and CCPLEX. The processing core may support a simplified power state entry sequence in software, with the work being offloaded to the microcode.

[0123] The GPU 908 may include an integrated GPU (alternatively, referred to herein as an "iGPU"). The GPU 908 may be programmable and efficient for parallel workloads. In some examples, the GPU 908 may use an enhanced tensor instruction set. The GPU 908 may include one or more streaming microprocessors, each of which may include an L1 cache (e.g., an L1 cache with at least 96KB storage capacity), and two or more of these streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB storage capacity). In some embodiments, the GPU 908 may include at least eight streaming microprocessors. The GPU 908 may use a computing application programming interface (API). In addition, the GPU 908 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0124] In the case of automotive and embedded use, GPU 908 can be power optimized to achieve optimal performance. For example, GPU 908 can be manufactured on fin field effect transistors (FinFETs). However, this is not intended to be limiting, and GPU 908 can be manufactured using other semiconductor manufacturing processes. Each streaming microprocessor can merge several mixed precision processing cores divided into multiple blocks. For example and without limitation, 64 PF32 cores and 32 PF64 cores can be divided into four processing blocks. In such an example, each processing block can be allocated 16 FP32 cores, 8 FP64 Cores, 16 INT32 cores, two mixed precision NVIDIA tensor cores for deep learning matrix arithmetic, L0 instruction cache, warp scheduler, dispatch unit and / or 64KB register file. In addition, the streaming microprocessor may include independent parallel integer and floating point data paths, To provide efficient execution of workloads using a mix of computation and addressing computations. The streaming microprocessor may include independent thread scheduling capabilities to allow for finer-grained synchronization and cooperation between parallel threads. The streaming microprocessor may include a combined L1 data cache and shared memory unit to improve performance while simplifying programming.

[0125] GPU 908 can include high bandwidth memory (HBM) and / or a 16GB HBM2 memory subsystem that provides a peak memory bandwidth of approximately 900 GB / s in some examples. In some examples, synchronous graphics random access memory (SGRAM), such as fifth generation graphics double data rate synchronous random access memory (GDDR5), can be used in addition to or in lieu of HBM memory.

[0126] The GPU 908 may include unified memory technology that includes access counters to allow memory pages to be more accurately migrated to the processor that accesses them most frequently, thereby improving the efficiency of memory ranges shared between processors. In some examples, address translation service (ATS) support can be used to allow the GPU 908 to directly access the CPU 906 page table. In such an example, when the GPU 908 memory management unit (MMU) experiences a miss, an address translation request can be transmitted to the CPU 906. In response, the CPU 906 can look up the virtual-physical mapping for the address in its page table and transmit the translation back to the GPU 908. In this way, unified memory technology can allow a single unified virtual address space to be used for the memory of both the CPU 906 and the GPU 908, thereby simplifying GPU 908 programming and porting applications to the GPU 908.

[0127] In addition, GPU 908 may include access counters that can track how often GPU 908 accesses the memory of other processors. Access counters can help ensure that memory pages are moved to the physical memory of the processor that accesses them most frequently.

[0128] SoC 904 may include any number of caches 912, including those described herein. For example, cache 912 may include an L3 cache available to both CPU 906 and GPU 908 (e.g., connected to both CPU 906 and GPU 908). Cache 912 may include a write-back cache that may track the state of a line, for example, by using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). Depending on the embodiment, the L3 cache may include 4MB or more, although smaller cache sizes may also be used.

[0129] SoC 904 may include one or more accelerators 914 (e.g., hardware accelerators, software accelerators, or combinations thereof). For example, SoC 904 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 SRAM) may enable the hardware acceleration cluster to accelerate neural networks and other calculations. The hardware acceleration cluster may be used to supplement GPU 908 and offload some tasks of GPU 908 (e.g., free up more cycles of GPU 908 for performing other tasks). As an example, accelerator 914 may be used for targeted workloads (e.g., perception, convolutional neural networks (CNNs), etc.) that are sufficiently stable to be easily controlled for acceleration. When used herein, the term "CNN" may include all types of CNNs, including region-based or regional convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0130] Accelerator 914 (e.g., a hardware acceleration cluster) may include a deep learning accelerator (DLA). DLA may include one or more tensor processing units (TPUs) that may be configured to provide an additional 10 trillion operations per second for deep learning applications and reasoning. TPU may be an accelerator configured to perform image processing functions (e.g., for CNN, RCNN, etc.) and optimized for performing image processing functions. DLA may be further optimized for a specific set of neural network types and floating point operations and reasoning. The design of DLA may provide higher performance per millimeter than a general-purpose GPU, and far exceeds the performance of a CPU. TPU may perform several functions, including a single instance convolution function, support, for example, INT8, INT16, and FP16 data types for both features and weights, and post-processor functions.

[0131] The DLA can quickly and efficiently execute neural networks, particularly CNNs, on processed or unprocessed data for any of a wide variety of functions, such as, but not limited to: a CNN for object recognition and detection using data from a camera sensor; a CNN for distance estimation using data from a camera sensor; a CNN for emergency vehicle detection and identification and detection using data from a microphone; a CNN for facial recognition and vehicle owner identification using data from a camera sensor; and / or a CNN for safety and / or security-related events.

[0132] The DLA can perform any function of the GPU 908, and by using an inference accelerator, for example, the designer can target any function to either the DLA or the GPU 908. For example, the designer can focus the processing of CNNs and floating point operations on the DLA, and leave other functions to the GPU 908 and / or other accelerators 914.

[0133] The accelerator 914 (e.g., a hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may be alternatively referred to herein as a computer vision accelerator. The PVA may be designed and configured to accelerate computer vision algorithms for advanced driver assistance systems (ADAS), autonomous driving, and / or augmented reality (AR) and / or virtual reality (VR) applications. The PVA may provide a balance between performance and flexibility. For example, each PVA may include, for example and without limitation, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0134] The RISC core can interact with an image sensor (e.g., an image sensor of any camera described herein), an image signal processor, and / or the like. Each of these RISC cores can include any amount of memory. Depending on the embodiment, the RISC core can use any of a number of protocols. In some examples, the RISC core can execute a real-time operating system (RTOS). The RISC core can be implemented using one or more integrated circuit devices, application specific integrated circuits (ASICs), and / or storage devices. For example, the RISC core can include an instruction cache and / or a tightly coupled RAM.

[0135] The DMA may enable components of the PVA to access system memory independently of the CPU 906. The DMA may support any number of features used to provide optimizations for the PVA, including, but not limited to, support for multi-dimensional addressing and / or circular addressing. In some examples, the DMA may support addressing in up to six or more dimensions, which may include block width, block height, block depth, horizontal block stepping, vertical block stepping, and / or depth stepping.

