DETECTION OF FAULTS IN THE SETTINGS OF SENSOR DEVICES

Abstract

In various examples, a technique for verifying sensor settings is described, comprising receiving sensor output data and at least one output sensor setting associated with the sensor output data from a sensor. The technique further comprises determining whether a queue of pending input sensor settings includes a matching input sensor setting that matches the output sensor setting. The technique also includes, in response to determining that the queue of pending input sensor settings includes the matching input sensor setting, identifying a pending input sensor setting in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue.The technique further includes determining whether a sensor setting failure has occurred based on the outstanding request count associated with the outstanding input setting. The technique further includes generating an error message based on the outstanding request count.

Description

BACKGROUND

[0001] An autonomous or semi-autonomous vehicle is typically equipped with cameras, depth sensors, and / or other sensors designed to detect and understand the vehicle's surroundings. The data collected by the sensors is used by downstream components to make critical decisions such as stopping the vehicle, starting the vehicle, controlling the vehicle's speed, determining the path the vehicle should take, deactivating an autonomous driving mode of the vehicle, and / or otherwise operating the vehicle. The accuracy of the sensors under certain conditions may depend on the correct configuration of the sensors. A sensor can accept one or more settings—that is, specific values ​​that configure the sensor's operation.The term “sensor setting” here can refer to a set of one or more setting values ​​that can be sent individually and / or as a group to a number of sensors, e.g., in a single configuration request. The setting values ​​can be values ​​of configuration parameters for the sensor. For example, a camera has an exposure parameter that controls the duration of the exposure of the camera sensor and a gain parameter that controls the amount of amplification of the pixels captured by the camera sensor. The appropriate values ​​for the exposure and gain parameters depend on various factors, including the characteristics of the scene being captured by the camera. Sensor settings for a camera, also referred to here as camera settings, can be specific values ​​for the exposure and gain parameters, e.g., 25 milliseconds and 10x, for example.

[0002] Sensor settings can be sent to the sensor from other system components. For example, a sensor controller can determine suitable sensor settings and send them to the sensor. For example, the sensor can have a configuration interface that receives specific settings, and the sensor controller can send the settings to the configuration interface to cause the sensor to use the sensor settings. When the sensor receives a specific set of settings, it configures the corresponding parameters to use the specified set of sensor settings. The processes of sending sensor settings to the sensor and / or configuring the sensor to use the sensor settings are referred to herein as applying the sensor settings to the sensor. After the sensor settings are applied, the sensor operates according to the settings.The sensor then receives information about the environment and generates detection information according to the sensor settings.

[0003] For example, a camera controller determines appropriate values ​​for the camera's exposure and gain settings based on scene statistics from a recent image captured by the camera and previous camera settings sent to the camera before the current image was captured. The camera controller can perform continuous automatic exposure (CAE), which determines the camera's exposure and gain settings to achieve optimized brightness in images captured by the camera. The camera controller applies the determined camera settings to the camera by sending the settings to the camera. If the camera settings are successfully applied to the camera, the camera uses the specified settings when capturing images.

[0004] If the sensor settings are not applied successfully, this may be due to unexpected conditions or behavior, e.g., if the hardware or software is not functioning correctly, e.g., does not meet the specifications. If the settings are not applied successfully, this may lead to incorrect results, e.g., inaccurate images being captured by the camera. A failure of a system component involved in sending and / or configuring the sensor to use the settings may, for example, result in the sensor settings not being applied. System components that, if operating incorrectly, may result in the sensor settings not being applied include the camera controller, a communications bus, or the sensor. A malfunction of the sensor, e.g.,If the sensor does not accept the settings, this can occur without generating an error message or other specific indication of a malfunction. A malfunction in the application of the sensor settings can also occur if sensor settings other than those specified for the sensor are used, which can cause the sensor to produce incorrect output.

[0005] Applications such as automotive systems require accurate and timely sensor output. Consequently, accurate and efficient verification of sensor settings is critical for reliable autonomous or semi-autonomous operation and navigation. For example, if the exposure time of a camera in a vehicle system is incorrectly set, this can result in captured images that are under- or overexposed, thus obscuring the surrounding objects. Such inaccurate or otherwise erroneous sensor data can prevent downstream components from correctly detecting objects in the vehicle's surroundings, or from detecting them at all, leading to incorrect or improper driving decisions.

[0006] One approach to verifying that sensor settings have been correctly applied to a sensor is to store a copy of the "input" settings sent to the sensor. When the sensor output is received, the "output" settings provided as part of the sensor output are compared with the stored copy of the input settings. If the input settings do not match the output settings within an error tolerance limit, the input settings have not been applied to the sensor, or if applied, they have not been applied correctly. The error tolerance limit is based on an expected quantization error that arises from converting the input settings into a quantized value in the sensor. The input settings, e.g.Exposure and / or gain may be rounded to different "quantized" values ​​by the sensor because there are only a limited number of bits available to represent such settings (or to represent values ​​that affect such settings). The sensor uses the quantized values ​​for the settings, so the quantized values ​​are provided by the sensor as output settings, which represent the settings used in generating the sensor output. Because the output settings can differ from the input settings by the amount of quantization error, the quantization error is used as the error tolerance (e.g., acceptable magnitude of difference) between the output settings and the stored copy of the input settings when checking that the input settings have been correctly applied to the camera.

[0007] However, this existing approach is ineffective when numerous different settings sets are used, as maintaining a copy of each setting set requires additional memory and computation time and is therefore not scalable to a larger number of settings. A larger number of settings is used, for example, by continuous autoexposure algorithms that dynamically generate different sensor settings based on changes in the input image. Therefore, the effectiveness of existing approaches decreases when the number of different settings exceeds a certain number.

[0008] Furthermore, existing approaches use a fixed error tolerance (e.g., a fixed percentage threshold) to determine whether the output settings match the input settings. In existing approaches, the stored copy of the input settings is linked or modified with a specific sensor quantization error that corresponds to the input settings. However, when numerous different sets of settings are used, the error tolerance (e.g., in percent) will be different for each different set of sensor settings, since different settings may have different amounts of quantization error introduced by the sensor. Continuously generated sensor settings, such as those used in continuous autoexposure, are subject to variable quantization over a range of sensor gain and exposure.It is impractical to adjust the error tolerance for each different set of sensor settings to account for the varying sensor quantization error. For example, a fixed set of settings might contain five settings, and a fixed error tolerance, such as 5%, might be associated with the five settings. However, for continuously generated sensor settings, the percentage error tolerance changes with the sensor settings. For example, lower gain and exposure settings might have a smaller error tolerance, while higher gain and exposure settings might have a larger error tolerance. Therefore, the fixed tolerance is ineffective for the range between the lower settings and the higher settings.The fixed-tolerance approach is therefore ineffective when settings change over time, as is the case with continuous autoexposure to adjust sensor settings. Therefore, existing approaches are inadequate to verify, within an acceptable timeframe, whether numerous sensor settings, such as those used in continuous autoexposure, have been correctly applied.

[0009] There is therefore a need for more effective techniques to verify the correct application of sensor settings, e.g. in autonomous or semi-autonomous systems. SUMMARY

[0010] The invention is defined by the claims. To illustrate the invention, aspects and embodiments are described herein, which may or may not fall within the scope of the claims.

[0011] Embodiments of the present disclosure relate to a technique for verifying sensor settings. The technique includes receiving sensor output data and at least one output sensor setting associated with the sensor output data from a sensor. The technique further includes determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting. The technique also includes, in response to determining that the queue of pending input sensor settings contains the matching input sensor setting, identifying a pending input sensor setting in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue.The technique further includes determining whether a sensor adjustment error has occurred based on a count of outstanding requests associated with the outstanding input adjustment. The technique further includes generating an error message based on the number of outstanding requests.

[0012] A technical advantage of the described techniques compared to previous solutions is the ability to detect a failure in the application of sensor settings when numerous different settings are applied, as is the case with continuous automatic exposure. Another technical advantage of the described techniques is the ability to identify and report the failure mode. The failure modes that can be detected include, among others, data corruption, lost settings that were sent but not used by the sensor, or lost output frames that were not received by the sensor. The detected failure mode can be used to find the cause of the failure and take corrective action. These technical advantages represent one or more technological improvements over existing approaches.

[0013] The disclosure extends to all novel aspects or features described and / or illustrated herein.

[0014] Further features of the disclosure are characterized by the independent and dependent claims.

[0015] Any feature of one aspect of the disclosure may be applied to other aspects of the disclosure, in any suitable combination. In particular, method aspects may be applied to device or system aspects, and vice versa.

[0016] Furthermore, functions implemented in hardware may also be implemented in software, and vice versa. Any reference to software and hardware functions herein should be interpreted accordingly.

[0017] Any system or device feature described here may also be provided as a method feature, and vice versa. System and / or device aspects described functionally (including means plus functional features) may alternatively be expressed by their corresponding structure, e.g., by an appropriately programmed processor and associated memory.

[0018] It should also be appreciated that certain combinations of the various features described and defined in each aspect of the disclosure may be implemented and / or provided and / or used independently of one another.

[0019] The disclosure also provides computer programs and computer program products comprising software code that, when executed on a data processing device, is adapted to perform any of the methods described herein and / or embody any of the apparatus and system features, including any or all substeps of any method.

[0020] The disclosure also provides a computer or computer system (including networked or distributed systems) having an operating system that supports a computer program for performing the methods described herein and / or for embodying the device or system features described herein.

[0021] The disclosure also provides a computer-readable medium on which one or more of the aforementioned computer programs are stored.

[0022] The disclosure also provides a signal carrying one or more of the aforementioned computer programs.

[0023] The disclosure extends to methods and / or devices and / or systems as described herein with reference to the accompanying drawings.

[0024] Aspects and embodiments of the disclosure will now be described, by way of example only, with reference to the accompanying drawings. SHORT DESCRIPTION OF THE DRAWING

[0025] The present systems and methods for secure detection of data changes in autonomous or semi-autonomous systems and applications are described in detail below with reference to the attached figures, where: Fig. 1 illustrates a computing device configured to implement one or more aspects of various embodiments; Fig. 2 is a more detailed representation of the sensor fault detector of Fig. 1, according to various embodiments; Fig. 3A illustrates a flowchart of a method for processing setup frames containing initialization data from a sensor in an initialization phase, according to various embodiments; Fig. 3B illustrates a flowchart of a method for processing setup frames containing sensor detection data in a runtime phase, according to various embodiments; Fig. 4 illustrates a flowchart of a method for queuing sensor settings into a settings history queue according to various embodiments; Fig. 5A and Fig. 5B show an example of the execution of an initialization phase of a sensor fault detector in fault-free operation of a sensor system according to some embodiments of the present disclosure Fig. 5C shows an example of executing an initialization phase of a sensor fault detector when corruption of sensor settings occurs, in accordance with some embodiments of the present disclosure. Fig. 6A shows an example of executing a runtime phase of a sensor fault detector when corruption of sensor settings occurs, in accordance with some embodiments of the present disclosure. Fig. 6B shows an example of executing a runtime phase of a sensor fault detector when a delay in sensor adjustments occurs, in accordance with some embodiments of the present disclosure. Fig. 6C shows an example of executing a runtime phase of a sensor failure detector when an output sensor frame is lost, in accordance with some embodiments of the present disclosure. Fig. 6D illustrates an example of executing a runtime phase of a sensor failure detector when sensor settings are lost, in accordance with some embodiments of the present disclosure; Fig. 7A is an illustration of an example of an autonomous vehicle in accordance with some embodiments of the present disclosure Fig. Figure 7B is an example of camera positions and fields of view for the autonomous vehicle from Fig. 7A, in accordance with some embodiments of the present disclosure; Fig. 7C is a block diagram of an example system architecture for the example autonomous vehicle of Fig. 7A, in accordance with some embodiments of the present disclosure Fig. Figure 7D is a system diagram for the communication between the cloud-based server(s) and the autonomous example vehicle of Fig. 7A, in accordance with some embodiments of the present disclosure; Fig. 8 is a block diagram of an example computing device suitable for use in implementing some embodiments of the present disclosure; and Fig. 9 is a block diagram of an example data center suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0026] Systems and methods for detecting disturbances to sensor settings of an autonomous or semi-autonomous vehicle or machine are described. Although the present disclosure is described with respect to an example of an autonomous or semi-autonomous vehicle or machine 700 (also referred to herein as "vehicle 700" or "ego-machine 700," see Fig. 7A-5D), this is not intended to be limiting. For example, the systems and methods described herein may be used, without limitation, by non-autonomous vehicles or machines, semi-autonomous vehicles or machines (e.g., in one or more adaptive driver assistance systems (ADAS)), autonomous vehicles or machines, guided and unguided robots or robotic platforms, warehouse vehicles, off-highway vehicles, vehicles coupled to one or more trailers, aircraft, boats, shuttles, emergency vehicles, motorcycles, electric or motorized bicycles, airplanes, construction vehicles, underwater vehicles, drones, and / or other types of vehicles. Although the present disclosure is described with respect to monitoring sensor performance in autonomous and / or semi-autonomous vehicles, this is not intended to be limiting.The systems and methods described herein may also be used in augmented reality, virtual reality, mixed reality, robotics, security and surveillance, autonomous or semi-autonomous machine applications, and / or any other technology area where sensor monitoring may be used.

[0027] Fig. 1 illustrates a computing device 100 configured to implement one or more aspects of various embodiments. In at least one embodiment, computing device 100 includes a desktop computer, a laptop computer, a smartphone, a personal digital assistant (PDA), a tablet computer, a server, one or more virtual machines, an embedded system, an embedded hardware module comprising a system on a chip, a system on a chip, a computing system of an autonomous, semi-autonomous, or non-autonomous machine, and / or any other type of computing device configured to receive input, process data, and optionally display images, and suitable for performing one or more embodiments.Computing device 100 includes memory 116, one or more processors 102, an interconnect 112, storage 114, an input / output (I / O) device interface 104, and a network interface 106. Computing device 100 also includes and / or communicates with one or more I / O devices 108. The I / O device(s) 108 communicate with the interconnect 112 via the I / O device interface 104. Memory 116 may be volatile random access memory or another suitable memory type.

[0028] In certain embodiments, a sensor fault detector 122 is stored in memory 116. The sensor fault detector 122 detects and reports errors in the application of sensor settings to sensors such as cameras, LIDAR sensors, RADAR sensors, or other types of sensors that can be configured by applying settings. To configure a sensor with an input setting, a system, such as a control system or a perception system for an autonomous or semi-autonomous machine, generates the input setting, e.g., based on environmental features such as lighting conditions, and requests that the input setting be applied to the sensor to cause the sensor to operate according to the input setting. For example, a control system may send a sensor configuration request containing the input setting to the sensor.If no interference occurs, the input setting is applied to the sensor and the sensor operates according to one or more values ​​contained in the input setting. Applying the input setting can, for example, mean configuring the sensor to use the input setting. The sensor then generates one or more output sensor frames. Each output sensor frame contains sensed data generated by the sensor based on the stimuli received by the sensor. For example, if the sensor is a camera, the sensed data may include images of the environment captured by the sensor. If the sensor is a RADAR or LIDAR sensor, the sensed data may also include a representation of detections, such as a point cloud or distances and directions to detected objects.Each output sensor frame also contains an output setting, which indicates which setting the sensor used to acquire the acquired data. The output setting may be different from the input setting, as the sensor may adjust the input setting before acquiring the data. For example, a setting value in the input setting may be rounded to different "quantized" values ​​by the sensor because only a limited number of bits are available to represent the setting value. The sensor uses the quantized values ​​for the setting value, so the quantized value is provided by the sensor in the output setting, which represents the setting used to generate the sensor output.

[0029] The sensor fault detector 122 monitors the input settings sent to the sensor from a sensor controller or other system component. For example, the sensor fault detector 122 may receive each input setting sent to the sensor from a sensor controller or other system component that configures the sensor. The sensor fault detector 122 also monitors the output settings received with each output data frame generated by the sensor, indicating the settings the sensor used to generate the output data frame. For example, the sensor fault detector 122 may receive each output data frame generated by the sensor and extract the output setting from each output data frame.