[0136] The vector processor can be a programmable processor that can be designed to efficiently and flexibly perform programming for computer vision algorithms and provide signal processing capabilities. In some examples, the PVA can include a PVA core and two vector processing subsystem partitions. The PVA core can include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem can operate as the main processing engine of the PVA, and can include a vector processing unit (VPU), an instruction cache, and / or a vector memory (e.g., VMEM). The VPU core can include a digital signal processor, such as, for example, a single instruction multiple data (SIMD), a very long instruction word (VLIW) digital signal processor. The combination of SIMD and VLIW can enhance throughput and rate.

[0137] Each of the vector processors may include an instruction cache and may be coupled to a dedicated memory. As a result, in some examples, each of the vector processors may be configured to execute independently of the other vector processors. In other examples, 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 examples, 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 sequence images or portions of an image. Among other things, any number of PVAs may be included in a hardware acceleration cluster, and any number of vector processors may be included in each of these PVAs. In addition, the PVA may include additional error correction code (ECC) memory to enhance overall system security.

[0138] The accelerator 914 (e.g., a hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high bandwidth, low latency SRAM for the accelerator 914. In some examples, the on-chip memory may include at least 4MB of SRAM consisting of, for example and not limited to, eight field-configurable memory blocks, which may be accessed by both the PVA and the DLA. Each pair of memory blocks may include an advanced peripheral bus (APB) interface, a configuration circuit system, a controller, and a multiplexer. Any type of memory may be used. The PVA and DLA may access the memory via a backbone that provides high-speed memory access to the PVA and DLA. The backbone may include an on-chip computer vision network that interconnects the PVA and DLA to the memory (e.g., using APB).

[0139] The on-chip computer vision network may include an interface that determines that both the PVA and DLA provide ready and valid signals before transmitting any control signals / addresses / data. Such an interface may provide separate phases and separate channels for transmitting control signals / addresses / data, as well as burst communication for continuous data transmission. This type of interface may comply with ISO 26262 or IEC 61508 standards, but other standards and protocols may also be used.

[0140] In some examples, SoC 904 may include a real-time ray tracing hardware accelerator such as described in U.S. Patent Application No. 16 / 101,232 filed on August 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the position and range of objects (e.g., within a world model) in order to generate real-time visualization simulations for RADAR signal interpretation, for sound propagation synthesis and / or analysis, for SONAR system simulations, for general wave propagation simulations, for comparison with LIDAR data for the purpose of positioning and / or other functions, and / or for other uses.

[0141] The accelerator 914 (e.g., a hardware accelerator cluster) has a wide range of uses in autonomous driving. The PVA can be a programmable visual accelerator that can be used in key processing stages in ADAS and autonomous vehicles. The capabilities of the PVA are a good match for algorithmic domains that require predictable processing, low power, and low latency. In other words, the PVA performs well on semi-dense or dense rule calculations, and even on small data sets that require predictable runtimes with low latency and low power. Therefore, in the context of a platform for autonomous vehicles, the PVA is designed to run classic computer vision algorithms because they are efficient at object detection and integer math.

[0142] For example, according to one embodiment of the technology, PVA is used to perform computer stereo vision. In some examples, algorithms based on semi-global matching can be used, but this is not intended to be limiting. Many applications for level 3-5 autonomous driving require instant motion estimation / stereo matching (e.g., structure from motion, pedestrian recognition, lane detection, etc.). PVA can perform computer stereo vision functions on input from two monocular cameras.

[0143] In some examples, PVA can be used to perform dense optical flow. According to the process raw RADAR data (e.g., using 4D fast Fourier transform) to provide processed RADAR. In other examples, PVA is used for time-of-flight depth processing, which is for example by processing raw time-of-flight data to provide processed time-of-flight data.

[0144] DLA can be used to run any type of network to enhance control and driving safety, including, for example, a neural network that outputs a confidence measure for each object detection. Such a confidence value can be interpreted as a probability, or as providing a relative "weight" of each detection compared to other detections. This confidence value enables the system to make further decisions about which detections should be considered true positive detections rather than false positive detections. For example, the system can set a threshold for confidence and only consider detections that exceed the threshold as true positive detections. In an automatic emergency braking (AEB) system, a false positive detection will cause the vehicle to automatically perform emergency braking, which is obviously undesirable. Therefore, only the most confident detection should be considered a trigger for AEB. DLA can run a neural network for regressing confidence values. The neural network can take at least some subset of parameters as its input, such as bounding box dimensions, ground plane estimates obtained (e.g., from another subsystem), inertial measurement unit (IMU) sensor 966 output related to vehicle 900 orientation and distance, 3D position estimates of objects obtained from neural networks and / or other sensors (e.g., LIDAR sensor 964 or RADAR sensor 960), etc.

[0145] SoC 904 may include one or more data stores 916 (e.g., memory). Data store 916 may be on-chip memory of SoC 904 that may store neural networks to be executed on the GPU and / or DLA. In some examples, data store 916 may be large enough to store multiple instances of the neural network for redundancy and safety. Data store 912 may include an L2 or L3 cache 912. References to data store 916 may include references to memory associated with a PVA, DLA, and / or other accelerator 914 as described herein.

[0146] SoC 904 may include one or more processors 910 (e.g., embedded processors). Processor 910 may include a boot and power management processor, which may be a dedicated processor and subsystem for handling boot power and management functions and related safety implementations. The boot and power management processor may be part of the SoC 904 boot sequence and may provide runtime power management services. The boot power and management processor may provide clock and voltage programming, auxiliary system low power state transitions, SoC 904 thermal and temperature sensor management, and / or SoC 904 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to the temperature, and SoC 904 may use the ring oscillator to detect the temperature of CPU 906, GPU 908, and / or accelerator 914. If it is determined that the temperature exceeds a threshold, the boot and power management processor may enter a temperature fault routine and place SoC 904 in a lower power state and / or place the vehicle 900 in a driver safety parking mode (e.g., parking the vehicle 900 safely).

[0147] The processor 910 may further include a set of embedded processors that may be used as an audio processing engine. The audio processing engine may be an audio subsystem that allows for full hardware support for multi-channel audio through multiple interfaces and a wide range of flexible audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core having a digital signal processor with dedicated RAM.

[0148] The processor 910 may further include an always-on processor engine that may provide the necessary hardware features to support low-power sensor management and wake-up use cases. The always-on processor engine may include a processor core, tightly coupled RAM, supporting peripherals (e.g., timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0149] The processor 910 may further include a safety cluster engine, which includes a dedicated processor subsystem that handles safety management of automotive applications. The safety cluster engine may include two or more processor cores, tightly coupled RAM, supporting peripherals (e.g., timers, interrupt controllers, etc.), and / or routing logic. In safety mode, the two or more cores may operate in lockstep mode and act as a single core with comparison logic to detect any differences between their operations.

[0150] Processor 910 may further include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0151] Processor 910 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 a camera processing pipeline.

[0152] The processor 910 may include a video image compositer, which may be a processing block (e.g., implemented on a microprocessor) that implements the video post-processing functions required for the video playback application to generate the final image for the player window. The video image compositer may perform lens distortion correction for the wide-angle camera 970, the surround camera 974, and / or for the in-cab surveillance camera sensor. The in-cab surveillance camera sensor is preferably monitored by a neural network running on another instance of the advanced SoC, configured to recognize in-cab events and respond accordingly. The in-cab system may perform lip reading to activate mobile phone service and make calls, dictate emails, change vehicle destinations, activate or change the vehicle's infotainment system and settings, or provide voice-activated web surfing. Certain functions are only available to the driver when the vehicle is operating in autonomous mode and are disabled in other cases.

[0153] The video image compositer may include enhanced temporal noise reduction for both spatial and temporal noise reduction. For example, in the case of motion in the video, the noise reduction appropriately weights the spatial information and reduces the weight of information provided by adjacent frames. In the case where an image or portion of an image does not include motion, the temporal noise reduction performed by the video image compositer may use information from previous images to reduce noise in the current image.

[0154] The video image compositor may also be configured to perform stereoscopic rectification on the input stereoscopic footage frames. The video image compositor may further be used for user interface composition when the operating system desktop is in use and the GPU 908 does not need to continuously render new surfaces. Even when the GPU 908 is powered on and active for 3D rendering, the video image compositor may be used to offload the GPU 908 to improve performance and responsiveness.

[0155] The SoC 904 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 SoC 904 may further include an input / output controller that may be controlled by software and may be used to receive I / O signals that are not committed to a specific role.

[0156] SoC 904 may further include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management and / or other devices. SoC 904 may be used to process data from cameras (connected via Gigabit multimedia serial links and Ethernet), sensors (e.g., LIDAR sensor 964, RADAR sensor 960, etc., which may be connected via Ethernet), data from bus 902 (e.g., speed of vehicle 900, steering wheel position, etc.), data from GNSS sensor 958 (connected via Ethernet or CAN bus). SoC 904 may further include dedicated high-performance mass storage controllers, which may include their own DMA engines, and which may be used to release CPU 906 from routine data management tasks.

[0157] SoC 904 can be an end-to-end platform with a flexible architecture that spans automation levels 3-5, providing a comprehensive functional safety architecture that leverages and efficiently uses computer vision and ADAS technologies to achieve diversity and redundancy, together with deep learning tools to provide a platform for a flexible and reliable driving software stack. SoC 904 can be faster, more reliable, and even more energy efficient and space efficient than conventional systems. For example, when combined with CPU 906, GPU 908, and data storage 916, accelerator 914 can provide a fast and efficient platform for level 3-5 autonomous vehicles.

[0158] The technology thus provides capabilities and functionality that cannot be achieved by conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using high-level programming languages ​​such as the C programming language to execute a wide variety of processing algorithms across a wide variety of visual data. However, CPUs often fail to meet the performance requirements of many computer vision applications, such as those related to, for example, execution time and power consumption. In particular, many CPUs are unable to execute complex object detection algorithms in real time, which is a requirement for in-vehicle ADAS applications and a requirement for practical Level 3-5 autonomous vehicles.

[0159] 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 executed simultaneously and / or sequentially, and the results to be combined together to achieve Level 3-5 autonomous driving functions. For example, a CNN executed on a DLA or dGPU (e.g., GPU 920) may include text and word recognition, allowing the supercomputer to read and understand traffic signs, including signs for which a neural network has not been specifically trained. The DLA may further include a neural network capable of recognizing, interpreting, and providing semantic understanding of the signs, and passing that semantic understanding to a path planning module running on the CPU complex.

[0160] As another example, as required for Level 3, 4, or 5 driving, multiple neural networks may be running simultaneously. For example, a warning sign consisting of "Caution: Flashing Lights Indicate Icing Conditions" along with electric lights may be interpreted by several neural networks independently or collectively. The sign itself may be recognized as a traffic sign by a deployed first neural network (e.g., a trained neural network), and the text "Flashing Lights Indicate Icing Conditions" may be interpreted by a deployed second neural network that informs the vehicle's path planning software (preferably executing on a CPU complex) that icing conditions exist when the flashing lights are detected. The flashing lights may be recognized by operating a deployed third neural network over multiple frames that informs the vehicle's path planning software of the presence (or absence) of the flashing lights. All three neural networks may be running simultaneously, for example, within the DLA and / or on the GPU 908.

[0161] In some examples, a CNN for facial recognition and owner recognition can use data from a camera sensor to identify the presence of an authorized driver and / or owner of the vehicle 900. 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, disable the vehicle when the owner leaves the vehicle. In this way, the SoC 904 provides security against theft and / or carjacking.

[0162] In another example, a CNN for emergency vehicle detection and identification can use data from microphone 996 to detect and identify emergency vehicle sirens. In contrast to conventional systems that use general classifiers to detect sirens and manually extract features, SoC 904 uses CNN to classify environmental and urban sounds and to classify visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative closing rate 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 958. Thus, for example, when operating in Europe, the CNN will seek to detect European sirens, and when in the United States, the CNN will seek to identify only North American sirens. Once an emergency vehicle is detected, with the assistance of the ultrasonic sensor 962, the control program can be used to execute emergency vehicle safety routines to slow down the vehicle, drive to the side of the road, stop the vehicle, and / or idle the vehicle until the emergency vehicle passes.

[0163] The vehicle may include a CPU 918 (e.g., a discrete CPU or dCPU) that may be coupled to the SoC 904 via a high-speed interconnect (e.g., PCIe). The CPU 918 may include, for example, an X86 processor. The CPU 918 may be used to perform any of a variety of functions, including, for example, arbitrating potentially inconsistent results between ADAS sensors and the SoC 904, and / or monitoring the status and health of the controller 936 and / or the infotainment SoC 930.

[0164] The vehicle 900 may include a GPU 920 (e.g., a discrete GPU or dGPU) that may be coupled to the SoC 904 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU 920 may provide additional artificial intelligence functionality, for example, by executing redundant and / or different neural networks, and may be used to train and / or update the neural network based on input from sensors of the vehicle 900 (e.g., sensor data).