[0030] The sensor fault detector 122 stores each received input setting in a sensor setting queue 124. The sensor setting queue 124 is stored in memory 116 and may be a first-in, first-out data structure containing elements. Each element may be a setting or a data structure containing, for example, an input setting and a counter. Each element may be implemented as a reference (e.g., a pointer) to an input setting or a data structure containing an input setting and a counter. The input setting that has been in the queue the longest is called the "head" of the queue. The input setting that has been in the queue the shortest time is called the "tail" of the queue.The sensor fault detector 122 uses the sensor setting queue 124 to detect errors that occur between sending an input setting to the sensor and receiving an output setting from the sensor. The sensor fault detector 122 identifies certain failure modes (e.g., types of errors) based on the contents of the sensor setting queue 124. Failure modes that may be identified include sensor settings being lost or corrupted during communication between the transmitter (e.g., a sensor controller) and the sensor, sensor settings being corrupted within the sensor, a frame captured by the sensor being lost en route to the sensor fault detector 122, sensor data in a captured frame being corrupted en route to a receiver (e.g., a sensor controller or the sensor fault detector 122), and / or a delay in sensor settings taking effect within the sensor.

[0031] After receiving an input setting, the sensor fault detector 122 stores the input setting in the setting queue 124, e.g., by adding the input setting to the end of the queue, and sets a counter associated with the input setting to a value indicating how many times the same setting has been consecutively sent to the sensor. The sensor fault detector 122 updates the counter as input settings are sent and output settings are received, so that the counter represents the difference between the number of consecutive times the input setting has been sent to the sensor and the number of times an output setting matching the input setting has been received. For example, the first time the setting is sent to the sensor, the counter is set to 1 (or another value indicating that the setting has been sent once).Each time the same input setting is sent to the sensor without another input setting being sent in between, the counter is incremented by 1 (or by another value indicating that the same setting has been sent one more time).

[0032] Upon receiving an output frame containing an output setting, the sensor fault detector 122 searches for the first, e.g., oldest, input setting in the queue of settings that matches the received output setting within a threshold tolerance. The threshold tolerance takes into account any quantization error introduced by the sensor and may be based on the value of the setting to be compared. The threshold tolerance may be determined using a series of setting value ranges, each associated with a different threshold tolerance. The particular threshold tolerance to be used to determine whether a particular setting value (e.g., gain) in an input setting matches a corresponding setting value in the output setting may be determined by finding the range containing the particular setting value.The threshold tolerance associated with the identified range is the specific threshold tolerance to be used for the specific setting value. For example, five ranges can be used for the gain setting, each with a value between 1 and 10. A tolerance T1 is associated with a range between 1 and 2, represented by the interval [1, 2]. A tolerance T2 is associated with a range between 2 and 4, represented by the interval [2, 4]. A tolerance T3 is associated with a range between 4 and 6, represented by the interval [4, 6]. A tolerance T4 is associated with a range between 6 and 8, represented by the interval [6, 8]. A tolerance T5 is associated with a range between 8 and 10, represented by the interval [8, 10]. In this example, an input setting with a gain of 3.5 falls within the range associated with tolerance T2.Tolerance can be represented as a tolerance value. In this case, an input setting value matches an output setting value if the absolute value of the difference between the input setting value and the output setting value is smaller than the tolerance value. In another example, tolerance can be represented as a percentage. If the tolerance T2 is 10%, the input setting with a gain of 3.5 matches a received output setting with a gain of 3.4, because 3.4 is within 10% of 3.5. However, an input setting with a gain of 3.5 does not match the received output setting with a gain of 3.1, because 3.1 is not within 10% of 3.5.Although the tolerance determination is described in the examples using a specific determination technique, the tolerance can be determined based on the input setting and / or the output setting using any suitable determination technique.

[0033] If the sensor fault detector 122 does not find a matching input setting in the sensor setting queue 124, a setting error has occurred. If a matching input setting is found, the sensor fault detector 122 decrements the matching input setting counter by 1. If the matching input setting is at the beginning of the queue, no input setting error has occurred. Otherwise, the matching input setting is not at the beginning of the queue, and any non-matching setting that precedes the matching setting is removed from the queue.

[0034] The counter for each mismatched setting is compared to 0 (or another value equal to 0) to determine if an error has occurred. If the mismatched setting counter is 0, no error has occurred. If the mismatched setting counter is less than 0, an error has occurred, and the type of error is determined from the counter of the next setting in the queue. If the next setting counter is zero, a setting update delay has occurred. If the next setting counter is zero, then a sensor settings failure has occurred. If the mismatched setting counter is greater than zero, an error has occurred, and the type of error is determined from the counter of the previous setting in the queue.If the previous setting counter is zero, a frame loss has occurred. If the previous setting counter is less than zero, a loss of sensor settings has occurred.

[0035] Error information describing the error mode and the associated input setting and / or output sensor frame can then be reported to inform an application or other system component that an error occurred in verifying sensor data. The error information may indicate, for example, that incorrect sensor data was received or that expected sensor data was not received. In various embodiments, prior to performing the operations described above, an initial verification phase is used to review and verify an initial sequence of settings generated by the sensor when the sensor is initialized.

[0036] It is noted that the computing device described herein is illustrative, and that all other technically feasible configurations are included within the scope of the present disclosure. For example, multiple instances of the sensor fault detector 122 may be executed on a number of nodes in a distributed and / or cloud computing system to implement the functionality of the computing device 100. Alternatively, the computing device 100 may be implemented similarly to the computing device of the exemplary autonomous or semi-autonomous machine 700, which may be used at least in the Fig. 7A-7D.

[0037] In at least one embodiment, computing device 100 includes, without limitation, an interconnect (bus) 112 connecting one or more processors 102, an input / output (I / O) interface 104 connected to one or more input / output (I / O) devices 108, memory 116, storage 114, and / or a network interface 106. The processor(s) 102 may comprise any suitable processor embodied as a central processing unit (CPU), graphics processing unit (GPU), application specific integrated circuit (ASIC), field programmable gate array (FPGA), artificial intelligence (AI) accelerator, deep learning accelerator (DLA), a parallel processing unit (PPU), a data processing unit (DPU), a vector or image processing unit (VPU), a programmable image processing accelerator (PVA), any other type of processing unit, or a combination of different processing units, such asa CPU(s) configured to operate in conjunction with a GPU(s). In general, the processor(s) 102 may comprise any technically feasible hardware unit capable of processing data and / or executing software applications. For the purposes of this disclosure, the computing elements depicted in computing device 100 may correspond to a physical computing system (e.g., a system in a data center or a machine) and / or a virtual computing instance executing in a computing cloud.

[0038] In at least one embodiment, I / O devices 108 include devices that can receive input, such as a keyboard, a mouse, a touchpad, a VR / MR / AR headset, a gesture recognition system, a steering wheel, mechanical, digital, or touch-sensitive buttons or input components, and / or a microphone, as well as devices that can provide output, such as a display device, a haptic device, and / or a speaker. In addition, I / O devices 108 may include devices that can both receive input and provide output, such as a touchscreen, a Universal Serial Bus (USB) port, etc. I / O devices 108 may be configured to receive various types of input from an end user (e.g., a designer) of computing device 100 and also provide various types of output to the end user of computing device 100, such as displayed digital images or digital videos or text.In some embodiments, one or more of the I / O devices 108 are configured to connect the computing device 100 to a network 110.

[0039] In at least one embodiment, network 110 is any technically feasible type of communications network that enables the exchange of data between computing device 100 and internal, local, remote, or external units or devices, such as a web server or other networked computing device. Network 110 may be, for example, a wide area network (WAN), a local area network (LAN), a wireless network (e.g., Wi-Fi), and / or the Internet, to name a few.

[0040] In at least one embodiment, storage 114 includes non-volatile memory for applications and data and may include fixed or removable disk drives, flash memory devices, and CD-ROM, DVD-ROM, Blu-ray, HD-DVD, or other magnetic, optical, or solid-state storage devices. The processing engine 122 and / or the analysis engine 124 may be stored in storage 114 and loaded into memory 116 upon execution.

[0041] In one embodiment, memory 116 includes a random access memory (RAM) module, a flash memory device, and / or another type of storage device, or a combination thereof. Processor(s) 102, I / O device interface 104, and network interface 106 may be configured to read and write data to memory 116. Memory 116 may contain various software programs or general-purpose software code executable by processor(s) 102, as well as application data associated with the software programs, including processing engine 122 and / or analytics engine 124.

[0042] Fig. Figure 2 is a more detailed illustration of the sensor fault detector 122 of Fig. 1, according to various embodiments. As shown, the sensor fault detector 122 is included in a sensor interface 202. The sensor interface 202 is used by a sensor setting generator 204 to send an input setting 260 to a sensor 212 and by a sensor frame end user 236 to receive an output sensor frame 216. For purposes of explanation, the term "input setting" (singular) is used herein to refer to a set of one or more setting values ​​that may be sent to the sensor 212 as a group, e.g., in a single configuration request. The term "input settings" (plural) refers herein to two or more sets of setting values. For example, two input settings 260 may be sent to a sensor 212 at different times.The sensor setting generator 204 and the sensor frame end user 236 may be components of a sensor controller, an application, or another system component (not shown).

[0043] The sensor interface 202 includes an input setting processor 206 that processes each input setting 260 by adding the input setting 260 to the sensor setting queue 124. The input setting processor 206 includes an input setting receiver 208 that receives the input setting 260 from the sensor setting generator 204. The input setting processor 206 also includes a setting enqueuer 210 that enqueues the received input setting 260 into a sensor setting queue 124.

[0044] The input setting processor 206 also sends the input setting 260 to a sensor 212 with a request to configure the sensor 212 according to the input setting 260. The sensor 212 receives the input setting 260 and applies the input setting 260 to the configuration of the sensor 212. The result of applying the input setting 260 to the sensor configuration is represented as the applied setting 214. The applied setting 214 may differ from the input setting 260, for example, if the sensor rounds the values ​​of a setting to a larger or smaller value compatible with the sensor, or if a disturbance (for example, data corruption) occurs during the configuration of the sensor.

[0045] After receiving an input setting 260, the input setting processor 206 stores the input setting 260 in the sensor setting queue 124, e.g., by adding the input setting to the end of the queue, and sets a counter associated with the input setting to a value indicating how many times the same setting has been consecutively sent to the sensor. For example, the setting enqueuer 210 may add the input setting 260 to the sensor setting queue 124 as a tail element 226N. The tail element 226N may, e.g., be the input setting 260 or a data structure containing the input setting 260. The sensor setting queue 124 also includes a head element 226A and a next element 226B between the head element 226A and the tail element 226N. Each element 226 in the sensor setting queue 124 may be a different sensor setting.

[0046] Because the counter associated with each input setting in the sensor setting queue 124 indicates how many times the same setting has been consecutively sent to the sensor 212, separate input settings with the same setting values ​​are not consecutively stored in the sensor setting queue 124. The sensor fault detector 122 updates the counter associated with each input setting 260 as input settings 260 are sent to the sensor 212 and output settings 222 are received from the sensor 212, such that the counter associated with each input setting 260 in the sensor setting queue 124 represents a difference between the number of consecutive times the input setting 260 has been sent to the sensor 212 and the number of times an output setting 222 matching the input setting 260 has been received from the sensor 212.For example, the first time an input setting 260 is sent to the sensor 212, the counter for the input setting 260 is set to 1 (or another value indicating that the setting has been sent once). Each time the same input setting 260 is sent to the sensor without another input setting 260 being sent in between, the counter is incremented by 1 (or another value indicating that the same setting has been sent once more). Alternatively, in other embodiments, each input setting 260 may be stored in the sensor setting queue 124, even if the input setting 260 has the same setting values ​​as another, adjacent input setting 260 in the queue. In this case, the counter need not be maintained, as the number of consecutive occurrences of the input setting in the input setting 260 queue may represent the count.

[0047] The sensor 212 generates a sequence of output sensor frames 216 in response to external stimuli such as light or other electromagnetic waves. Each output sensor frame 216 may contain sensed data 220 generated by the sensor 212 based on external stimuli according to the applied setting 214. The sensed data 220 may be, for example, a camera image, a point cloud, or other suitable representation of the data sensed by the sensor 212. The output sensor frame 216 also contains an output setting 222 containing one or more setting values ​​that the sensor 212 uses to acquire the sensed data 220. The output setting 222 may be different from the applied setting 214.For example, the sensor 212 may round the one or more setting values ​​in the applied setting 214 to corresponding quantized values ​​due to the limited number of bits available to represent the setting values ​​in the sensor 212. The sensor uses the quantized values ​​for the settings when acquiring the acquired data 220, so that the quantized values ​​are provided by the sensor as an output setting 222, which represents the settings used in generating the acquired data 220.

[0048] The captured frame processor 228 detects and reports failures that occur when sending or applying input settings 260 to the sensor 212 or when receiving output settings 222 from the sensor 212. The captured frame processor 228 includes a sensor frame receiver 230 that receives each output sensor frame 216 from the sensor 212 and forwards the output sensor frame 216 to an output setting analyzer 234. Upon receiving an output sensor frame 216, the output setting analyzer 234 searches the sensor setting queue 124 for the first, e.g., oldest, queue item 226 that contains an input setting that matches the output setting 222 contained in or associated with the output sensor frame 216.

[0049] In some embodiments, an input setting in the sensor setting queue 124 may match the output setting 222 if each setting value in the input setting is equal to a corresponding setting value for the same parameter in the output setting 222. For example, if the input setting includes the exposure parameter value of 25 milliseconds and the gain parameter value of 10, then the input setting matches the output setting 222 if the output setting also includes the exposure parameter value of 25 milliseconds and the gain parameter value of 10. In other embodiments, the input setting matches the output setting 222 if each setting value in the input setting is within a threshold tolerance of a corresponding setting value for the same parameter in the output setting 222.For example, an input setting containing an exposure parameter value of 25 milliseconds and a gain parameter value of 10 matches an output setting 222 containing an exposure parameter value of 26 milliseconds and a gain parameter value of 11 within a tolerance threshold of 10%. The tolerance threshold takes into account the quantization error caused by the sensor and can be based on the value of the setting being compared.

[0050] If the output settings analyzer 234 does not find a matching input setting, the output settings analyzer 234 determines that a "setting corruption" fault has occurred and sends a failure mode identifier 242 to an error data generator 244. The error data generator 244 sends the error data 246, which identifies the failure mode 242 and the output sensor frame 216 for which no match was found, to an error handler 248. The error handler 248 may forward the error data 246 to a system component, such as a sensor controller or an application. The system component may then process the error based on the error data 246. For example, the error handler 248 determines that acquired data 220 for which a settings error occurred should be discarded without further processing.The system component then discards the captured data 220 associated with the output sensor frame 216.

[0051] If the output setting analyzer 234 finds a matching input setting in the sensor setting queue 124 that matches the output setting 222, the sensor fault detector 122 decrements the counter associated with the matching input setting by 1. If the matching input setting is at the head of the queue, then no error has occurred with the input setting, and the output setting analyzer 234 continues to search the sensor setting queue 124 for another input setting that matches the output setting 222. Otherwise, the matching input setting is not at the head of the queue, and a setting dequeuer 240 removes, e.g., delists, any non-matching input setting that precedes the matching setting from the sensor setting queue 124.

[0052] The output setting analyzer 234 compares the counter associated with each mismatched setting in the queue to 0 (or another value equivalent to 0) to determine if an error has occurred. If the counter for a mismatched setting in the queue is 0, no error has occurred, and the output setting analyzer 234 compares the next mismatched unlisted setting in the queue to 0. If the counter for each of the mismatched unlisted settings in the queue is 0, the output setting analyzer 234 determines that no error has occurred with the mismatched unlisted settings.

[0053] If the output setting analyzer 234 determines that the counter for a mismatched unlisted setting in the queue is less than 0, then a fault / error has occurred, and the output setting analyzer 234 identifies the particular type of fault based on the counter of the next setting in the queue. For example, the next setting may be the setting next to and after the mismatched enlisted setting in the queue. If the next setting counter is zero, a setting update delay has occurred. If the next setting counter is equal to zero, then a loss of sensor settings has occurred.

[0054] If the output setting analyzer 234 determines that the counter for a mismatched setting in the queue is greater than 0, a failure has occurred, and the output setting analyzer 234 identifies the type of error based on the counter of the previous setting in the queue. The previous setting can, for example, be the setting that is directly next to and before the mismatched delisted setting in the queue. If the previous setting counter is zero, then a frame loss has occurred. If the previous setting counter is less than zero, then a loss of sensor settings has occurred. The output setting analyzer 234 sends an error mode 242 indicating the identified error mode (e.g.,setting update delay, sensor setting loss, or frame loss) to the error data generator 244, which sends error data 246 specifying the identified error mode 242 to the error handler 248.