[0165] The vehicle 900 may further include a network interface 924, which may include one or more wireless antennas 926 (e.g., one or more wireless antennas for different communication protocols, such as cellular antennas, Bluetooth antennas, etc.). The network interface 924 can be used to enable wireless connection with the cloud (e.g., with a server 978 and / or other network devices), with other vehicles, and / or with a computing device (e.g., a passenger's client device) via the Internet. In order to communicate with other vehicles, a direct link can be established between the two vehicles, and / or an indirect link can be established (e.g., across a network and through the Internet). The direct link can be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link can provide the vehicle 900 with information about vehicles approaching the vehicle 900 (e.g., vehicles in front of, to the side of, and / or behind the vehicle 900). This functionality can be part of the collaborative adaptive cruise control functionality of the vehicle 900.

[0166] The network interface 924 may include a SoC that provides modulation and demodulation functions and enables the controller 936 to communicate via a wireless network. The network interface 924 may include a radio frequency front end for up-conversion from baseband to radio frequency and down-conversion from radio frequency to baseband. The frequency conversion may be performed by a known process, and / or may be performed using a super-heterodyne process. In some examples, the radio frequency front end function may be provided by a separate chip. The network interface may include a wireless function for communicating via LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0167] The vehicle 900 may further include data storage 928, which may include off-chip storage (e.g., outside the SoC 904). The data storage 928 may include one or more storage elements, including RAM, SRAM, DRAM, VRAM, flash memory, hard disk, and / or other components and / or devices that can store at least one bit of data.

[0168] The vehicle 900 may further include a GNSS sensor 958. The GNSS sensor 958 (e.g., GPS and / or assisted GPS sensor) is used to assist in mapping, perception, occupancy grid generation, and / or path planning functions. Any number of GNSS sensors 958 may be used, including, for example and without limitation, GPS using a USB connector with an Ethernet to serial (RS-232) bridge.

[0169] The vehicle 900 may further include a RADAR sensor 960. The RADAR sensor 960 may be used by the vehicle 900 for remote vehicle detection even in darkness and / or inclement weather conditions. The RADAR functional safety level may be ASIL B. The RADAR sensor 960 may use CAN and / or bus 902 (e.g., to transmit data generated by the RADAR sensor 960) for control and access to object tracking data, accessing Ethernet in some examples to access raw data. A variety of RADAR sensor types may be used. For example and without limitation, the RADAR sensor 960 may be suitable for front, rear, and side RADAR use. In some examples, a pulse Doppler RADAR sensor is used.

[0170] The RADAR sensor 960 may include different configurations, such as long-range with narrow field of view, short-range with wide field of view, short-range side coverage, etc. In some examples, the long-range RADAR may be used for adaptive cruise control functions. The long-range RADAR system may provide a wide field of view (e.g., within 250m) achieved by two or more independent scans. The RADAR sensor 960 may help distinguish between static objects and moving objects, and may be used by the ADAS system for emergency braking assistance and forward collision warnings. The long-range RADAR sensor may include a single-station multimode RADAR with multiple (e.g., six or more) fixed RADAR antennas and high-speed CAN and FlexRay interfaces. In an example with six antennas, the central four antennas may create a focused beam pattern designed to record the surroundings of the vehicle 900 at a higher rate with minimal traffic interference from adjacent lanes. The other two antennas may expand the field of view, making it possible to quickly detect vehicles entering or leaving the lane of the vehicle 900.

[0171] As an example, a medium-range RADAR system may include a range of up to 960m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 950 degrees (rear). A short-range RADAR system may include, but is not limited to, a RADAR sensor designed to be mounted at both ends of a rear bumper. When mounted at both ends of a rear bumper, such a RADAR sensor system may create two beams that continuously monitor the blind spots behind and beside the vehicle.

[0172] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0173] The vehicle 900 may further include an ultrasonic sensor 962. The ultrasonic sensor 962, which may be placed on the front, rear, and / or side of the vehicle 900, may be used for parking assistance and / or creating and updating an occupancy grid. A variety of ultrasonic sensors 962 may be used, and different ultrasonic sensors 962 may be used for different detection ranges (e.g., 2.5 m, 4 m). The ultrasonic sensor 962 may operate at a functional safety level of ASIL B.

[0174] The vehicle 900 may include a LIDAR sensor 964. The LIDAR sensor 964 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The LIDAR sensor 964 may be ASIL B for functional safety level. In some examples, the vehicle 900 may include multiple LIDAR sensors 964 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0175] In some examples, LIDAR sensor 964 may be able to provide a list of objects and their distances for a 360-degree field of view. Commercially available LIDAR sensor 964 may have, for example, an advertising range of approximately 900m, an accuracy of 2cm-3cm, and support for 900Mbps Ethernet connection. In some examples, one or more non-protruding LIDAR sensors 964 may be used. In such examples, LIDAR sensor 964 may be implemented as a small device that can be embedded in the front, back, side, and / or corner of vehicle 900. In such examples, LIDAR sensor 964 may provide up to 920 degrees horizontal and 35 degrees vertical field of view even for low reflectivity objects, with a range of 200m. Front mounted LIDAR sensor 964 may be configured for a horizontal field of view between 45 and 135 degrees.

[0176] In some examples, LIDAR technologies such as 3D flash LIDAR may also be used. 3D flash LIDAR uses a flash of laser as an emission source to illuminate the vehicle's surroundings up to about 200m. The flash LIDAR unit includes a receiver that records the laser pulse transmission time and the reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LIDAR can allow highly accurate and distortion-free images of the surrounding environment to be generated with each laser flash. In some examples, four flash LIDAR sensors can be deployed, one on each side of the vehicle 900. Available 3D flash LIDAR systems include solid-state 3D staring array LIDAR cameras (e.g., non-scanning LIDAR devices) with no moving parts other than fans. The flash LIDAR device can use 5 nanosecond Class I (eye-safe) laser pulses per frame, and can capture the reflected laser light in the form of a 3D range point cloud and co-registered intensity data. By using flash LIDAR, and because flash LIDAR is a solid-state device with no moving parts, LIDAR sensor 964 may be less susceptible to motion blur, vibration, and / or shock.

[0177] The vehicle may further include an IMU sensor 966. In some examples, the IMU sensor 966 may be located at the center of the rear axle of the vehicle 900. The IMU sensor 966 may include, for example and without limitation, an accelerometer, a magnetometer, a gyroscope, a magnetic compass, and / or other sensor types. In some examples, for example, in a six-axis application, the IMU sensor 966 may include an accelerometer and a gyroscope, and in a nine-axis application, the IMU sensor 966 may include an accelerometer, a gyroscope, and a magnetometer.