[0055] Fig. 3A shows a flowchart of a method 300 for processing setup frames containing initialization data from a sensor in an initialization phase, according to various embodiments. Each block of the method 300 described herein comprises a computational process that may be performed using any combination of hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in memory. The method may also be stored in the form of computer-usable instructions on a computer storage medium. The method may be provided as a standalone application, as a service or hosted service (standalone or in combination with another hosted service), or as a plug-in for another product, to name a few examples. Furthermore, the method 300 is exemplary for the system of Fig. 1-2. However, this method may additionally or alternatively be performed by any system or combination of systems, including, but not limited to, the systems described herein. Furthermore, the acts in method 300 may be omitted, repeated, and / or performed in any order without departing from the scope of the present disclosure.

[0056] As in Fig. 3, method 300 begins with operation 302, in which a sensor fault detector 122 receives an output setting from an output sensor frame. In operation 304, the sensor fault detector 122 determines whether a current execution phase of the sensor fault detector 122 is an initialization phase, in which each output sensor frame 216 contains an output setting 222 with an initialization setting value (also referred to herein as "Si"). Each output setting 222 during the initialization phase may have the same initial value. If the sensor fault detector 122 is not in the initialization phase, then in operation 306, the sensor fault detector 122 performs operations of a runtime phase by executing the flowchart 350 of Fig. 3B, and method 300 ends. If sensor fault detector 122 is in the initialization phase, then sensor fault detector 122 performs operation 308. In operation 308, sensor fault detector 122 determines whether a current frame number ("f") is less than or equal to an expected number of initial output sensor frames. The current frame number corresponds to the output sensor frame received in operation 302 and may be 1 for the first received frame, 2 for the second received frame, etc. Each initial output sensor frame is an output sensor frame 216 for which output setting 222 is a predetermined initial output setting, rather than an output setting 222 determined from input settings 260. The initial output setting may be a series of predetermined setting values, e.g., a predetermined exposure and a predetermined gain.The initial output setting is referred to herein as "Sinitial." If the sensor fault detector 122 determines in operation 308 that the current frame count f is less than or equal to the expected number M of output frames, then the method 300 ends. The method 300 is subsequently called again when a different output setting is received in a sensor output frame.

[0057] If the sensor failure detector 122 determines in operation 308 that the current frame number f is not less than or equal to the number M of initialization frames, then the sensor failure detector 122 determines in operation 310 whether the current frame number f is equal to the number of initialization frames plus 1 ("M+1"). If so, the sensor failure detector 122 performs operation 312.

[0058] In operation 312, the sensor fault detector 122 determines whether the output setting 222 of the current output sensor frame 216 is the initialization setting Sinitial. If so, the sensor fault detector 122 performs operation 318. In operation 318, the sensor fault detector 122 decrements a counter associated with a queue element with the initialization setting (e.g., a queue element corresponding to the initialization setting or a queue element containing the initialization setting). The queue element with the initialization setting may, for example, be the first (oldest) element found in the sensor setting queue 124 with the setting Sinitial. If the sensor failure detector 122 determines in operation 312 that the current output sensor frame 216 does not have the initialization setting, the sensor failure detector 122 reports an error in operation 316.

[0059] If the sensor failure detector 122 determines in operation 310 that the number of current frames is not equal to the number of initialization frames M plus 1, then the sensor failure detector 122 performs operation 314. In operation 314, the sensor failure detector 122 determines whether the counter of the queue element with the initialization setting (Sinitial) is greater than or equal to 0. If not, then the sensor failure detector 122 reports an error in operation 316.

[0060] If the sensor fault detector 122 determines in operation 314 that the counter of the next queue element with the initialization setting is greater than or equal to 0, then the sensor fault detector 122 determines in operation 320 whether the current frame from sensor 212 has the initialization setting. If so, the sensor fault detector 122 performs operation 318, which decrements the counter of the queue element with the initialization setting (Sinitial). Otherwise, operation 320 determines that the current frame does not have the initialization setting and instead has a runtime-determined output setting 222. The output setting 222 is determined by the sensor 212 based on an input setting 260. In operation 322, the sensor fault detector 122 removes any queue element with an initialization setting from the sensor setting queue 124.In operation 324, the sensor failure detector 122 sets the current execution phase of the sensor failure detector 122 to a runtime phase of execution to indicate that the frames with the initialization settings have been exhausted and frames with the output setting 222 determined at runtime can now be processed. In operation 326, the sensor failure detector 122 executes the runtime phase by executing flowchart 350. Fig. 3B.

[0061] Fig. 3B shows a flowchart of a method 350 for processing setting frames containing sensor detection data in a runtime phase, according to various embodiments. Each block of the method 350 described herein comprises a computational process that may be performed using any combination of hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in memory. The method may also be stored in the form of computer-usable instructions on a computer storage medium. The method may be provided as a standalone application, as a service or hosted service (standalone or in combination with another hosted service), or as a plug-in for another product, to name a few examples. Furthermore, the method 350 is exemplary for the system of Fig. 1-2. However, this method may additionally or alternatively be performed by any system or combination of systems, including, but not limited to, the systems described herein. Furthermore, the acts in method 350 may be omitted, repeated, and / or performed in any order without departing from the scope of the present disclosure.

[0062] As in Fig. 3B, the method 350 begins with operation 352 in which the sensor fault detector 122 determines whether the sensor adjustment queue 124 contains an item that is consistent with the operation 302 of the method 300 of Fig. 3A received output setting. Alternatively, in operation 352, the sensor failure detector 122 may determine whether the sensor setting queue 124 contains an item that matches an output setting received in an output sensor frame from any suitable source. If the sensor failure detector 122 determines in operation 350 that the sensor setting queue 124 does not contain an item that matches the output setting, then in operation 354 the sensor failure detector 122 generates an error message indicating frame or setting corruption. Otherwise, the sensor setting queue 124 contains an item that matches the output setting, and the sensor failure detector 122 performs operation 356.

[0063] In operation 356, the sensor failure detector 122 decrements a counter of the matching queue element, e.g., by subtracting 1 from the counter. In operation 358, the sensor failure detector 122 determines whether the matching queue element is at the head of the sensor setup queue 124. If so, the method 350 ends. Otherwise, the matching queue element is not at the head of the sensor setup queue 124, and the sensor failure detector 122 performs operation 360. In operation 360, the sensor failure detector 122 unlists each queue element that is in the queue before the matching queue element (e.g., that was added to the queue before the matching element). The sensor failure detector 122 performs operation 362 for each queue element that was unlisted from the queue in operation 360.

[0064] In operation 362, the sensor failure detector 122 determines whether the counter associated with the destacked queue element is less than 0. If so, in operation 364, the sensor failure detector 122 determines whether the value of the next queue element (e.g., the next element in the sensor settings queue 124) is 0, in which case the sensor failure detector 122 performs operation 366, or greater than 0, in which case the sensor failure detector 122 performs operation 370. In operation 366, the sensor failure detector 122 generates an error indicating that a sensor settings update delay has occurred. In operation 370, the sensor failure detector 122 generates an error message indicating that a sensor settings loss has occurred.

[0065] If the sensor failure detector 122 determines in operation 362 that the counter associated with the delisted queue element is less than 0, then the sensor failure detector 122 performs operation 372. In operation 372, the sensor failure detector 122 determines whether the counter associated with the element in the queue is greater than 0. If so, the sensor failure detector 122 performs operation 376. In operation 376, the sensor failure detector 122 determines whether the value of the counter associated with the previous queue element (e.g., the previous element in the sensor setup queue 124) is less than 0, in which case the / 12 performs operation 370, or equal to 0, in which case the sensor failure detector 122 performs operation 378. In operation 378, the sensor failure detector 122 generates an error indicating that an output sensor frame loss has occurred.The error(s) generated in operations 366, 370, and / or 376 may, for example, be error data 246 containing a failure mode identifier 242 indicating the type of error. Following operations 366, 370, or 376, sensor failure detector 122 performs operation 368, which determines whether there is another queue item in sensor setup queue 124 before the matching queue item. If so, sensor failure detector 122 re-executes operation 360 and subsequent operations. If not, method 350 terminates. If the sensor failure detector 122 determines in operation 372 that there is another queue element before the matching queue element in the sensor setup queue 124, the sensor failure detector 122 performs operation 360 and subsequent operations again.

[0066] Fig. 4 illustrates a flowchart of a method 400 for adding sensor settings to a settings history queue according to various embodiments. Each block of the method 400 described herein comprises a computational process that may be performed using any combination of hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in memory. The method may also be stored in the form of computer-usable instructions on a computer storage medium. The method may be provided as a standalone application, as a service or hosted service (standalone or in combination with another hosted service), or as a plug-in for another product, to name a few examples. Furthermore, the method 400 is exemplary for the system of Fig. 1-2. However, this method may additionally or alternatively be performed by any system or combination of systems, including, but not limited to, the systems described herein. Furthermore, the acts in method 400 may be omitted, repeated, and / or performed in any order without departing from the scope of the present disclosure.

[0067] As in Fig. 4, the method 400 begins with operation 402, in which the / 12 receives an input setting 260, e.g., from a sensor setting generator 204. In operation 404, the sensor fault detector 122 determines whether the input setting matches (e.g., is identical to) the previous input setting in the sensor fault detector 122. The previous input setting may be the setting most recently added prior to receiving the input setting in operation 402. If the input setting is determined to match the previous input setting in operation 404, the sensor fault detector 122 performs operation 406. In operation 406, the sensor fault detector 122 adds (e.g., enqueues) a queue item to the end of the sensor setting queue 124. The added queue item contains the input setting and is associated with a counter value of 1.If it is determined in operation 404 that the input setting does not match the previous setting, the sensor failure detector 122 increments the counter of the queue element at the end of the sensor setting queue 124 in operation 408.

[0068] Fig. 5A and Fig. 5B illustrate an embodiment 500 of an initialization phase of a sensor fault detector during normal operation of a sensor system in accordance with some embodiments of the present disclosure. In the example of a normal operation 500, an input setting sequence 502 generated by a sensor setting generator 204 is transmitted to a sensor fault detector 122. The input setting sequence 502 is also provided to a sensor 212, e.g., by the sensor fault detector 122 or the sensor fault detector 122. Each element of the input setting sequence 502 corresponds to an input setting 260. The input setting sequence 502 includes the elements Si, Si, Si, Si, S1, S2, S2, S3, each of which represents an input setting 260.Si (also referred to herein as Sinitial) represents a predefined initialization setting generated by the sensor setting generator 204 during an initialization phase, and S1, S2, S3 represent three different settings. The initialization setting Si is sent M=5 times to the sensor fault detector 122 (and also to the sensor 212) in this example. The initialization settings may be sent, for example, to prepare the sensor for operation. The sensor fault detector 122 receives and processes each element (e.g., each input setting 260) of the input setting sequence 502 during a corresponding period of time corresponding to one frame and updates the sensor setting queue 124 accordingly, as shown in FIG. Fig. 5A and Fig. 5B. As shown in Fig. 3A, it is assumed that the first M (e.g., 5) output frames received from sensor 212 have the initialization setting Si and can be discarded.

[0069] The sensor 212 generates an output setting sequence 504 in response to receiving the input setting sequence 502. Each element of the output setting sequence 504 corresponds to an output setting 222 of an output sensor frame 216. The output setting sequence 502 includes the elements Sr, Sr, Si, Si, Si, Si, S1, S2, S2, S3. The setting Sr corresponds to random or unknown data generated by the sensor setting generator 204 prior to initialization. Thus, the sequence of output settings 502 includes two occurrences of the random setting Sr, three occurrences of the initialization setting Si, one occurrence of a setting S1, two occurrences of a setting S2, and one occurrence of a setting S3. The sensor error detector 122 receives and processes each element (e.g.,each output setting 222) of the output setting sequence 504 during a corresponding period of time corresponding to one frame, and updates the sensor setting queue 124 accordingly, as shown in FIG. Fig. 5A and Fig. 5B shown.

[0070] The input setting sequence 502 is processed by the input setting enqueuer 210. An example of a method for processing the input settings is shown in Fig. 4. The output setting sequence 504 is processed by the output setting analyzer 234. An example of a method for processing the output settings is shown in Fig. 3A and Fig. 3B. The input setting sequence 502 may be processed in parallel with the output setting sequence 504.

[0071] In frame "1" 520-1, the sensor fault detector 122 receives the first of the five Si input settings. The sensor fault detector 122 queues the input setting into the sensor setting queue 522 if the input setting differs from the previous input setting. Since there is no previous setting in frame 1, the input setting Si is added to the queue, resulting in a queue 522-1 containing the setting Si. The sensor fault detector 122 also sets the counter for the setting Si to 1 because the input setting differs from the previous setting (as shown in Fig. 4). Queue 522-1 thus contains a single element Si with counter=1. The counter can be stored in queue 522-1 (e.g., in the same queue element as the setting Si) or separately from the queue. Furthermore, in frame 1 520-1, sensor fault detector 122 receives the first of the two Sr output settings from sensor 212. Since the frame number, which is 1, is less than or equal to the number of expected initialization frames M=5, sensor fault detector 122 performs no further action for frame 520-1.

[0072] In frame "2" 520-2, sensor failure detector 122 receives the second of the five Si input settings. Since Si is equal to the previous setting (the first Si from frame 1), sensor failure detector 122 increments the counter of the previously queued Si (which is at the end of the queue). Therefore, queue 522-2 contains the Si setting with counter value = 2. Also, in frame 2 520-2, sensor failure detector 122 receives the second of the two Sr output settings. Since frame number 2 is less than or equal to the number of expected output frames, M = 5, sensor failure detector 122 performs no further action for frame 2.

[0073] In frame "3" 520-3, sensor failure detector 122 receives the third of the five Si input settings. Since Si is equal to the previous setting (the second Si, from frame 2), sensor failure detector 122 increments the number of Si previously queued. Consequently, queue 522-3 contains the Si setting with the associated counter=3. Furthermore, in frame 3 520-3, sensor failure detector 122 receives the first of the five output Si settings. Since frame number 3 is less than or equal to the number of expected output frames M=5, sensor failure detector 122 performs no further action for frame 3.

[0074] In frame "4" 520-4, the sensor failure detector 122 receives the fourth of the five Si input settings. Since Si is equal to the previous setting (the third Si, from frame 3), the sensor failure detector 122 increments the number of previously queued Si. Consequently, the queue 522-4 contains the Si setting with the associated counter=4. Also, in frame 4 520-4, the sensor failure detector 122 receives the second of the five Si output settings. Since the number of frames with 4 is less than or equal to the number of expected initialization frames, M=5, the sensor failure detector 122 performs no further action for frame 4.

[0075] In frame "5" 520-5, sensor failure detector 122 receives the fifth of the five Si input settings. Since Si is equal to the previous setting (the fourth Si, from frame 4), sensor failure detector 122 increments the number of Si previously queued. Consequently, queue 522-5 contains the Si setting with the associated counter = 5. Also, in frame 5 520-5, sensor failure detector 122 receives the third of the five Si output settings. Since the number of frames with 5 is less than or equal to the number of expected initialization frames, M = 5, sensor failure detector 122 performs no further action for frame 5.

[0076] In frame "6" 520-6, the sensor fault detector 122 receives the input setting S1. Since S1 does not match the previous setting (the fifth Si, from frame 5), the sensor fault detector 122 inserts the input setting S1 with counter=1 into the settings queue 522-6. Therefore, the queue 522-6 contains the setting Si with the corresponding counter=5 as the first element (at the beginning of the queue) and the setting S1 with the corresponding counter=1 as the second element (at the end of the queue). In addition, in frame 6 520-6, the sensor fault detector 122 receives the fourth of the five Si output settings. Since the frame number 6 is equal to the number of expected initial frames M=5 plus 1 (see operation 310 in Fig. 3A) and the current setting of the output frame is Si (see operation 312), the sensor fault detector 122 performs the action of decrementing the counter of the queue element with the setting = Si (see operation 318). After executing this action, the queue 524-6 contains the setting Si with counter=4 at the beginning of the queue. Since the current setting of the received output frame is Si (see operation 312), no error occurs in frame 6 (status: OK).