[0178] In some embodiments, the IMU sensor 966 can be implemented as a miniature high-performance GPS-assisted inertial navigation system (GPS / INS) that combines a micro-electromechanical system (MEMS) inertial sensor, a high-sensitivity GPS receiver, and an advanced Kalman filter algorithm to provide estimates of position, velocity, and attitude. Thus, in some examples, the IMU sensor 966 can enable the vehicle 900 to estimate heading by directly observing and correlating velocity changes from the GPS to the IMU sensor 966 without the need for input from a magnetic sensor. In some examples, the IMU sensor 966 and the GNSS sensor 958 can be combined into a single integrated unit.

[0179] The vehicle may include a microphone 996 positioned in and / or around the vehicle 900. The microphone 996 may be used for, among other things, emergency vehicle detection and identification.

[0180] The vehicle may further include any number of camera types, including a stereo camera 968, a wide-angle camera 970, an infrared camera 972, a surround camera 974, a long-range and / or mid-range camera 998, and / or other camera types. These cameras can be used to capture image data around the entire periphery of the vehicle 900. The type of camera used depends on the embodiment and the requirements of the vehicle 900, and any combination of camera types can be used to provide the necessary coverage around the vehicle 900. In addition, the number of cameras can vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. As an example and not limitation, the cameras can support Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the cameras is described herein with respect to Fig. 9A and Fig. 9B Described in more detail.

[0181] The vehicle 900 may further include a vibration sensor 942. The vibration sensor 942 may measure vibrations of components of the vehicle, such as axles. For example, changes in vibration may indicate changes in the road surface. In another example, when two or more vibration sensors 942 are used, the difference between the vibrations may be used to determine friction or slip of the road surface (e.g., when there is a vibration difference between a powered drive shaft and a freely rotating shaft).

[0182] The vehicle 900 may include an ADAS system 938. In some examples, the ADAS system 938 may include a SoC. The ADAS system 938 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward collision warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keeping assist (LKA), blind spot warning (BSW), rear cross traffic warning (RCTW), collision warning system (CWS), lane centering (LC), and / or other features and functions.

[0183] The ACC system may use a RADAR sensor 960, a LIDAR sensor 964, and / or a camera. The ACC system may include a longitudinal ACC and / or a lateral ACC. The longitudinal ACC monitors and controls the distance to the vehicle immediately in front of the vehicle 900, and automatically adjusts the vehicle speed to maintain a safe distance from the vehicle in front. The lateral ACC performs distance keeping and suggests that the vehicle 900 change lanes when necessary. The lateral ACC is related to other ADAS applications such as LCA and CWS.

[0184] CACC uses information from other vehicles, which can be received indirectly from other vehicles via a network interface 924 and / or a wireless antenna 926 via a wireless link or through a network connection (e.g., via the Internet). A direct link can be provided by a vehicle-to-vehicle (V2V) communication link, while an indirect link can be an infrastructure-to-vehicle (I2V) communication link. Typically, the V2V communication concept provides information about the vehicle immediately ahead (e.g., the vehicle immediately ahead of the vehicle 900 and in the same lane as it), while the I2V communication concept provides information about traffic farther ahead. The CACC system may include either or both of the I2V and V2V information sources. Given information about the vehicle ahead of the vehicle 900, CACC can be more reliable, and it is possible to improve the smoothness of traffic flow and reduce road congestion.

[0185] The FCW system is designed to alert the driver to hazards so that the driver can take corrective action. The FCW system uses a front camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker and / or vibrating components. The FCW system can provide warnings in the form of, for example, sound, visual warnings, vibrations and / or rapid brake pulses.

[0186] The AEB system detects an impending front collision with another vehicle or other object and can automatically apply the brakes if the driver does not take corrective action within specified time or distance parameters. The AEB system can use a front camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA and / or ASIC. When the AEB system detects a hazard, it typically first alerts the driver to take corrective action to avoid the collision, and if the driver does not take corrective action, the AEB system can automatically apply the brakes in an effort to prevent or at least mitigate the effects of the predicted collision. The AEB system may include technologies such as dynamic brake support and / or collision approach braking.

[0187] The LDW system provides visual, audible, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle 900 crosses a lane marking. When the driver indicates an intention to leave the lane, by activating a turn signal, the LDW system is not activated. The LDW system may use a front-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibrating component.

[0188] The LKA system is a variation of the LDW system. If the vehicle 900 begins to leave the lane, the LKA system provides steering input or braking to correct the vehicle 900.

[0189] The BSW system detects and warns the driver of vehicles in the car's blind spot. The BSW system can provide visual, auditory and / or tactile alerts to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses a turn signal. The BSW system can use a rear-facing camera and / or RADAR sensor 960 coupled to a dedicated processor, DSP, FPGA and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker and / or vibration components.

[0190] The RCTW system can provide visual, audible, and / or tactile notifications when an object is detected outside the range of the rear camera while the vehicle 900 is in reverse. Some RCTW systems include AEB to ensure that the vehicle brakes are applied to avoid a crash. The RCTW system can use one or more rear RADAR sensors 960 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration components.

[0191] Conventional ADAS systems may be prone to false positive results, which may annoy and distract the driver, but are typically not catastrophic because the ADAS system alerts the driver and allows the driver to decide whether the safety condition really exists and take action accordingly. However, in the autonomous vehicle 900, in the case of conflicting results, the vehicle 900 itself must decide whether to pay attention to the results from the main computer or the auxiliary computer (e.g., the first controller 936 or the second controller 936). For example, in some embodiments, the ADAS system 938 can be a backup and / or auxiliary computer for providing perception information to the backup computer rationality module. The backup computer rationality monitor can run redundant and diverse software on hardware components to detect faults in perception and dynamic driving tasks. The output from the ADAS system 938 can be provided to the supervisory MCU. If the outputs from the main computer and the auxiliary computer conflict, the supervisory MCU must determine how to coordinate the conflict to ensure safe operation.

[0192] In some examples, the primary computer can be configured to provide a confidence score to the supervisory MCU, indicating the primary computer's confidence in the selected result. If the confidence score exceeds a threshold, the supervisory MCU can follow the direction of the primary computer, regardless of whether the secondary computer provides conflicting or inconsistent results. In the event that the confidence score does not meet the threshold and the primary computer and the secondary computer indicate different results (e.g., conflicts), the supervisory MCU can arbitrate between these computers to determine the appropriate result.