[0077] In frame "7" 520-7, the sensor fault detector 122 receives the first of the two S2 input settings. Since S2 is not equal to the previous setting (S1), the sensor fault detector 122 inserts the input setting S2 with counter=1 into the settings queue 522-7. Thus, the queue 522-7 contains the setting Si with counter=4 as the first element (at the head of the queue), the setting S1 with counter=1 as the second element, and the setting S2 with counter=1 as the third element (at the tail of the queue). Also, in frame 7 520-7, the sensor fault detector 122 receives the fifth of the five Si output settings. Since the frame number 7 is greater than or equal to the number of expected initial frames M=5 plus 2 (in operation 310 of Fig. 3A), the counter of the queue element with the setting Si is greater than or equal to 0 (see operation 314) and the current input frame setting is Si (see operation 320), the sensor fault detector 122 performs the action of removing each queue element with the setting = Si from the queue (see operation 320). After performing this action, the queue 524-7 contains the setting Si with the associated counter = 3 at the beginning of the queue. Since the current setting of the received output frame is Si (see operation 312), no error occurs in frame 6 (status: OK).

[0078] In frame "8" 520-8, the sensor fault detector 122 receives the second of the two S2 input settings. Since S2 is equal to the previous setting (S2), the sensor fault detector 122 increments the counter corresponding to S2 in the queue 522-8. As a result, the queue 522-8 contains the setting Si with counter=3 as the first element (at the beginning of the queue), the setting S1 with counter=1 as the second element, and the setting S2 with counter=2 as the third element (at the end of the queue). In addition, in frame 8 520-8, the sensor fault detector 122 receives the S1 output setting. Since the frame number 8 is greater than or equal to the number of expected initial frames M=5 plus 2 (in operation 310 of Fig. 3A), the counter of the queue element with the setting Si is greater than or equal to 0 (see operation 314) and the current input frame setting (S1) is not equal to Si (see operation 320), the sensor fault detector 122 performs the action of removing the queue element with the setting = Si from the queue (see operation 322) and decrementing the counter associated with the current input frame setting (S1). The sensor fault detector 122 also sets the current phase indicator to "Runtime" (operation 324) so ​​that subsequent input and output settings are determined by the runtime phasing method of Fig. 3B are processed. Following the execution of the action, setting queue 524-8 contains setting S1 with counter=0 at the head of the queue and setting S2 with counter=2 at the tail of the queue. No error occurs in frame 6 (status: OK).

[0079] In frame "9" 520-9, the sensor fault detector 122 receives the input setting S3. Since S3 is not equal to the previous setting (S2), the sensor fault detector 122 adds S3 to the settings queue 522-9. Thus, the queue 522-9 contains the setting S1 with counter=0 as the first element (at the head of the queue), the setting S2 with counter=2 as the second element, and the setting S3 with counter=1 as the third element. Furthermore, in frame 9 520-9, the sensor fault detector 122 executes the runtime phase method 350 of Fig. 3B instead of the initialization phase method 300 of Fig. 3A because the current phase indicator was set to "runtime" in frame 8. Accordingly, sensor failure detector 122 receives output setting S2 and determines whether queue 522-9 contains an element with a setting that matches the received output setting S2 (see operation 352). Since queue 522-9 contains S2, sensor failure detector 122 decrements the counter of S2 (see operation 356). Sensor failure detector 122 determines whether the matching queue element (S2) is at the head of the queue (see operation 358). Since S2 is not at the head of the queue, sensor failure detector 122 removes any element from the queue that precedes the matching queue element (see operation 360). Sensor failure detector 122 compares the counter of the removed element S2 to 0 (see operations 362 and 372). Since the counter is zero, no error messages are generated for frame 9 (status: OK).After the action has been performed, queue 524-8 contains the setting S2 with counter=1 at the beginning of the queue and the setting S3 with counter=1 at the end of the queue.

[0080] Fig. 5C shows an embodiment 501 of an initialization phase of a sensor fault detector, wherein corruption of sensor settings occurs, in accordance with some embodiments of the present disclosure. In embodiment 501, an input setting sequence 506 is transmitted to a sensor fault detector 122 and to a sensor 212. The input setting sequence 506 corresponds to the input setting sequence 502 of Fig. 5A and contains the elements Si, Si, Si, Si, Si, S1, S2, S2, S3. The output sequence 508 contains the elements Sr, Sr, Si, Si, Si, S100, S1, S2, S2, S3. Frames "1" 530-1 to "5" 530-5 are the same as frames "1" 520-1 to "5" 520-5 of the Fig. 5A and Fig. 5B. In frame "6" 530-6, the sensor fault detector 122 receives the input setting S1. The sensor fault detector 122 queues the input setting S1 in the sensor setting queue 532-6 because S1 is different from the previous input setting (Si). The sensor fault detector 122 also sets the counter for the setting S1 to 1. The queue 532-6 therefore contains an element Si with counter=5 (at the beginning of the queue) and an element S1 with counter=1. In addition, in frame 6 530-6, the sensor fault detector 122 receives the output setting S100 from the sensor 212. The sensor fault detector 122 searches the queue 532-6 for the received setting S100 (see operation 352 in Fig. 3B). Since queue 532-6 does not contain setting S100, sensor fault detector 122 generates a setting corruption error message (see operation 354).

[0081] Fig. 6A shows an embodiment 600 of a runtime phase of a sensor fault detector when corruption of sensor settings occurs, in accordance with some embodiments of the present disclosure. In embodiment 600, the input setting sequence 602 includes elements S-1, S0, S1, S2, S3, S4 (where S-1 represents S-1). The output setting sequence 604 includes elements S-1, S0, S100, S2, S3, S4.

[0082] In frame N 620-1, sensor failure detector 122 receives input setting S1. Since S1 does not match the previous setting, sensor failure detector 122 inserts input setting S1 with counter=1 into setting queue 622-1. Thus, queue 522-6 contains setting S-1 with counter=1 (at the beginning of the queue), setting S0 with counter=1, and setting S1 with counter=1. Also, in frame N 620-1, sensor failure detector 122 receives output setting S-1. Sensor failure detector 122 executes in the runtime phase. Sensor failure detector 122 finds a queue element that matches output setting S-1 (see operation 352) and decrements the counter of the matching queue element (see operation 356). Since the matching queue element is at the beginning of the queue, no dequeuing is performed and no error occurs in frame N (status: OK).After the action is executed, the queue 624-1 contains the setting S-1 with counter=0 at the head of the queue, the setting S0 with counter=1 in the second queue element and the setting S1 with counter=1 in the third queue element.

[0083] In frame N+1 620-2, sensor failure detector 122 receives input setting S2. Since S2 is not equal to the previous setting, sensor failure detector 122 adds input setting S2 with counter=1 to setting queue 622-2. Therefore, queue 622-2 contains setting S-1 with counter=0 (at the head of the queue), setting S0 with counter=1, setting S1 with counter=1, and setting S2 with counter=1. Also, in frame N+1 620-2, sensor failure detector 122 receives output setting S0. Sensor failure detector 122 executes in the runtime phase. Sensor failure detector 122 finds a queue element that matches output setting S0 (see operation 352) and decrements the counter of the matching queue element (see operation 356).Since the matching queue element is not at the head of the queue, each element that precedes the matching element S0 is dequeued (see operation 360), and the counter of each queue element that has been dequeued is compared to 0. Since the setting S-1 precedes S0 in the queue, the setting S-1 is 0. Since the counter of the dequeued queue element S-1 is 0, no error occurs in frame N (status: OK). After executing the action, queue 624-2 contains the setting S0 with counter=1 in the head element, the setting S1 with counter=1 in the second element, and the setting S2 with counter=1 in the third element. Since the matching queue element S0 is at the head of the queue, no dequeuing is performed and no error occurs in frame N (status: OK).

[0084] In frame N+2 620-3, the sensor failure detector 122 receives the input setting S3. Since S3 does not match the previous setting, the sensor failure detector 122 inserts the input setting S3 with counter=1 into the setting queue 622-3. Thus, the queue 622-3 contains the setting S0 with counter=1 (at the beginning of the queue), the setting S1 with counter=1, the setting S2 with counter=1, and the setting S3 with counter=1. Furthermore, in frame N+2 620-3, the sensor failure detector 122 receives the output setting S100. The sensor failure detector 122 searches the queue 622-3 for a queue element that matches the output setting S100, but the setting S100 is not found because it does not exist in the queue. Since the matching queue element is not found, the sensor failure detector 122 generates a "setting corruption" error message (see operation 354).

[0085] Fig. 6B shows an embodiment 605 of a runtime phase of a sensor fault detector when a sensor setting update delay occurs, in accordance with some embodiments of the present disclosure. In the example embodiment 605, the input setting sequence 606 includes the elements S-1, S0, S1, S2, S3, S4, S5 (where S-1 represents S-1). The sequence of output settings 608 includes the elements S-1, S0, S1, S1, S2, S3. The setting S2 has been delayed so that S1 appears twice in the received setting sequence 608. The frames N 630-1 and N+1 630-2 of Fig. 6B are the same as frame N 620-1 and N+1 620-2 of Fig. 6A.

[0086] In frame N+2 630-3, the sensor fault detector 122 receives the input setting S3. Since S3 is not equal to the previous setting (S2), the sensor fault detector 122 inserts the input setting S3 with counter=1 into the setting queue 632-3. Thus, the queue 632-3 contains the setting S0 with counter=0 (at the beginning of the queue), followed by the settings S1, S2, and S3, each with counter=1. Furthermore, in frame N+2 630-3, the sensor fault detector 122 receives the first of the two S1 output settings. The sensor fault detector 122 finds a queue element that matches the output setting S1 and decrements the counter of the matching queue element S1. Since the matching queue element is not at the head of the queue, the sensor failure detector 122 removes any element (S0) that precedes the matching queue element.The sensor failure detector 122 compares the counter of the element S0 that is in the queue with 0. Since the counter is zero, no error messages are generated for frame N+1 630-3 (status: OK).

[0087] In frame N+3 630-4, the sensor fault detector 122 receives the input setting S4. Since S4 does not match the previous setting (S3), the sensor fault detector 122 adds the input setting S4 to the queue 632-4 with counter=1. Furthermore, in frame N+3 630-3, the sensor fault detector 122 receives the second S1 output setting. The sensor fault detector 122 finds a queue element that matches the output setting S1 and decrements the counter of the matching queue element S1. This decrement sets the counter associated with S1 to -1. Since the matching queue element is at the beginning of the queue, the sensor fault detector 122 does not remove any elements from the queue in frame N+3 630-4, and no errors are generated in frame N+3 630-4.

[0088] In frame N+4 630-5, the sensor fault detector 122 receives the input setting S5. Since S5 does not match the previous setting (S4), the sensor fault detector 122 adds the input setting S5 to the queue 632-5 with counter=1. Furthermore, in frame N+4 630-3, the sensor fault detector 122 receives the second output setting S2. The sensor fault detector 122 finds a queue element that matches the output setting S2 and decrements the counter of the matching queue element S2. Decrementing sets the counter for S2 to 0. Since the matching queue element S2 is not at the head of the queue, the sensor fault detector 122 removes from the queue any element (S1) that precedes the matching queue element S2. The sensor failure detector 122 compares the counter of the destacked element S1 with 0.Since the counter of S1 is less than 0 and the counter of the next queue element S2 is 0, the sensor failure detector 122 determines that a setting update delay has occurred (see operation 366 in . Fig. 3B).

[0089] Fig. 6C illustrates an embodiment 607 of a runtime phase of a sensor failure detector when a loss of an output sensor frame occurs, in accordance with some embodiments of the present disclosure. In the example embodiment 607, the input setup sequence 610 includes elements S-1, S0, S1, S1, S2, S3, S4 (where S-1 represents S-1). The output setup sequence 612 includes elements S-1, S0, S1, S2, S3, S4. The second S1 is not present in the output setup sequence 612, so an error has occurred in the output setup. Frame N 640-1 in Fig. 6B is the same as frame N 630-1 in Fig. 6B.

[0090] In frame N+1 640-2, sensor failure detector 122 receives input setting S1. Since S1 is equal to the previous setting (S1), sensor failure detector 122 increments the counter of S1 in queue 642-2, resulting in counter=2. Also in frame N+1 630-2, sensor failure detector 122 receives the S0 output setting. Sensor failure detector 122 finds a queue element that matches output setting S0 and decrements the counter of the matching queue element S0. Decrementing sets the counter for S0 to 0. Since the matching queue element is not at the head of the queue, sensor failure detector 122 removes from the queue any element (S-1) that precedes the matching queue element (S0). The sensor failure detector 122 compares the counter of the element S-1 in the queue with 0.Since the counter is zero, no error messages are generated for frame N+1 640-2 (status: OK).

[0091] In frame N+2 640-3, the sensor failure detector 122 receives the input setting S2. Since S2 does not match the previous setting (S1), the sensor failure detector 122 adds the input setting S2 with counter=1 to the setting queue 642-3. Thus, the queue 642-3 contains the setting S0 with counter=0 (at the head of the queue), followed by the settings S1 with counter=2 and S2 with counter=1. Also in frame N+2 640-3, the sensor failure detector 122 receives the setting S1. The sensor failure detector 122 finds a queue element that matches the output setting S1 and decrements the counter of the matching queue element S1. Since the matching queue element is not at the head of the queue, the sensor failure detector 122 removes any element (S0) that precedes the matching queue element. The sensor failure detector 122 compares the counter of the element S0 in the queue with 0.Since the counter is zero, no error messages are generated for frame N+2 640-3 (status: OK).

[0092] In frame N+3 640-4, the sensor fault detector 122 receives the input setting S3. Since S3 is not equal to the previous setting (S2), the sensor fault detector 122 adds the input setting S3 to the queue 642-4 with counter=1. Furthermore, in frame N+3 640-4, the sensor fault detector 122 receives the output setting S2. The sensor fault detector 122 finds a queue element that matches the output setting S2 and decrements the counter of the matching queue element S2. This decrement sets the counter of S2 to 0.

[0093] Since the matching queue element is not at the beginning of the queue, the sensor failure detector 122 removes any element (S1) from the queue that is before the matching queue element (S2). The sensor failure detector 122 compares the counter of the element S1 that is in the queue with 0. Since the counter is 1, i.e., greater than 0, and the counter of the previous delisted element S2 is 0, a frame loss error message is generated for frame N+3 (see operation 378 in Fig. 3B).

[0094] Fig. 6D illustrates an example implementation 609 of a runtime phase of a sensor fault detector when a sensor setting loss occurs, in accordance with some embodiments of the present disclosure. In the example implementation 609, the input setting sequence 614 includes the elements S-1, S0, S1, S2, S3, S4 (where S-1 represents S-1). The output setting sequence 616 includes the elements S-1, S0, S1, S1, S2, S3. The input setting S2 was lost, so S1 appears twice in the received setting sequence 616. Frames N 650-1 through N+3 650-4 of Fig. 6D are the same as frames N 630-1 to N+3 630-4 of Fig. 6B.

[0095] In frame N+4 650-5, sensor failure detector 122 receives input setting S5. Since S5 does not match the previous setting (S4), sensor failure detector 122 adds input setting S5 with counter=1 to setting queue 652-5. Therefore, queue 652-5 contains setting S1 with counter=-1 (at the beginning of the queue), followed by settings S2, S3, S4, and S5, each with counter=1. Also, in frame N+4 650-5, sensor failure detector 122 receives setting S3. Sensor failure detector 122 finds a queue element that matches output setting S3 and decrements the counter of the matching queue element S3. Since the matching queue element S3 is not at the beginning of the queue, the sensor failure detector 122 removes each element (S1 and S2) that precedes the matching queue element. The sensor failure detector 122 compares the counter of the delisted element S1 to 0.Since the counter of S1 is less than 0 and the counter of the next queue element S2 is greater than 0, an input setting loss error message is generated for frame N+4 650-5 (see operation 370).

[0096] It should be understood that these and other arrangements described herein are only examples. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, groupings of functions, etc.) may be used in addition to or instead of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional units that may be implemented as individual or distributed components or in conjunction with other components, and in any suitable combination and location. Various functions performed by units described herein may be performed by hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in memory.In some embodiments, the systems, methods, and processes described herein may be implemented with similar components, features, and / or functions as those of the example autonomous vehicle 700 of FIG. Fig. 7A-7D, the example computing device 800 of Fig. 8 and / or the example data center 900 of Fig. 9.

[0097] In summary, the described techniques use a sensor settings verification system to detect and report errors that occur when sending or applying sensor input settings to sensors, or when receiving sensor output settings from sensors. The settings verification system monitors "input" settings sent to the sensor from a sensor controller or other system component. The settings verification system also monitors "output" settings received with each output data frame generated by the sensor, indicating the settings the sensor used to generate the output data frame. The settings verification system archives each input setting in a settings queue when the setting is sent to the sensor and assigns a counter to each input setting in the queue.The counter associated with each input setting in the queue is updated when input settings are sent to the sensor and output settings are received from the sensor, and represents the difference between the number of consecutive times the input setting has been sent to the sensor and the number of times an output setting matching the input setting has been received.

[0098] The setting verification system uses the queue to detect setting failures. Setting failures include, for example, the failure to receive the expected input settings at the sensor ("setting loss"), the failure to receive a sensor output data frame containing an output setting matching a corresponding input setting ("frame loss"), a delay in updating the sensor to use input settings ("setting delay"), or corruption of input or output settings ("setting corruption"). If no failure occurs, the counter for a particular input setting in the queue is set to zero before an output data frame with a different output setting is received that matches another input setting that was subsequently sent to the sensor.Setting corruption can be detected when an output setting is received. If no input setting is found in the queue that matches the output setting, setting corruption has occurred. Other types of errors, such as setting loss, frame loss, or setting delay, can be detected when an output setting is received and the output setting matches an input setting that is in the queue but not at the beginning of the queue. In this case, there is at least one mismatched input setting before the matching output setting in the queue. If the counter for the mismatched input setting is non-zero, a fault has occurred. The fault is identified by whether the counter for the matching output setting is less than zero (e.g.Sensor setting update delay or sensor setting loss) or greater than zero (e.g., frame loss or sensor setting loss). The particular fault can be identified by a counter associated with the next setting in the queue or the previous setting in the queue. When a fault is detected, the setting verification system generates an error message to inform other system components, such as a sensor controller or application, about the nature of the fault and the output and / or input settings affected by the fault.

[0099] A technical advantage of the described techniques over the prior art is the ability to detect a malfunction in the application of sensor settings when numerous different settings are applied, as is the case, for example, with continuous automatic exposure. Another technical advantage of the described techniques is the ability to identify and report the failure mode. The failure modes that can be detected include, among others, data corruption, lost settings that were sent but not used by the sensor, or lost output frames that were not received by the sensor. The detected failure mode can be used to find the cause of the failure and take corrective action. These technical advantages represent one or more technological improvements over the prior art.

[0100] The systems and methods described herein may be used, without limitation, by non-autonomous vehicles or machines, semi-autonomous vehicles or machines (e.g., in one or more adaptive driver assistance systems (ADAS)), autonomous vehicles or machines, guided and unguided robots or robotic platforms, warehouse vehicles, off-road vehicles, vehicles coupled to one or more trailers, aircraft, boats, shuttles, emergency vehicles, motorcycles, electric or motorized bicycles, airplanes, construction vehicles, underwater vehicles, drones, and / or other types of vehicles.Furthermore, the systems and methods described herein may be used for a variety of purposes, including, without limitation, machine control, machine locomotion, machine driving, synthetic data generation, model training, perception, augmented reality, virtual reality, mixed reality, robotics, security and surveillance, simulation and digital twinning, autonomous or semi-autonomous machine applications, deep learning, simulation of environments, simulation of objects or actors and / or digital twinning, data center processing, conversational AI, light transport simulation (e.g., ray tracing, path tracing, etc.), collaborative content creation for 3D assets, cloud computing, and / or other suitable applications.

[0101] The described embodiments may be included in a variety of different systems, such as automotive systems (e.g., a control system for an autonomous or semi-autonomous machine, a perception system for an autonomous or semi-autonomous machine), systems implemented using a robot, aviation systems, media systems, boat systems, intelligent area surveillance systems, systems for performing deep learning operations, systems for performing simulation operations, systems for performing digital twin operations, systems implemented using an edge device, systems containing one or more virtual machines (VMs), systems for generating synthetic data, systems implemented at least partially in a data center, systems for performing conversational AI operations, systems for performing light transport simulations,Systems for collaborative content creation for 3D assets, systems implemented at least partially using cloud computing resources, and / or other types of systems. EXAMPLE AUTONOMOUS VEHICLE

[0102] Fig. 7A is an illustration of an example of an autonomous vehicle 700 in accordance with some embodiments of the present disclosure. The autonomous vehicle 700 (alternatively referred to herein as "vehicle 700") may be, without limitation, a passenger vehicle, such as a passenger car, a truck, a bus, an emergency vehicle, a shuttle, an electric or motorized bicycle, a motorcycle, a fire engine, a police vehicle, an ambulance, a boat, a construction vehicle, an underwater vehicle, a robotic vehicle, a drone, an aircraft, a vehicle coupled to a trailer (e.g., a semi-trailer truck for transporting goods), and / or another type of vehicle (e.g., an unmanned vehicle and / or a vehicle with one or more passengers).Autonomous vehicles are generally described in terms of levels of automation defined by the National Highway Traffic Safety Administration (NHTSA), a division of the U.S. Department of Transportation, and the Society of Automotive Engineers (SAE) "Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles" (Standard No. J3016-201806, published June 15, 2018, Standard No. J3016-201609, published September 30, 2016, and prior and future versions of this standard). The vehicle 700 may exhibit features according to one or more of the Levels 3 through 7 of the autonomous driving levels. The vehicle 700 may exhibit features according to one or more of the Level 1 - Level 7 of the autonomous driving levels.For example, depending on the embodiment, the vehicle 700 may provide driver assistance (Level 1), partial automation (Level 2), conditional automation (Level 3), high automation (Level 4), and / or full automation (Level 7). The term "autonomous" as used herein may encompass any and / or all types of autonomy for the vehicle 700 or other machine, such as fully autonomous, highly autonomous, conditionally autonomous, partially autonomous, assisted autonomy, semi-autonomous, primarily autonomous, or another designation.

[0103] The vehicle 700 may include components such as a chassis, a vehicle body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other components of a vehicle. The vehicle 700 may include a propulsion system 750, such as an internal combustion engine, a hybrid electric drive, a pure electric motor, and / or another type of drive. The propulsion system 750 may be connected to a drivetrain of the vehicle 700, which may include a transmission to enable propulsion of the vehicle 700. The propulsion system 750 may be controlled in response to receiving signals from the throttle / accelerator device 752.

[0104] A steering system 754, which may include a steering wheel, may be used to steer the vehicle 700 (e.g., along a desired path or route) when the propulsion system 750 is operating (e.g., when the vehicle is moving). The steering system 754 may receive signals from a steering actuator 756. The steering wheel may be optional in full automation (Level 7).

[0105] The brake sensor system 746 may be used to apply the vehicle brakes in response to receiving signals from the brake actuators 748 and / or the brake sensors.

[0106] The control unit(s) 736, which comprises one or more System-on-Chip (SoCs) 704 ( Fig. 7C) and / or GPU(s), may send signals (e.g., representing commands) to one or more components and / or systems of the vehicle 700. For example, the control unit(s) may send signals to apply the vehicle brakes via one or more brake actuators 748, to apply the steering system 754 via one or more steering actuators 756, to apply the propulsion system 750 via one or more throttle / accelerator devices 752. The control unit(s) 736 may include one or more integrated computing devices (e.g., supercomputers) that process sensor signals and issue operational commands (e.g., signals representing commands) to enable autonomous driving and / or to assist a human driver in driving the vehicle 700.The control unit(s) 736 may include a first control unit 736 for autonomous driving functions, a second control unit 736 for functional safety functions, a third control unit 736 for artificial intelligence functions (e.g., computer vision), a fourth control unit 736 for infotainment functions, a fifth control unit 736 for emergency redundancy, and / or other control units. In some examples, a single control unit 736 may perform two or more of the above functions, two or more control units 736 may perform a single function, and / or any combination thereof.

[0107] The control unit(s) 736 may provide the signals to control one or more components and / or systems of the vehicle 700 in response to sensor data received from one or more sensors (e.g., sensor inputs). The sensor data may, for example and without limitation, come from GNSS sensors 758 (e.g., Global Positioning System sensors), RADAR sensors 760, ultrasonic sensors 762, LiDAR sensors 764, IMU sensors 766 (e.g., accelerometers, gyroscopes, magnetic compasses, magnetometers, etc.), microphone(s) 796, stereo camera(s) 768, wide-angle camera(s) 770 (e.g., fisheye cameras), infrared camera(s) 772, environmental camera(s) 774 (e.g., 360-degree cameras), long-range and / or medium-range camera(s) 798, speed sensor(s) 744 (e.g., for measuring the speed of the vehicle 700), vibration sensor(s) 742, steering sensor(s) 740, brake sensor(s) (e.g., as part of the brake sensor system 746) and / or other sensor types.The controller(s) 736 may include one or more instances of the sensor fault detector 122 to monitor the performance of the sensors based on the corresponding sensor data.

[0108] One or more of the control units 736 may receive inputs (e.g., in the form of input data) from an instrument cluster 732 of the vehicle 700 and provide outputs (e.g., in the form of output data, display data, etc.) via a human-machine interface (HMI) display 734, an audible annunciator, a speaker, and / or via other components of the vehicle 700. The outputs may include information such as vehicle speed vector, speed, time, map data (e.g., the high definition ("HD") map 722 of Fig. 7C), location data (e.g., the location of the vehicle 700, e.g., on a map), direction, location of other vehicles (e.g., an occupancy grid), information about objects and the status of objects as perceived by the controller(s) 736, etc. For example, the HMI display 734 may display information about the presence of one or more objects (e.g., a road sign, a warning sign, a changing traffic light, etc.) and / or information about maneuvers the vehicle has performed, is currently performing, or will perform (e.g., change lanes now, take exit 34B in two miles, etc.).

[0109] The vehicle 700 also includes a network interface 724 that may use one or more wireless antenna(s) 726 and / or modem(s) to communicate over one or more networks. The network interface 724 may, for example, enable communication over Long-Term Evolution ("LTE"), Wideband Code Division Multiple Access ("WCDMA"), Universal Mobile Telecommunications System ("UMTS"), Global System for Mobile Communication ("GSM"), IMT-CDMA Multi-Carrier ("CDMA2000"), etc. The wireless antenna(s) 726 may also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using local area networks such as Bluetooth, Bluetooth Low Energy ("LE"), Z-Wave, ZigBee, etc., and / or Low Power Wide Area Networks ("LPWANs") such as LoRaWAN, SigFox, etc.

[0110] Fig. 7B is an example of camera positions and fields of view for the autonomous vehicle 700 of Fig. 7A, in accordance with some embodiments of the present disclosure. The cameras and the respective fields of view are an example embodiment and are not intended to be limiting. For example, additional and / or alternative cameras may be present and / or the cameras may be located at various locations on the vehicle 700.

[0111] The camera types for the cameras may include, but are not limited to, digital cameras that can be adapted for use with the components and / or systems of the vehicle 700. The camera(s) may operate at Security Level B (ASIL) and / or another ASIL. The camera types may have any image capture rate, such as 60 frames per second (fps), 120 fps, 240 fps, etc., depending on the embodiment. The cameras may use rolling shutter, global shutter, another shutter type, or a combination thereof. In some examples, the color filter array may include a red-clear-clear-clear color filter array (RCCC), a red-clear-clear-blue color filter array (RCCB), a red-blue-green-clear color filter array (RBGC), a Foveon X3 color filter array, a Bayer sensor color filter array (RGGB), a monochrome sensor color filter array, and / or another type of color filter array.In some embodiments, cameras with clear pixels, such as cameras with an RCCC, an RCCB, and / or an RBGC color filter array, may be used to increase light sensitivity.

[0112] In some examples, one or more of the cameras can be used to run advanced driver assistance systems (ADAS) (e.g., as part of a redundant or fail-safe design). For example, a multifunction mono camera can be installed, providing functions such as lane departure warning, traffic sign assist, and intelligent headlight control. One or more of the cameras (e.g., all cameras) can simultaneously record and provide image data (e.g., video).

[0113] One or more of the cameras can be mounted in a bracket, e.g., a custom-made (three-dimensionally ("3D") printed) mount, to eliminate stray light and reflections from inside the vehicle (e.g., reflections from the dashboard reflected in the windshield mirrors) that could interfere with the camera's image data acquisition. Regarding the mounting of the exterior mirrors, the exterior mirrors can be custom 3D printed so that the camera mounting plate is adapted to the shape of the exterior mirror. In some examples, the camera can be integrated into the exterior mirror. For side-view cameras, the cameras can also be integrated into the four pillars at each corner of the cabin.

[0114] Cameras with a field of view that includes portions of the environment in front of the vehicle 700 (e.g., forward-facing cameras) can be used for the surrounding view to identify forward paths and obstacles and, with the assistance of one or more controllers 736 and / or control SoCs, provide information critical for establishing an occupancy grid and / or determining preferred vehicle paths. Forward-facing cameras can be used for many of the same ADAS functions as LiDAR, including emergency braking, pedestrian detection, and collision avoidance. Forward-facing cameras can also be used for ADAS functions and systems such as lane departure warning ("LDW"), autonomous cruise control ("ACC"), and / or other functions such as traffic sign recognition.

[0115] A variety of cameras can be used in a forward-facing configuration, e.g., a monocular camera platform containing a CMOS color image sensor. Another example is the wide-angle camera(s) 770, which can be used to detect objects entering the field of view from the periphery (e.g., pedestrians, crossing vehicles, or bicycles). Although Fig. 7B depicts only one wide-angle camera, the vehicle 700 may be equipped with any number (including zero) of wide-angle cameras 770. Furthermore, any number of wide-angle cameras 798 (e.g., a pair of wide-angle stereo cameras) may be used for depth-based object detection, particularly for objects for which a neural network has not yet been trained. The long-range camera(s) 798 may also be used for object detection and classification, as well as basic object tracking.

[0116] Any number of stereo cameras 768 may also be included in a forward-facing configuration. In at least one embodiment, one or more of the stereo camera(s) 768 may include an integrated control unit comprising a scalable processing unit that can provide programmable logic ("FPGA") and a multi-core microprocessor with an integrated Controller Area Network ("CAN") or Ethernet interface on a single chip. Such a unit may be used to create a 3D map of the vehicle's surroundings, including a distance estimate for all points in the image. An alternative stereo camera(s) 768 may include a compact stereo vision sensor(s) containing two camera lenses (one each on the left and right) and an image processing chip that measures the distance between the vehicle and the target object and processes the generated information (e.g.,Metadata) is used to activate the autonomous emergency braking and lane departure warning functions. In addition to or as an alternative to the stereo cameras described here, other types of stereo cameras 768 may also be used.

[0117] Cameras with a field of view that includes parts of the environment to the side of the vehicle 700 (e.g., side cameras) can be used for the environmental view and provide information used to create and update the occupancy grid and to generate collision warnings in the event of a side impact. For example, the environmental camera(s) 774 (e.g., four environmental cameras 774, as shown in Fig. 7B) may be positioned on the vehicle 700. The surround camera(s) 774 may include wide-angle camera(s) 770, fisheye camera(s), 360-degree camera(s), and / or the like. For example, four fisheye cameras may be mounted on the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle may utilize three surround cameras 774 (e.g., left, right, and rear) and employ one or more other cameras (e.g., a forward-facing camera) as a fourth surround camera.

[0118] Cameras with a field of view that includes portions of the environment behind the vehicle 700 (e.g., rearview cameras) may be used for parking assistance, surrounding view, rear collision warnings, and occupancy grid creation and updating. A variety of cameras may be used, including, but not limited to, cameras that are also suitable as forward-facing cameras (e.g., long-range and / or mid-range camera(s) 798, stereo camera(s) 768, infrared camera(s) 772, etc.), as described herein.

[0119] Fig. 7C is a block diagram of an example system architecture for the example autonomous vehicle 700 of Fig. 7A, in accordance with some embodiments of the present disclosure. It should be understood that these and other arrangements described herein are only examples. Other configurations and elements (e.g., engines, interfaces, functions, arrangements, groupings of functions, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Furthermore, many of the elements described herein are functional units that may be implemented as individual or distributed components, or in conjunction with other components, and in any suitable combination and location. Various functions performed by units described herein may be performed by hardware, firmware, and / or software. For example, various functions may be performed by a processor executing instructions stored in memory.