[0193] The supervisory MCU can be configured to run a neural network that is trained and configured to determine conditions under which the auxiliary computer provides a false alarm based on outputs from the main computer and the auxiliary computer. Therefore, the neural network in the supervisory MCU can learn when the output of the auxiliary computer can be trusted and when it cannot. For example, when the auxiliary computer is a RADAR-based FCW system, the neural network in the supervisory MCU can learn when the FCW system is identifying a metal object that is not actually dangerous, such as a drainage grid or manhole cover that triggers an alarm. Similarly, when the auxiliary computer is a camera-based LDW system, the neural network in the supervisory MCU can learn to ignore the LDW when a cyclist or pedestrian is present and lane departure is actually the safest strategy. In an embodiment including a neural network running on the supervisory MCU, the supervisory MCU may include at least one of a DLA or GPU suitable for running the neural network with associated memory. In a preferred embodiment, the supervisory MCU may include a component of SoC 904 and / or be included as a component of SoC 904.

[0194] In other examples, the ADAS system 938 may include an auxiliary computer that uses traditional computer vision rules to perform ADAS functions. In this way, the auxiliary computer can use classic computer vision rules (if-then), and the presence of a neural network in the supervisory MCU can improve reliability, safety, and performance. For example, diverse implementations and intentional non-identity make the entire system more fault-tolerant, especially for failures caused by software (or software-hardware interface) functions. For example, if there is a software vulnerability or error in the software running on the main computer and the non-identical software code running on the auxiliary computer provides the same overall result, the supervisory MCU can be more confident that the overall result is correct and that the vulnerability in the software or hardware on the main computer does not cause a substantial error.

[0195] In some examples, the output of the ADAS system 938 can be fed to the perception block of the main computer and / or the dynamic driving task block of the main computer. For example, if the ADAS system 938 indicates a forward collision warning due to an object immediately ahead, the perception block can use this information when identifying the object. In other examples, the auxiliary computer can have its own neural network that is trained and thus reduces the risk of false positives as described herein.

[0196] The vehicle 900 may further include an infotainment SoC 930 (e.g., an in-vehicle infotainment system (IVI)). Although illustrated and described as an SoC, the infotainment system may not be an SoC and may include two or more discrete components. The infotainment SoC 930 may include a combination of hardware and software that can be used to provide audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming media, etc.), phone (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.), and / or information services (e.g., navigation system, rear parking assistance, radio data system, vehicle-related information such as fuel level, total distance covered, brake fuel level, oil level, door open / closed, air filter information, etc.) to the vehicle 900. For example, the infotainment SoC 930 may include a radio, a disc player, a navigation system, a video player, USB and Bluetooth connections, an onboard computer, onboard entertainment, Wi-Fi, steering wheel audio controls, hands-free voice controls, a head-up display (HUD), an HMI display 934, 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 930 may further be used to provide information (e.g., visual and / or auditory) to a user of the vehicle, such as information from an ADAS system 938, autonomous driving information such as planned vehicle maneuvers, trajectories, surrounding environment information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0197] The infotainment SoC 930 may include GPU functionality. The infotainment SoC 930 may communicate with other devices, systems, and / or components of the vehicle 900 via a bus 902 (e.g., a CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 930 may be coupled to a supervisory MCU so that in the event of a failure of a main controller 936 (e.g., a main and / or backup computer of the vehicle 900), the GPU of the infotainment system may perform some self-driving functions. In such an example, the infotainment SoC 930 may place the vehicle 900 in a driver safe parking mode as described herein.

[0198] The vehicle 900 may further include an instrument cluster 932 (e.g., a digital instrument panel, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 932 may include a controller and / or a supercomputer (e.g., a separate controller or a supercomputer). The instrument cluster 932 may include a set of instruments, such as a speedometer, a fuel level, an oil pressure, a tachometer, an odometer, a turn indicator, a shift position indicator, a seat belt warning light, a parking brake warning light, an engine fault light, an airbag (SRS) system information, a lighting control, a safety system control, navigation information, etc. In some examples, information may be displayed and / or shared between the infotainment SoC 930 and the instrument cluster 932. In other words, the instrument cluster 932 may be included as part of the infotainment SoC 930, or vice versa.

[0199] Fig.9D For a cloud-based server and Fig. 9A 900. System 976 may include a server 978, a network 990, and a vehicle including the vehicle 900. The server 978 may include a plurality of GPUs 984 (A) -984 (H) (collectively referred to herein as GPUs 984), PCIe switches 982 (A) -982 (H) (collectively referred to herein as PCIe switches 982), and / or CPUs 980 (A) -980 (B) (collectively referred to herein as CPUs 980). The GPUs 984, CPUs 980, and PCIe switches may be interconnected with high-speed interconnects and / or PCIe connections 986 such as, for example and without limitation, the NVLink interface 988 developed by NVIDIA. In some examples, the GPUs 984 are connected via NVLink and / or NVSwitch SoCs, and the GPUs 984 and PCIe switches 982 are connected via PCIe interconnects. Although eight GPUs 984, two CPUs 980, and two PCIe switches are illustrated, this is not intended to be limiting. Depending on the embodiment, each of the servers 978 may include any number of GPUs 984, CPUs 980, and / or PCIe switches. For example, each of the servers 978 may include eight, sixteen, thirty-two, and / or more GPUs 984.

[0200] Server 978 may receive image data over network 990 and from a vehicle, the image data representing images showing unexpected or changed road conditions, such as recently begun road work. Server 978 may transmit neural network 992, updated neural network 992, and / or map information 994, including information about traffic and road conditions, over network 990 and to the vehicle. Updates to map information 994 may include updates to HD map 922, such as information about construction sites, potholes, curves, flooding, or other obstacles. In some examples, neural network 992, updated neural network 992, and / or map information 994 may have been generated from new training and / or data received from any number of vehicles in the environment and / or based on experience from training performed at a data center (e.g., using server 978 and / or other servers).

[0201] Server 978 can be used to train a machine learning model (e.g., a neural network) based on training data. The training data can be generated by the vehicle, and / or can be generated in simulation (e.g., using a game engine). In some examples, the training data is labeled (e.g., where the neural network benefits from supervised learning) and / or undergoes other preprocessing, while in other examples, the training data is not labeled and / or preprocessed (e.g., where the neural network does not require supervised learning). Once the machine learning model is trained, the machine learning model can be used by the vehicle (e.g., transmitted to the vehicle via network 990), and / or the machine learning model can be used by server 978 to remotely monitor the vehicle.

[0202] In some examples, server 978 can receive data from the vehicle and apply the data to the latest real-time neural network for real-time intelligent reasoning. Server 978 may include a deep learning supercomputer and / or a dedicated AI computer powered by GPU 984, such as DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 978 may include a deep learning infrastructure of a data center powered only by CPUs.