[0120] All components, functions and systems of the vehicle 700 in Fig. 7C are connected via bus 702. Bus 702 may include a Controller Area Network (CAN) data interface (also referred to herein as a "CAN bus"). A CAN bus may be a network within vehicle 700 used to control various features and functions of vehicle 700, such as the application of brakes, acceleration, braking, steering, windshield wipers, etc. A CAN bus may be configured to have dozens or even hundreds of nodes, each with its own unique identifier (e.g., a CAN ID). The CAN bus may be read to determine steering wheel angle, vehicle speed, engine speed (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0121] Although bus 702 is described herein as a CAN bus, this is not a limitation. For example, FlexRay and / or Ethernet may be used in addition to or as an alternative to the CAN bus. Even though bus 702 is represented by a single line, this is not a limitation. For example, there may be any number of buses 702, which may include one or more CAN buses, one or more FlexRay buses, one or more Ethernet buses, and / or one or more other bus types using a different protocol. In some examples, two or more buses 702 may be used to perform different functions and / or provide redundancy. For example, a first bus 702 may be used for collision avoidance and a second bus 702 may be used for drive control.In each example, each bus 702 may communicate with any component of the vehicle 700, and two or more buses 702 may communicate with the same components. In some examples, each SoC 704, each control unit 736, and / or each computer within the vehicle may have access to the same input data (e.g., inputs from sensors of the vehicle 700) and may be connected to a common bus, such as the CAN bus.

[0122] The vehicle 700 may include one or more control units 736 as described herein in Fig. 7A. The control unit(s) 736 may be used for a variety of functions. The control unit(s) 736 may be coupled to the various other components and systems of the vehicle 700 and may be used for controlling the vehicle 700, the artificial intelligence of the vehicle 700, the infotainment of the vehicle 700, and / or the like.

[0123] The vehicle 700 may include a system(s) on a chip (SoC) 704. The SoC 704 may include CPU(s) 706, GPU(s) 708, processor(s) 710, cache(s) 712, accelerator 714, data storage 716, and / or other components and functions not shown. The SoC(s) 704 may be used to control the vehicle 700 in a variety of platforms and systems. For example, the SoC(s) 704 may be combined in a system (e.g., the system of the vehicle 700) with an HD card 722 that may be accessed via a network interface 724 from one or more servers (e.g., server(s) 778 Fig. 7D) can receive map updates and / or updates.

[0124] The CPU(s) 706 may comprise a CPU cluster or CPU complex (alternatively referred to herein as a "CCPLEX"). The CPU(s) 706 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU(s) 706 may comprise eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU(s) 706 may comprise four dual-core clusters, each cluster having a dedicated L2 cache (e.g., a 2 MB L2 cache). The CPU(s) 706 (e.g., the CCPLEX) may be configured to support concurrent operation of clusters, such that any combination of clusters of the CPU(s) 706 may be active at any given time.

[0125] The CPU(s) 706 may implement power management features that include one or more of the following: clock gating may be automatically enabled for individual hardware blocks when the hardware blocks are idle to conserve dynamic power; power gating may be enabled for each core when the core is not actively executing instructions due to the execution of WFI / WFE instructions; power gating may be enabled for each core independently; clock gating may be enabled for each core cluster independently if all cores have clock gating or power gating active; and / or power gating may be enabled for each core cluster independently if all cores have power gating active.The CPU(s) 706 may also implement an enhanced power state management algorithm, where permissible power states and expected wake-up times are determined, and the hardware / microcode determines the best power state for the core, cluster, and CCPLEX. The processor cores may support simplified power state entry sequences in software, offloading the work to the microcode.

[0126] The GPU(s) 708 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The GPU(s) 708 may be programmable and efficient for parallel workloads. The GPU(s) 708 may, in some examples, utilize an enhanced Tensor instruction set. The GPU(s) 708 may include one or more streaming microprocessors, where each streaming microprocessor may include an L1 cache (e.g., an L1 cache with at least 96 KB of memory capacity) and two or more of the streaming microprocessors may share an L2 cache (e.g., an L2 cache with 712 KB of memory capacity). In some embodiments, the GPU(s) 708 may include at least eight streaming microprocessors. The GPU(s) 708 may utilize the application programming interface(s) (API(s)) of the computing application. In addition, the GPU(s) 708 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0127] The graphics processor(s) 708 may be optimized for best performance in automotive and embedded use cases. For example, the GPU(s) 708 may be fabricated on a fin field-effect transistor (FinFET). However, this is not a limitation, and the GPU(s) 708 may also be fabricated using other semiconductor fabrication techniques. Each streaming microprocessor may include a number of mixed-precision compute cores divided into multiple blocks. For example, without loss of generality, 64 PF32 cores and 32 PF64 cores may be divided into four processing blocks. In such an example, each processing block can be assigned 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two NVIDIA TENSOR COREs with mixed precision for deep learning matrix arithmetic, an L0 instruction cache, a warp scheduler, a dispatch unit, and / or a 64 KB register file.Additionally, streaming microprocessors can include independent parallel integer and floating-point datapaths to enable efficient execution of workloads with a mix of computations and addressing calculations. Streaming microprocessors can include independent thread scheduling to enable finer synchronization and collaboration between parallel threads. Streaming microprocessors can include a combined L1 data cache and shared memory unit to improve performance while simplifying programming.

[0128] The graphics processor(s) 708 may include high-bandwidth memory (HBM) and / or a 16 GB HBM2 memory subsystem to provide, in some examples, a peak memory bandwidth of approximately 900 GB / second. In some examples, synchronous graphics random-access memory (SGRAM), e.g., synchronous graphics double-data-rate random-access memory type 5 (GDDR5), may be used in addition to or as an alternative to HBM memory.

[0129] The GPU(s) 708 may include unified memory technology with access counters to enable more accurate migration of memory pages to the processor that accesses them most frequently, thereby improving efficiency for memory regions shared between processors. In some examples, address translation service (ATS) support may be used to allow the GPU(s) 708 to directly access the page tables of the CPU(s) 706. In such examples, if the memory management unit (MMU) of the GPU(s) 708 detects a fault, an address translation request may be communicated to the CPU(s) 706. In response, the CPU(s) 706 may look up the virtual-physical mapping for the address in its page tables and pass the translation back to the GPU(s) 708.Unified Memory technology enables a single, unified virtual address space for the memory of the CPU(s) 706 and the GPU(s) 708, thus simplifying the programming of the GPU(s) 708 and the porting of applications to the GPU(s) 708.

[0130] Additionally, the GPU(s) 708 may include an access counter that tracks the frequency with which the GPU(s) 708 access other processors' memory. The access counter may help move memory pages to the physical memory of the processor that accesses those pages most frequently.

[0131] The SoC(s) 704 may include any number of cache(s) 712, including those described herein. For example, the cache(s) 712 may include an L3 cache available to both the CPU(s) 706 and the GPU(s) 708 (e.g., connected to both the CPU(s) 706 and the GPU(s) 708). The cache(s) 712 may include a write-back cache that can track line states, e.g., by using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). The L3 cache may be 4 MB or more, depending on the embodiment, although smaller cache sizes may be used.

[0132] The SoC(s) 704 may include an arithmetic logic unit(s) (ALU(s)) that may be utilized in performing processing related to any of the various tasks or operations of the vehicle 700, such as processing DNNs. Additionally, the SoC(s) 704 may include one or more floating-point units (FPU(s))—or other mathematical coprocessors or numerical coprocessors—for performing mathematical operations within the system. For example, the SoC(s) 704 may include one or more FPUs integrated as execution units within a CPU(s) 706 and / or GPU(s) 708.

[0133] The SoC(s) 704 may include one or more accelerators 714 (e.g., hardware accelerators, software accelerators, or a combination thereof). The SoC(s) 704 may include, for example, a hardware acceleration cluster comprising optimized hardware accelerators and / or a large on-chip memory. The large on-chip memory (e.g., 4 MB SRAM) may enable the hardware acceleration cluster to accelerate neural networks and other computations. The hardware acceleration cluster may be used to complement the GPU(s) 708 and offload some of the tasks from the GPU(s) 708 (e.g., to free up more cycles of the GPU(s) 708 to perform other tasks). For example, the accelerator(s) 714 may be used for targeted workloads (e.g., perception, convolutional neural networks (CNNs), etc.) that are stable enough to be accelerated.The term “CNN” as used here can encompass all types of CNNs, including region-based or regional convolutional neural networks (RCNs) and fast RCNs (e.g., for object detection).

[0134] The accelerator(s) 714 (e.g., the hardware acceleration cluster) may include a deep learning accelerator (DLA). The DLA(s) may include one or more tensor processing units (TPUs), which may be configured to provide an additional ten trillion operations per second for deep learning applications and inferencing. The TPUs may be accelerators configured and optimized to perform image processing functions (e.g., for CNNs, RCNNs, etc.). The DLA(s) may further be optimized for a specific number of neural network types and floating-point operations, as well as for inferencing. The design of the DLA(s) may provide more performance per millimeter than a general-purpose GPU, far exceeding the performance of a CPU.The TPU(s) can perform several functions, including a single-instance convolution function supporting, for example, INT8, INT16, and FP16 data types for both features and weights, as well as post-processing functions.

[0135] The DLA(s) may quickly and efficiently execute neural networks, in particular CNNs, on processed or unprocessed data for a variety of functions, including, for example and without limitation: a CNN for object identification and recognition using data from camera sensors; a CNN for distance estimation using data from camera sensors; a CNN for emergency vehicle detection and identification using data from microphones; a CNN for facial recognition and vehicle owner identification using data from camera sensors; and / or a CNN for safety-related events.

[0136] The DLA(s) can perform any function of the GPU(s) 708, and by using an inference accelerator, for example, a developer can use either the DLA(s) or the GPU(s) 708 for any desired function. For example, the developer can concentrate the processing of CNNs and floating-point operations on the DLA(s) and leave other functions to the GPU(s) 708 and / or other accelerators 714.

[0137] The accelerator(s) 714 (e.g., the hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may alternatively be referred to herein as a computer vision accelerator. The PVA(s) may be developed 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(s) may provide a balance between performance and flexibility. For example, and without limitation, each PVA may include any number of reduced instruction set computers (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0138] The RISC cores may interact with image sensors (e.g., the image sensors of one of the cameras described herein), image signal processors, and / or the like. Each of the RISC cores may include any amount of memory. The RISC cores may use a variety of protocols depending on the embodiment. In some examples, the RISC cores may execute a real-time operating system (RTOS). The RISC cores may be implemented with one or more integrated circuits, application-specific integrated circuits (ASICs), and / or memory devices. The RISC cores may include, for example, an instruction cache and / or tightly coupled RAM.

[0139] The DMA may enable components of the PVA(s) to access system memory independently of the CPU(s) 706. The DMA may support any number of features designed to optimize the PVA, including, but not limited to, support for multi-dimensional addressing and / or circular addressing. In some examples, the DMA may support up to six or more dimensions of addressing, including block width, block height, block depth, horizontal block stepping, vertical block stepping, and / or depth stepping.

[0140] The vector processors may be programmable processors that can be designed to efficiently and flexibly execute the programming of computer vision algorithms and provide signal processing functions. In some examples, the PVA may include a PVA core and two vector processing subsystem partitions. The PVA core may include a processor subsystem, DMA engine(s) (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem may act as the primary processing unit of the PVA and may include a vector processing unit (VPU), an instruction cache, and / or vector memory (e.g., VMEM). A VPU core may include a digital signal processor, such as a single instruction multiple data (SIMD) and very long instruction word (VLIW) digital signal processor. Combining SIMD and VLIW can increase throughput and speed.

[0141] Each of the vector processors may include an instruction cache and be connected to dedicated memory. Therefore, in some examples, each of the vector processors may be configured to operate independently of the other vector processors. In other examples, the vector processors included in a particular PVA may be configured to use data parallelism. For example, in some embodiments, the multiple vector processors in a single PVA may execute the same computer vision algorithm, but for different regions of an image. In other examples, the vector processors included in a particular PVA may concurrently execute different computer vision algorithms on the same image, or even execute different algorithms on consecutive images or portions of an image.Among other things, any number of PVAs can be included in the hardware acceleration cluster, and any number of vector processors can be included in each of the PVAs. Furthermore, the PVA(s) can contain additional ECC (Error Correcting Code) memory to increase system security.

[0142] The accelerator(s) 714 (e.g., the hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high-bandwidth, low-latency SRAM to the accelerator(s) 714. In some examples, the on-chip memory may include at least 4 MB of SRAM, consisting, for example and without limitation, of eight field-configurable memory blocks accessible by both the PVA and the DLA. Each pair of memory blocks may include an Advanced Peripheral Bus (APB) interface, configuration circuitry, a controller, and a multiplexer. Any memory type may be used. The PVA and the DLA may access the memory via a backbone that provides the PVA and the DLA with high-speed access to the memory. The backbone may include an on-chip computer vision network that connects the PVA and the DLA to the memory (e.g.,via the APB).

[0143] The on-chip computer vision network can include an interface that determines whether both the PVA and the DLA are delivering ready and valid signals before transmitting control signals / addresses / data. Such an interface can provide separate phases and channels for transmitting control signals / addresses / data, as well as burst communication for continuous data transmission. This type of interface can comply with ISO 26262 or IEC 61508, although other standards and protocols can also be used.

[0144] In some examples, the SoC(s) 704 may include a real-time ray tracing hardware accelerator, as described in U.S. Patent Application No. 16 / 101,232, filed August 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the positions and extents of objects (e.g., within a world model), generate real-time visualization simulations, interpret radar signals, synthesize and / or analyze sound propagation, simulate sonar systems, general wave propagation simulation, compare with lidar data for localization and / or other functions, and / or for other purposes. In some embodiments, one or more tree traversal units (TTUs) may be used to perform one or more ray tracing-related operations.

[0145] The accelerator(s) 714 (e.g., the hardware accelerator cluster) have a wide range of uses for autonomous driving. The PVA can be a programmable image processing accelerator that can be used for critical processing steps in ADAS and autonomous vehicles. The capabilities of the PVA are well suited to algorithmic areas that require predictable processing with low power and low latency. In other words, the PVA is well suited for semi-dense or dense regular computations, even on small datasets, that require predictable runtimes with low latency and low power. In the context of autonomous vehicle platforms, the PVAs are therefore designed to execute classical computer vision algorithms because they are efficient at object detection and operate with integer mathematical methods.

[0146] For example, in one embodiment of the technology, the PVA is used to perform computer stereo vision. In some examples, a semi-global matching algorithm may be used, although this is not a limitation. Many Level 3-5 autonomous driving applications require motion estimation / stereo matching while driving (e.g., structure from motion, pedestrian detection, lane detection, etc.). The PVA can perform a computer stereo vision function based on inputs from two monocular cameras.

[0147] In some examples, PVA can be used to enable dense optical flow after processing raw radar data (e.g., with a 4D Fast Fourier Transform) to obtain processed radar data. In other examples, PVA is used for depth-of-flight processing, e.g., by processing raw time-of-flight data to yield processed time-of-flight data.

[0148] Any type of network can be powered by the DLA to improve control and driving safety, e.g., a neural network that outputs a confidence measure for each object detection. Such a confidence value can be interpreted as a probability, or as the relative "weight" of each detection compared to other detections. This confidence value allows the system to make further decisions about which detections should be considered true positives and which should be considered false positives. For example, the system can set a confidence threshold and only consider detections that exceed this threshold as true positives. In an automatic emergency braking (AEB) system, false positive detections would cause the vehicle to automatically perform emergency braking, which is obviously undesirable.Therefore, only the most confident detections should be considered as triggers for AEB. The DLA may employ a neural network to regress the confidence value. The neural network may take as input at least a subset of parameters, such as the dimensions of the bounding box, the ground plane estimate obtained (e.g., from another subsystem), the output of the IMU (Inertial Measurement Unit) sensor 766 correlated with the orientation of the vehicle 700, the distance, the 3D position estimates of the object obtained by the neural network and / or other sensors (e.g., LiDAR sensor(s) 764 or RADAR sensor(s) 760), and others.

[0149] The SoC(s) 704 may include data storage 716 (e.g., memory). The data storage(s) 716 may be on-chip memory of the SoC(s) 704, in which neural networks to be executed on the GPU and / or the DLA may be stored. In some examples, the capacity of the data storage(s) 716 may be large enough to store multiple instances of neural networks for redundancy and security. The data storage(s) 712 may include L2 or L3 cache(s) 712. The reference to the data storage(s) 716 may also include a reference to the memory associated with the PVA, DLA, and / or other accelerators 714, as described herein.

[0150] The SoC(s) 704 may include one or more processor(s) 710 (e.g., embedded processors). The processor(s) 710 may include a boot and power management processor, which may be a dedicated processor and subsystem to handle the boot and power management functions and associated security enforcement. The boot and power management processor may be part of the boot sequence of the SoC(s) 704 and may provide runtime power management services. The boot and power management processor may provide clock and voltage programming, support for transitions to a low-power state, management of SoC(s) 704 temperatures and temperature sensors, and / or management of the SoC(s) 704 power states.Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and the SoC(s) 704 may use the ring oscillators to sense the temperatures of the CPU(s) 706, GPU(s) 708, and / or accelerator 714. If temperatures are determined to exceed a threshold, the boot and power management processor may enter a temperature fault routine and place the SoC(s) 704 into a lower-power state and / or place the vehicle 700 into a chauffeur-to-safe-stop mode (e.g., bring the vehicle 700 to a safe stop).

[0151] The processor(s) 710 may also include a number of embedded processors that can serve as an audio processing engine. The audio processing engine may be an audio subsystem that enables full hardware support for multi-channel audio across multiple interfaces, as well as a wide and flexible range of audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core with a digital signal processor with dedicated RAM.

[0152] The processor(s) 710 may also include an "always on" processor engine that may provide the necessary hardware functions to support low-power sensor management and wake-up functions. The "always on" processor engine may include a processor core, tightly coupled memory, supporting peripherals (e.g., timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0153] The processor(s) 710 may also include a security cluster engine, which includes a dedicated processor subsystem for security management of automotive applications. The security cluster engine may include two or more processor cores, tightly coupled memory, supporting peripherals (e.g., timers, an interrupt controller, etc.), and / or routing logic. In a security mode, the two or more cores may operate in a lockstep mode, functioning as a single core with comparison logic to detect differences between their operations.

[0154] The processor(s) 710 may also include a real-time camera engine that includes a dedicated processor subsystem for managing the real-time camera.

[0155] The processor(s) 710 may also include a high dynamic range signal processor, which may include an image signal processor that is a hardware engine that is part of the camera processing pipeline.

[0156] The processor(s) 710 may include a video image compositor, which may be a processing block (e.g., implemented on a microprocessor) that implements video post-processing functions required by a video playback application to generate the final image for the player window. The video image compositor may perform lens distortion correction on the wide-angle camera(s) 770, the surround camera(s) 774, and / or the in-cabin surveillance camera sensors. The in-cabin surveillance camera sensor is preferably monitored by a neural network running on another instance of the Advanced SoC and configured to detect and respond to events in the cabin.An in-cabin system can perform lip reading to activate cellular service and place a call, dictate emails, change the vehicle's destination, activate or change the vehicle's infotainment system and settings, or enable voice-activated web browsing. Certain functions are available to the driver only when the vehicle is operating in autonomous mode and are disabled otherwise.

[0157] The video compositor can incorporate enhanced temporal noise reduction to reduce both spatial and temporal noise. For example, if motion is present in a video, the noise reduction weights the spatial information accordingly, reducing the weight of information provided by neighboring frames. If an image or part of an image contains no motion, the video compositor's temporal noise reduction can use information from the previous frame to reduce noise in the current frame.

[0158] The video image compositor can also be configured to perform stereo rectification of the input stereo image frames. The video image compositor can also be used for user interface design when the operating system desktop is in use and the GPU(s) 708 do not need to constantly render new surfaces. Even when the graphics processor(s) 708 are turned on and actively performing 3D rendering, the video image compositor can be used to offload the graphics processor(s) 708, thus improving performance and responsiveness.

[0159] The SoC(s) 704 may also include a serial MIPI (Mobile Industry Processor Interface) camera interface for receiving video and input from cameras, a high-speed interface, and / or a video input block that may be used for camera and related pixel input functions. The SoC(s) 704 may also include an input / output controller that may be controlled by software and used to receive I / O signals not associated with a specific function.

[0160] The SoC(s) 704 may also include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management, and / or other devices. The SoC(s) 704 may be used to process data from cameras (e.g., via Gigabit Multimedia Serial Link and Ethernet), sensors (e.g., LiDAR sensor(s) 764, RADAR sensor(s) 760, etc., which may be connected via Ethernet), data from bus 702 (e.g., speed of vehicle 700, steering wheel position, etc.), and data from GNSS sensor(s) 758 (e.g., connected via Ethernet or KANN bus). The SoC(s) 704 may also include dedicated high-performance mass storage controllers, which may include their own DMA engines and which may be used to offload routine data management tasks from the CPU(s) 706.

[0161] The SoC(s) 704 may be an end-to-end platform with a flexible architecture spanning automation levels 3 through 5, thereby providing a comprehensive functional safety architecture that leverages computer vision and ADAS techniques for diversity and redundancy, and provides a platform for a flexible, reliable driving software stack along with deep learning tools. The SoC(s) 704 may be faster, more reliable, and even more power and space efficient than conventional systems. For example, the accelerator(s) 714, in combination with the CPU(s) 706, the GPU(s) 708, and the data memory(s) 716, may form a fast, efficient platform for Level 3-5 autonomous vehicles.

[0162] The technology thus offers capabilities and functions that cannot be achieved with conventional systems. For example, computer vision algorithms can be executed on CPUs, which can be configured using high-level programming languages, such as the C programming language, to execute a variety of processing algorithms on a wide variety of visual data. However, CPUs are often unable to meet the performance requirements of many image processing applications, e.g., in terms of execution time and power consumption. In particular, many CPUs are unable to execute complex object detection algorithms in real time, which is a prerequisite for in-vehicle ADAS applications and a requirement for practical Level 3-5 autonomous vehicles.

[0163] In contrast to conventional systems, the technology described here enables the simultaneous and / or sequential execution of multiple neural networks and the combination of the results to enable Level 3-5 autonomous driving functions by providing a CPU complex, a GPU complex, and a hardware acceleration cluster. For example, a CNN running on the DLA or dGPU (e.g., the GPU(s) 720) may include text and word recognition, allowing the supercomputer to read and understand traffic signs, even those for which the neural network has not been specifically trained. The DLA may also include a neural network capable of identifying, interpreting, and semantically understanding the signs and passing this semantic understanding to the path planning modules running on the CPU complex.The DLA can also use metrics related to sensor performance as input to one or more neural networks.

[0164] Another example is that multiple neural networks can run simultaneously, as required for Level 3, 4, or 7 driving. A warning sign reading "Caution: Flashing lights indicate black ice," along with an electric light, can be interpreted independently or jointly by multiple neural networks. The sign itself can be identified as a traffic sign by a first deployed neural network (e.g., a trained neural network), and the text "Flashing lights indicate black ice" can be interpreted by a second deployed neural network, which informs the vehicle's path-planning software (preferably running on the CPU complex) that the flashing lights indicate black ice.The turn signal can be identified by running a third neural network across multiple frames, where the third neural network informs the vehicle's path planning software about the presence (or absence) of turn signals. All three neural networks can run simultaneously, e.g., within the DLA and / or on the GPU(s) 708.

[0165] In some examples, a facial recognition and vehicle owner identification CNN may use data from camera sensors to identify the presence of an authorized driver and / or owner of the vehicle 700. The "always on" sensor processing engine may be used to unlock the vehicle when the owner approaches the driver's door and turns on the lights, and in security mode, to disable the vehicle when the owner exits the vehicle. In this way, the SoC(s) 704 provide security against theft and / or carjacking.

[0166] In another example, a CNN for emergency vehicle detection and identification may use data from microphones 796 to detect and identify emergency vehicle sirens. Unlike conventional systems that use general classifiers to detect sirens and manually extract features, the SoC(s) 704 utilize the CNN to classify ambient and urban noise, as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to detect the relative approach speed of the emergency vehicle (e.g., using the Doppler effect). The CNN may also be trained to detect emergency vehicles specific to the local region in which the vehicle is traveling, as identified by the GNSS sensors 758.For example, the CNN will attempt to detect European sirens when deployed in Europe and only North American sirens when deployed in the United States. Once an emergency vehicle is detected, a control program can be used to execute an emergency vehicle safety routine, slow the vehicle, pull over to the side of the road, park the vehicle, and / or idle the vehicle using ultrasonic sensors 762 until the emergency vehicle(s) has passed.

[0167] The vehicle may include one or more CPU(s) 718 (e.g., discrete CPU(s) or dCPU(s)) that may be connected to the SoC(s) 704 via a high-speed connection (e.g., PCIe). The CPU(s) 718 may be, for example, an x86 processor. The CPU(s) 718 may be used to perform any of a variety of functions, including reconciling potentially conflicting results between ADAS sensors and the SoC(s) 704 and / or monitoring the status and health of the controller(s) 736 and / or the infotainment SoC 730, for example.

[0168] The vehicle 700 may include one or more GPU(s) 720 (e.g., discrete GPU(s) or dGPU(s)) that may be connected to the SoC(s) 704 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU(s) 720 may provide additional artificial intelligence functionality, e.g., by executing redundant and / or distinct neural networks, and may be used to train and / or update neural networks based on inputs (e.g., sensor data) from sensors of the vehicle 700.

[0169] The vehicle 700 may also include the network interface 724, which may include one or more wireless antennas 726 (e.g., one or more wireless antennas for various communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 724 may be used to enable a wireless connection over the internet to the cloud (e.g., to the server(s) 778 and / or other network devices), to other vehicles, and / or to computing devices (e.g., passenger client devices). To communicate with other vehicles, a direct connection may be established between the two vehicles and / or an indirect connection (e.g., via networks and the internet). Direct connections may be established via a vehicle-to-vehicle communication link.The vehicle-to-vehicle communication link may provide information to the vehicle 700 about vehicles in the vicinity of the vehicle 700 (e.g., vehicles in front of, beside, and / or behind the vehicle 700). This function may be part of a cooperative adaptive cruise control function of the vehicle 700.

[0170] The network interface 724 may include an SoC that provides modulation and demodulation functions and enables the controller(s) 736 to communicate over wireless networks. The network interface 724 may include a radio frequency front-end for upconversion from baseband to radio frequency and downconversion from radio frequency to baseband. The frequency conversions may be performed using known methods and / or super-heterodyne methods. In some examples, the radio frequency front-end functionality may be provided by a separate chip. The network interface may include wireless functions for communication over LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-Wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0171] The vehicle 700 may also include data storage 728 that may be stored off-chip (e.g., outside the SoC(s) 704). The data storage(s) 728 may include one or more memory elements such as RAM, SRAM, DRAM, VRAM, flash, hard drives, and / or other components and / or devices capable of storing at least one bit of data.

[0172] The vehicle 700 may also include GNSS sensor(s) 758. The GNSS sensor(s) 758 (e.g., GPS, assisted GPS sensors, differential GPS (DGPS), etc.) assist in mapping, sensing, occupancy grid creation, and / or path planning. Any number of GNSS sensors 758 may be used, e.g., a GPS with a USB port and an Ethernet-to-serial (RS-232) bridge.

[0173] The vehicle 700 may also include RADAR sensor(s) 760. The RADAR sensor(s) 760 may be used by the vehicle 700 for vehicle detection over long distances, even in darkness and / or poor weather conditions. The RADAR sensor(s) 760 may use the CAN bus and / or bus 702 (e.g., for transmitting the data generated by the RADAR sensor(s) 760) for control and access to object tracking data, with access to the raw data via Ethernet in some examples. A variety of RADAR sensor types may be used. For example, and without limitation, the RADAR sensor (or sensors) 760 may be suitable for front, rear, and side RADAR deployment. In some examples, pulse Doppler RADAR sensors are used.

[0174] The RADAR sensor(s) 760 may include various configurations, e.g., long range with a narrow field of view, short range with a wide field of view, short range with side coverage, etc. In some examples, long range RADAR may be used for adaptive cruise control. Long range RADAR systems may provide a wide field of view realized by two or more independent scans, e.g., within a range of 250 m. The RADAR sensor(s) 760 may help distinguish between static and moving objects and may be used by ADAS systems for emergency braking and forward collision warning. Long range RADAR sensors may include monostatic multimodal RADARs with multiple (e.g., six or more) fixed RADAR antennas and a high-speed CAN and FlexRay interface.In a six-antenna example, the middle four antennas can create a focused beam pattern that covers the vehicle's surroundings at higher speeds with minimal interference from traffic in adjacent lanes. The other two antennas can expand the field of view so that vehicles entering or exiting the lane of vehicle 700 can be quickly detected.

[0175] Medium-range radar systems, for example, can have a range of up to 760 m (front) or 80 m (rear) and a field of view of up to 42 degrees (front) or 750 degrees (rear). Short-range radar systems include radar sensors that can be mounted on either end of the rear bumper. When installed on either end of the rear bumper, such a radar sensor system can create two beams that continuously monitor the blind spot behind and to the side of the vehicle.

[0176] Short-range radar systems can be used in an ADAS system for blind spot detection and / or as a lane change assistant.

[0177] The vehicle 700 may also include ultrasonic sensor(s) 762. The ultrasonic sensor(s) 762, which may be mounted at the front, rear, and / or sides of the vehicle 700, may be used for parking assistance and / or for creating and updating an occupancy grid. A variety of ultrasonic sensors 762 may be used, and different ultrasonic sensors 762 may be used for different detection ranges (e.g., 2.5 m, 4 m). The ultrasonic sensor(s) 762 may operate according to functional safety levels of ASIL B.

[0178] The vehicle 700 may include LiDAR sensor(s) 764. The LiDAR sensor(s) 764 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. The LiDAR sensor(s) 764 may be ASIL B functional safety rated. In some examples, the vehicle 700 may include multiple LiDAR sensors 764 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to deliver data to a Gigabit Ethernet switch).

[0179] In some examples, the LiDAR sensor(s) 764 may be capable of providing a list of objects and their distances for a 360-degree field of view. Commercially available LiDAR sensors 764 may have a range of approximately 700 m, with an accuracy of 2 cm-3 cm, and with support for a 700 Mbps Ethernet connection, for example. In some examples, one or more non-protruding LiDAR sensors 764 may be used. In such examples, the LiDAR sensor (or LiDAR sensors) 764 may be implemented as a small device that may be embedded in the front, rear, sides, and / or corners of the vehicle 700. In such examples, the LiDAR sensor(s) 764 may provide a horizontal field of view of up to 120 degrees and a vertical field of view of up to 35 degrees, with a range of 200 m, even for objects with low reflectivity.The front-mounted LiDAR sensor(s) 764 can be configured for a horizontal field of view between 45 degrees and 135 degrees.

[0180] In some examples, LiDAR technologies such as 3D Flash LiDAR may also be used. 3D Flash LiDAR uses a laser flash as a transmission source to illuminate the vehicle's surroundings up to a distance of approximately 200 m. A Flash LiDAR unit contains a receptor that records the time of flight of the laser pulse and the reflected light on each pixel, which in turn corresponds to the distance between the vehicle and the objects. Flash LiDAR can enable highly precise and distortion-free images of the surroundings to be generated with each laser flash. In some examples, four Flash LiDAR sensors may be deployed, one on each side of the vehicle 700. Available 3D Flash LiDAR systems include a fixed 3D star array LiDAR camera that contains no moving parts other than a fan (e.g., a non-scanning LiDAR device).The Flash LiDAR device can use a 7-nanosecond Class I (eye-safe) laser pulse per image and capture the reflected laser light as 3D range point clouds and co-registered intensity data. By using Flash LiDAR and being a solid-state device with no moving parts, the 764 LiDAR sensor(s) are less susceptible to motion blur, vibration, and / or shock.

[0181] The vehicle may also include IMU sensor(s) 766. In some examples, the IMU sensor(s) 766 may be located at the center of the rear axle of the vehicle 700. The IMU sensor(s) 766 may include, for example and without limitation, one or more accelerometers, one or more magnetometers, one or more gyroscopes, one or more magnetic compasses, and / or other types of sensors. In some examples, such as in six-axis applications, the IMU sensor(s) 766 may include accelerometers and gyroscopes, while in nine-axis applications, the IMU sensor(s) 766 may include accelerometers, gyroscopes, and magnetometers.

[0182] In some embodiments, the IMU sensor(s) 766 may be implemented as a miniaturized, high-performance GPS-based inertial navigation system (GPS / INS) that combines microelectromechanical (MEMS) inertial sensors, a high-sensitivity GPS receiver, and advanced Kalman filtering algorithms to provide estimates of position, velocity, and attitude. In some examples, the IMU sensors 766 may enable the vehicle 700 to estimate heading without requiring input from a magnetic sensor by directly observing velocity changes from the GPS and correlating them with the IMU sensors 766. In some examples, the IMU sensor(s) 766 and the GNSS sensor(s) 758 may be combined into a single integrated unit.

[0183] The vehicle may include one or more microphones 796 mounted in and / or around the vehicle 700. The microphone(s) 796 may be used, among other things, for detecting and identifying emergency vehicles.