[0203] The deep learning infrastructure of server 978 may be capable of rapid real-time inference, and may use this capability to assess and verify the health of the processors, software, and / or associated hardware in vehicle 900. For example, the deep learning infrastructure may receive periodic updates from vehicle 900, such as a sequence of images and / or objects located in the sequence of images that vehicle 900 has located (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural network to identify objects and compare them to the objects identified by vehicle 900, and if the results do not match and the infrastructure concludes that the AI ​​in vehicle 900 has failed, then server 978 may transmit a signal to vehicle 900 instructing the fail-safe computer of vehicle 900 to take control, notify passengers, and complete a safe parking maneuver.

[0204] For reasoning, the server 978 may include a GPU 984 and one or more programmable reasoning accelerators (e.g., NVIDIA's TensorRT). The combination of GPU-powered servers and reasoning acceleration can make real-time responses possible. In other examples, such as when performance is not so important, CPU, FPGA, and other processor-powered servers can be used for reasoning.

[0205] Example computing device

[0206] Fig.10 1 is a block diagram of an example computing device 1000 suitable for implementing some embodiments of the present disclosure. The computing device 1000 may include a bus 1002 that directly or indirectly couples the following devices: a memory 1004, one or more central processing units (CPUs) 1006, one or more graphics processing units (GPUs) 1008, a communication interface 1010, an input / output (I / O) port 1012, an input / output component 1014, a power supply 1016, and one or more presentation components 1018 (e.g., a display).

[0207] although Fig.10 The various blocks of are shown as being connected via a bus 1002 having wires, but this is not intended to be limiting and is merely for clarity. For example, in some embodiments, presentation components 1018 such as a display device may be considered I / O components 1014 (e.g., if the display is a touch screen). As another example, CPU 1006 and / or GPU 1008 may include memory (e.g., memory 1004 may represent a storage device in addition to the memory of GPU 1008, CPU 1006, and / or other components). In other words, Fig.10The computing devices described herein are illustrative only. No distinction is made between categories such as "workstations," "servers," "laptops," "desktops," "tablets," "client devices," "mobile devices," "handheld devices," "game consoles," "electronic control units (ECUs)," "virtual reality systems," and / or other device or system types, as all are considered Fig.10 within the range of computing devices.

[0208] The bus 1002 may represent one or more buses, such as an address bus, a data bus, a control bus, or a combination thereof. The bus 1002 may include one or more bus 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.

[0209] Memory 1004 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 1000. Computer-readable media may include volatile and nonvolatile media and removable and non-removable media. By way of example and not limitation, computer-readable media may include computer storage media and communication media.

[0210] Computer storage media may include volatile and nonvolatile media and / or removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, and / or other data types. For example, memory 1004 may store computer-readable instructions (e.g., representing programs and / or program elements, such as an operating system). Computer storage media may include, but are not limited to, RAM, ROM, EEPROM, flash memory or other storage technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by computing device 1000. As used herein, computer storage media does not include the signals themselves.

[0211] Communication media may contain computer readable instructions, data structures, program modules, and / or other data types in a modulated data signal such as a carrier wave or other transmission mechanism, and include any information transport medium. The term "modulated data signal" may refer to a signal that has one or more of its characteristics set or changed in such a way that information is encoded into the signal. By way of example and not limitation, communication media may include wired media such as a wired network or a direct wired connection, and wireless media such as sound, RF, infrared, and other wireless media. Any combination of the above should also be included within the scope of computer readable media.

[0212] CPU 1006 may be configured to execute computer-readable instructions to control one or more components of computing device 1000 to perform one or more of the methods and / or processes described herein.

[0213] Each of the CPUs 1006 may include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of processing a large number of software threads simultaneously. The CPUs 1006 may include any type of processor, and may include different types of processors, depending on the type of computing device 1000 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 1000, the processor may be an ARM processor implemented using reduced instruction set computing (RISC) or an x86 processor implemented using complex instruction set computing (CISC). The computing device 1000 may include one or more CPUs 1006 in addition to one or more microprocessors or supplementary coprocessors such as math coprocessors.

[0214] GPU 1008 can be used by computing device 1000 to render graphics (e.g., 3D graphics). GPU 1008 can include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. GPU 1008 can generate pixel data for outputting an image in response to a rendering command (e.g., a rendering command from CPU 1006 received via a host interface). GPU 1008 can include a graphics memory such as a display memory for storing pixel data. Display memory can be included as part of memory 1004. GPU 1008 can include two or more GPUs operating in parallel (e.g., via a link). When combined together, each GPU 1008 can generate pixel data for different parts of an output image or for different output images (e.g., a first GPU for a first image, a second GPU for a second image). Each GPU can include its own memory, or can share memory with other GPUs.

[0215] In examples where computing device 1000 does not include GPU 1008 , CPU 1006 may be used to render graphics.

[0216] The communication interface 1010 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 1000 to communicate with other computing devices via an electronic communication network, including wired and / or wireless communications. The communication interface 1010 may include components and functionality that 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), a low power wide area network (e.g., LoRaWAN, SigFox, etc.), and / or the Internet.

[0217] I / O ports 1012 can enable computing device 1000 to be logically coupled to other devices including I / O components 1014, presentation components 1018, and / or other components, some of which can be built into (e.g., integrated into) computing device 1000. Illustrative I / O components 1014 include microphones, mice, keyboards, joysticks, game pads, game controllers, satellite dishes, scanners, printers, wireless devices, and the like. I / O components 1014 can provide a natural user interface (NUI) that processes air gestures, voice, or other physiological inputs generated by a user. In some instances, the input can be transmitted to an appropriate network element for further processing. NUI can implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, gesture recognition on and adjacent to the screen, air gestures, head and eye tracking, and touch recognition associated with the display of computing device 1000 (as described in more detail below). Computing device 1000 may include a depth camera such as a stereo camera system, an infrared camera system, an RGB camera system, touch screen technology, and combinations of these for gesture detection and recognition. Additionally, computing device 1000 may include an accelerometer or gyroscope to enable motion detection (e.g., as part of an inertial measurement unit (IMU)). In some examples, the output of the accelerometer or gyroscope may be used by computing device 1000 to render immersive augmented reality or virtual reality.

[0218] The power supply 1016 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 1016 may provide power to the computing device 1000 to enable the components of the computing device 1000 to operate.

[0219] The presentation component 1018 may include a display (e.g., a monitor, a touch screen, a television screen, a head-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The presentation component 1018 may receive data from other components (e.g., GPU 1008, CPU 1006, etc.) and output the data (e.g., as an image, video, sound, etc.).

[0220] The present disclosure may be described in the general context of machine-usable instructions or computer codes executed by a computer or other machine such as a personal digital assistant or other handheld device, including computer-executable instructions such as program modules. Typically, program modules including routines, programs, objects, components, data structures, etc. refer to codes that perform specific tasks or implement specific abstract data types. The present disclosure may be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. The present disclosure may also be practiced in a distributed computing environment in which tasks are performed by remote processing devices linked through a communication network.

[0221] As used herein, the description of "and / or" about two or more elements should be interpreted as referring to only one element or a combination of elements. For example, "element A, element B and / or element C" can include only element A, only element B, only element C, element A and element B, element A and element C, element B and element C, or elements A, B and C. In addition, "at least one of element A or element B" can include at least one of element A, at least one of element B, or at least one of element A and at least one of element B. Further, "at least one of element A and element B" can 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.

[0222] The subject matter of the present disclosure is described in detail herein to meet statutory requirements. However, the description itself is not intended to limit the scope of the present disclosure. On the contrary, the inventors have contemplated that the claimed subject matter may also be embodied in other ways to include steps different from the steps described herein in conjunction with other current or future technologies or combinations of similar steps. Moreover, although the terms "step" and / or "block" may be used herein to imply different elements of the method employed, these terms should not be interpreted as implying any particular order among or between the various steps disclosed herein, unless the order of the steps is explicitly described.

Claims

1. A method comprising: receiving first data representing a boundary shape corresponding to an object depicted in an image; receiving second data representing at least a portion of the boundary shape, the second data being determined to include drivable free space; cropping the portion of the bounding shape to generate an object fence from within the bounding shape; receiving third data representing positions of a plurality of lane markings in the image; generating a lane mask using the third data; determining an overlap between the object fence and one or more lanes in the image defined by the lane mask; as well as The object is assigned to at least one of the one or more lanes based at least in part on the overlap.

2. The method of claim 1, wherein the step of clipping the portion of the border comprises: clipping a first portion of the boundary shape to generate a clipped boundary shape, A second portion of the bounding shape is cropped to generate an object fence from the cropped bounding shape.

3. The method of claim 1 , wherein the determining the overlap comprises: determining distances between pairs of vertices of the object fence; calculating the sum of the distances; as well as The overlap is determined at least in part using the sum of the distances.

4. The method of claim 1 , wherein said determining said overlap comprises: At least one fence pixel corresponding to the object fence in the image is determined to be the same pixel as at least one lane pixel corresponding to a lane of the one or more lanes. 5 . The method of claim 1 , wherein the third data is determined from at least one of a lane detection algorithm or fourth data from a map application.

6. The method of claim 1 , wherein said determining said overlap comprises: A number of pixels shared between the object fence and a lane of the one or more lanes is compared to exceed a threshold number of pixels.

7. A method comprising: applying first data generated by the one or more sensors to an object fence generation algorithm; identifying, using the object fence generation algorithm, a world space object fence of objects within a sensor field of view of at least one of the one or more sensors; converting the world-space object fence to an image-space object fence based at least in part on sensor parameters corresponding to at least one sensor of the one or more sensors; receiving second data representing positions of a plurality of lane markings in the image represented by the first data; generating a lane mask using the second data; determining an overlap between the image-space object fence and one or more lanes defined by the lane mask; as well as Based at least in part on the overlap, the object is assigned to at least one lane of the one or more lanes.

8. The method of claim 7, wherein the second data is received from one or more of a mapping application, a lane detection algorithm, or a neural network.

9. The method of claim 7, wherein said generating said lane mask comprises: Vertices between lane lines in the plurality of lane lines corresponding to the same lane are connected in a cross pattern using triangulation.

10. The method according to claim 7, wherein: The image space object fence is represented by a plurality of vertices of a polygon; as well as The determining the overlap of the image space object fence with the one or more lanes comprises: For each pair of vertices in the plurality of vertices, calculating a distance between the vertices; adding the distances for each pair of vertices in the plurality of vertices to generate a sum of distances; and The overlap is determined based at least in part on the sum of the distances.

11. The method according to claim 7, wherein: The image space object fence is represented by multiple vertices of a polygon; as well as When the image space object fence overlaps with the first lane and the second lane, determining the overlap includes: determining vertex pairs among the plurality of vertices that are divided by a lane line among the plurality of lane lines; determining pixels between the vertex pairs; calculating a first sum of distances between the pixel in the first lane and each of the plurality of vertices and a second sum of distances between the pixel in the second lane and each of the plurality of vertices; and An overlap ratio of the image space object fence in each of the first lane and the second lane is determined using the first sum of the distances and the second sum of the distances.

12. The method of claim 7, wherein an image space object fence is determined on a plurality of consecutive images of the object, and parameters of a first image space object fence in the image space object fence are used to determine second parameters of a second image space object fence in the image space object fence.

13. The method of claim 7, wherein said generating said lane mask comprises: A lane of the one or more lanes is extended beyond a range represented by the second data, extending the lane comprising at least one of curve fitting or extrapolation.

14. A system comprising: one or more sensors for generating sensor data of an environment surrounding the vehicle; A computing device comprising one or more processing devices and one or more storage devices communicatively coupled to the one or more processing devices, the one or more storage devices storing thereon programming instructions that, when executed by the processor, cause the instantiation of: an object fence generator for generating an object fence corresponding to an object in the environment based at least in part on the sensor data; a lane allocator for comparing the object fence to a lane mask representing at least a first lane and a second lane of lanes in the environment to determine a lane assignment for the object; as well as A control component is configured to use the lane assignment to determine one or more operations for controlling the vehicle within the environment. 15 . The system of claim 14 , further comprising a lane identifier for identifying the lane using at least one of the sensor data or lane data representing one or more of the lanes in the environment.

16. The system of claim 14, wherein the object fence is generated by cutting out a drivable free space to have a boundary shape corresponding to the object.

17. The system of claim 14, wherein the object fence is generated using a fence generation algorithm that uses at least a portion of the sensor data as input.

18. The system of claim 14, wherein the lane allocator is further configured to: determining, based at least in part on the object fence and the lane mask, that an object fence in the object fence at least partially overlaps with each of the first lane and the second lane; identifying pairs of pixels corresponding to vertices of the object fence, the pairs of pixels being separated by a lane line dividing the first lane and the second lane; determining a pixel corresponding to the lane line between the pixel pairs; calculating a first sum of distances between the pixel in the first lane and each of a plurality of first vertices of the object fence and a second sum of distances between the pixel in the second lane and each of a plurality of second vertices of the object fence; as well as A first overlap ratio of the object fence in the first lane is determined based at least in part on the first sum of the distances, and a second overlap ratio of the object fence in the second lane is determined based at least in part on the second sum of the distances.

19. The system of claim 18, wherein the lane allocator further normalizes the first sum of the distances and the second sum of the distances.

20. The system of claim 14, wherein the one or more operations correspond to path planning for the vehicle through the lane.

Images