[0184] The vehicle may also include any number of camera types, including stereo camera(s) 768, wide-angle camera(s) 770, infrared camera(s) 772, surround camera(s) 774, long-range and / or medium-range camera(s) 798, and / or other camera types. The cameras may be used to capture image data around the entire perimeter of the vehicle 700. The type of cameras used depends on the embodiments and requirements of the vehicle 700, and any combination of camera types may be used to provide the necessary coverage around the vehicle 700. The number of cameras may also vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. The cameras may support, for example and without limitation, Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet.Each of the cameras is described here with reference to . Fig. 7A and Fig. 7B is described in more detail.

[0185] The vehicle 700 may also include one or more vibration sensors 742. The vibration sensor(s) 742 may measure vibrations from components of the vehicle, such as the axle(s). Changes in vibration may, for example, indicate a change in the road surface. In another example, when using two or more vibration sensors 742, the differences between the vibrations may be used to determine the friction or slippage of the road surface (e.g., if the difference in vibration is between a driven axle and a free-spinning axle).

[0186] The vehicle 700 may include an ADAS system 738. The ADAS system 738 may include an SoC in some examples. The ADAS system 738 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 keep assist (LKA), blind spot warning (BSW), rear cross traffic alert (RCTW), forward collision warning (CWS), lane centering (LC), and / or other features and functions.

[0187] The ACC systems may use radar sensor(s) 760, LiDAR sensor(s) 764, and / or one or more cameras. The ACC systems may include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to the vehicle immediately in front of vehicle 700 and automatically adjusts vehicle speed to maintain a safe distance from preceding vehicles. Lateral ACC maintains the distance and advises vehicle 700 to change lanes if necessary. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0188] CACC utilizes information from other vehicles, which may be received via the network interface 724 and / or the wireless antenna(s) 726 from other vehicles over a wireless connection or indirectly via a network connection (e.g., over the Internet). Direct connections may be established via a vehicle-to-vehicle (V2V) communication link, while indirect connections may be an infrastructure-to-vehicle (I2V) communication link. In general, the V2V communication concept provides information about the immediately preceding vehicles (e.g., vehicles immediately in front of and in the same lane as vehicle 700), while the I2V communication concept provides information about traffic further ahead. CACC systems may include one or both I2V and V2V information sources.Given the information about the vehicles ahead of vehicle 700, CACC can be more reliable and has the potential to improve traffic flow and reduce congestion on the road.

[0189] FCW systems are designed to warn the driver of a hazard so they can take corrective action. FCW systems utilize a forward-facing camera and / or radar sensor(s) coupled with a dedicated processor, DSP, FPGA, and / or ASIC that is electrically connected to provide feedback to the driver, e.g., via a display, speaker, and / or vibrating component. FCW systems can provide a warning, e.g., in the form of a sound, a visual warning, a vibration, and / or a rapid braking pulse.

[0190] AEB systems detect an impending forward collision with another vehicle or object and can automatically apply the brakes if the driver does not take corrective action within a specified time or distance parameter. AEB systems can use forward-facing camera(s) and / or radar sensor(s) 760 connected 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. If the driver does not take corrective action, the AEB system can automatically apply the brakes to prevent or at least mitigate the effects of the predicted collision. AEB systems can incorporate techniques such as dynamic brake support and / or crash-preparatory braking.

[0191] Lane departure warning systems warn the driver visually, audibly, and / or tactilely, e.g., through vibrations of the steering wheel or seat, when the vehicle 700 crosses lane markings. Lane departure warning systems will not activate if the driver indicates an intentional lane departure by activating the turn signal. LDW systems may use forward-facing cameras connected to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically connected to driver feedback, e.g., a display, speaker, and / or vibrating component.

[0192] LKA systems are a variant of LDW systems. LKA systems correct the vehicle 700 through steering inputs or braking when the vehicle 700 leaves the lane.

[0193] BSW systems detect and warn the driver of vehicles in the vehicle's blind spot. BSW systems can provide visual, audible, and / or tactile warnings that merging or changing lanes is unsafe. The system can provide an additional warning when the driver activates a turn signal. BSW systems can utilize rear-facing camera(s) and / or radar sensor(s) 760 coupled with a dedicated processor, DSP, FPGA, and / or ASIC electrically connected to driver feedback, e.g., a display, speaker, and / or vibrating component.

[0194] RCTW systems can provide visual, audible, and / or tactile notification when an object outside the range of the rear camera is detected while the vehicle 700 is reversing. Some RCTW systems include AEB to ensure the vehicle brakes are applied to avoid a crash. RCTW systems can utilize one or more rear-facing RADAR sensors 760 coupled to a dedicated processor, DSP, FPGA, and / or ASIC electrically connected to driver feedback, such as a display, speaker, and / or vibrating component.

[0195] Conventional ADAS systems can produce false positive results, which can be annoying and distracting for the driver, but are generally not catastrophic because ADAS systems warn the driver and give them the opportunity to decide whether a safety-threatening situation truly exists and act accordingly. However, in an autonomous vehicle 700, in the event of conflicting results, the vehicle 700 must decide for itself whether to consider the result of a primary or a secondary computer (e.g., a first control unit 736 or a second control unit 736). In some embodiments, the ADAS system 738 may, for example, be a backup and / or secondary computer that provides perception information to a rationality module of the backup computer.The backup computer's rationality monitor can run redundantly diverse software on hardware components to detect errors in perception and dynamic driving tasks. The outputs of the ADAS system 738 can be forwarded to a monitoring MCU. If the outputs of the primary computer and the secondary computer conflict, the monitoring MCU must determine how to resolve the conflict to ensure safe operation.

[0196] In some examples, the primary computer may be configured to provide the monitoring MCU with a confidence value indicating the primary computer's confidence in the chosen outcome. If the confidence value exceeds a threshold, the monitoring MCU may follow the primary computer's instruction regardless of whether the secondary computer provides a conflicting or inconsistent outcome. If the confidence value does not reach the threshold and the primary and secondary computers provide different outcomes (e.g., a conflict), the monitoring MCU may mediate between the computers to determine the correct outcome.

[0197] The monitoring MCU can be configured to run one or more neural networks trained and configured to determine, based on the outputs of the primary computer and the secondary computer, the conditions under which the secondary computer triggers false alarms. This allows the neural network in the monitoring MCU to learn when the output of the secondary computer can and cannot be trusted. For example, if the secondary computer is a RADAR-based FCW system, a neural network in the monitoring MCU can learn when the FCW system identifies metallic objects that are not actually hazardous, such as a drain grate or manhole cover, which triggers an alarm.If the secondary computer is a camera-based lane departure warning system, a neural network in the monitoring MCU can learn to override the lane departure warning system when cyclists or pedestrians are present and lane departure is actually the safest maneuver. In embodiments where a neural network runs on the monitoring MCU, the monitoring MCU can include at least one DLA or a GPU suitable for running the neural network(s) with associated memory. In preferred embodiments, the monitoring MCU can comprise and / or be included as a component of the SoC(s) 704.

[0198] In other examples, the ADAS system 738 may include a secondary computer that executes the ADAS functionality using conventional computer vision rules. Thus, the secondary computer may use classic computer vision rules (if-then), and the presence of one or more neural networks in the monitoring MCU may improve reliability, safety, and performance. Thus, the diverse implementation and intentional non-identity make the overall system more fault-tolerant, especially against errors caused by software features (or software-hardware interfaces).For example, if a software bug occurs in the primary computer's software and the non-identical software code on the secondary computer produces the same overall result, the monitoring MCU can assume with greater confidence that the overall result is correct and that the bug in the primary computer's software or hardware does not cause a significant failure.

[0199] In some examples, the output of the ADAS system 738 may be fed into the perception block of the main computer and / or the dynamic driving task block of the main computer. For example, if the ADAS system 738 displays a collision warning due to an object immediately ahead, the perception block may use this information in identifying objects. In other examples, the secondary computer may have its own neural network trained to reduce the risk of false alarms, as described herein.

[0200] The vehicle 700 may also include the infotainment SoC 730 (e.g., an in-vehicle infotainment (IVI) system). Although the infotainment system is illustrated and described as an SoC, it need not be a SoC, but may consist of two or more discrete components. The infotainment SoC 730 may include a combination of hardware and software responsible for providing audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming, etc.), phone (e.g., hands-free calling), network connectivity (e.g., LTE, Wi-Fi, etc.), and / or information services (e.g., navigation systems, parking assistance, a radio data system, vehicle-related information such as fuel level, total distance traveled, brake fluid level, fuel level, oil level, door open / close, air filter information, etc.) to the vehicle 700.The infotainment SoC 730 may include, for example, radios, record players, navigation systems, video players, USB and Bluetooth connections, car computers, in-car entertainment, Wi-Fi, steering wheel audio controls, a hands-free system, a heads-up display (HUD), an HMI display 734, a telematics device, a control panel (e.g., for controlling and / or interacting with various components, functions, and / or systems), and / or other components. The infotainment SoC 730 may also be used to provide information (e.g., visual and / or audible) to one or more users of the vehicle, such as information from the ADAS system 738, autonomous driving information such as planned vehicle maneuvers, trajectories, information about the environment (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0201] The infotainment SoC 730 may include GPU functions. The infotainment SoC 730 may communicate with other devices, systems, and / or components of the vehicle 700 via the bus 702 (e.g., CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 730 may be coupled to a supervisory MCU so that the infotainment system's GPU may perform some self-driving functions in the event that the primary control unit(s) 736 (e.g., the primary and / or backup computers of the vehicle 700) fail. In such an example, the infotainment SoC 730 may place the vehicle 700 into a stop-to-safe chauffeur mode, as described herein.

[0202] The vehicle 700 may also include an instrument cluster 732 (e.g., a digital instrument panel, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 732 may include a control unit and / or a supercomputer (e.g., a discrete control unit or a supercomputer). The instrument cluster 732 may include a number of instruments, such as a speedometer, fuel level, oil pressure, tachometer, odometer, turn signals, shift position indicator, seat belt warning light(s), parking brake warning light(s), engine malfunction light(s), airbag system (SRS) information, lighting controls, safety system controls, navigation information, etc. In some examples, information from the infotainment SoC 730 and the instrument cluster 732 may be displayed and / or shared. In other words, the instrument cluster 732 may be part of the infotainment SoC 730, or vice versa.

[0203] Fig. 7D is a system diagram for communication between the cloud-based server(s) and the autonomous example vehicle 700 of Fig. 7A, in accordance with some embodiments of the present disclosure. The system 776 may include the server(s) 778, the network(s) 790, and vehicles, including the vehicle 700. The server(s) 778 may include a plurality of GPUs 784(A)-784(H) (collectively referred to herein as GPUs 784), PCIe switches 782(A)-782(H) (collectively referred to herein as PCIe switches 782), and / or CPUs 780(A)-780(B) (collectively referred to herein as CPUs 780). The GPUs 784, the CPUs 780, and the PCIe switches may be interconnected via high-speed interconnects, such as the NVLink interfaces 788 developed by NVIDIA and / or PCIe interconnects 786. In some examples, the GPUs 784 are connected via NVLink and / or NVSwitch SoC, and the GPUs 784 and the PCIe switches 782 are connected via PCIe connections. Although eight GPUs 784, two CPUs 780, and two PCIe switches are shown, this is not a limitation.Depending on the embodiment, each of the servers 778 may include any number of GPUs 784, CPUs 780, and / or PCIe switches. For example, the servers 778 may each include eight, sixteen, thirty-two, and / or more GPUs 784.

[0204] The server(s) 778 may receive, via the network(s) 790 and from the vehicles, image data representative of images depicting unexpected or changed road conditions, such as recently commenced roadwork. The server(s) 778 may transmit, via the network(s) 790 and to the vehicles, neural networks 792, updated neural networks 792, and / or map information 794, including information about traffic and road conditions. The updates to the map information 794 may include updates to the HD map 722, such as information about construction, potholes, detours, flooding, and / or other obstacles.In some examples, the neural networks 792, the updated neural networks 792, and / or the map information 794 may be the result of new training and / or new experience contained in the data received from any number of vehicles in the environment and / or based on training performed in a data center (e.g., using the server(s) 778 and / or other servers).

[0205] The server(s) 778 may be used to train machine learning models (e.g., neural networks) based on training data. The training data may be generated by the vehicles and / or in a simulation (e.g., using a game engine). In some examples, the training data is tagged (e.g., when the neural network benefits from supervised learning) and / or subjected to other preprocessing, while in other examples, the training data is not tagged and / or preprocessed (e.g., when the neural network does not require supervised learning).Training may be performed using one or more classes of machine learning techniques, including, but not limited to, supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, federated learning, transfer learning, feature learning (including principal component and cluster analysis), multilinear subspace learning, nonlinear dimensionality reduction, representation learning (including sparse dictionary learning), rule-based machine learning, anomaly detection, and any variations or combinations thereof. Once the machine learning models are trained, the machine learning models may be used by the vehicles (e.g., by transmitting them to the vehicles via the network(s) 790) and / or the machine learning models may be used by the server(s) 778 to remotely monitor the vehicles.

[0206] In some examples, servers 778 may receive data from the vehicles and apply the data to real-time neural networks to draw intelligent conclusions in real time. The server(s) 778 may include deep learning supercomputers and / or dedicated AI computers powered by GPU(s) 784, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, servers 778 may also include deep learning infrastructures using only CPU-powered data centers.

[0207] The deep learning infrastructure of server(s) 778 may be capable of rapid, real-time inference and may utilize this capability to evaluate and verify the state of the processors, software, and / or associated hardware in vehicle 700. For example, the deep learning infrastructure may receive regular updates from vehicle 700, such as an image sequence and / or objects that vehicle 700 has located in that image sequence (e.g., via computer vision and / or other machine object classification techniques). The deep learning infrastructure may run its own neural network to identify the objects and compare them to the objects identified by vehicle 700.If the results do not match and the infrastructure concludes that the AI ​​in the vehicle 700 is not functioning properly, the server(s) 778 may send a signal to the vehicle 700 instructing a fail-safe computer of the vehicle 700 to take control, notify the passengers, and perform a safe parking maneuver.

[0208] For inference, server 778 may include GPU(s) 784 and one or more programmable inference accelerators (e.g., NVIDIA's TensorRT). The combination of GPU-driven servers and inference acceleration may enable real-time responsiveness. In other examples, e.g., where performance is less critical, servers powered by CPUs, FPGAs, and other processors may be used for inference. EXAMPLE COMPUTER DEVICE

[0209] Fig. 8 is a block diagram of an example computing device 800 suitable for use in implementing some embodiments of the present disclosure. Computing device 800 may include an interconnect system 802 that directly or indirectly interconnects the following devices: memory 804, one or more central processing units (CPUs) 806, one or more graphics processing units (GPUs) 808, a communications interface 810, input / output (I / O) ports 812, input / output components 814, a power supply 816, one or more presentation components 818 (e.g., display(s)), and one or more logic units 820. In at least one embodiment, computing device(s) 800 may include one or more virtual machines (VMs), and / or each of the components thereof may include virtual components (e.g., virtual hardware components).As non-limiting examples, one or more GPUs 808 may include one or more vGPUs, one or more CPUs 806 may include one or more vCPUs, and / or one or more logic units 820 may include one or more virtual logic units. A computing device (or multiple computing devices) 800 may include discrete components (e.g., a full GPU for computing device 800), virtual components (e.g., a portion of a GPU for computing device 800), or a combination thereof.

[0210] Although the different blocks in Fig. 8 are shown as being connected with lines via the interconnect system 802, this is not to be understood as a limitation and is for clarity only. For example, in some embodiments, a presentation component 818, such as a display, may be considered an I / O component 814 (e.g., if the display is a touchscreen). Another example is that the CPUs 806 and / or the GPUs 808 may include memory (e.g., the memory 804 may represent any memory device in addition to the memory of the GPUs 808, the CPUs 806, and / or other components). In other words, the computing device of Fig. 8 is merely illustrative. No distinction is made between categories such as ‘workstation’, ‘server’, ‘laptop’, ‘desktop’, ‘tablet’, ‘client device’, ‘mobile device’, ‘handheld device’, ‘game console’, ‘electronic control unit (ECU)’, ‘virtual reality system’ and / or other device or system types, since all components of the computing device are Fig. 8 should be considered.

[0211] The interconnect system 802 may represent one or more connections or buses, such as an address bus, a data bus, a control bus, or a combination thereof. The interconnect system 802 may include one or more bus or connection 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 bus or connection type. In some embodiments, there are direct connections between the components. For example, the CPU 806 may be directly connected to the memory 804. Additionally, the CPU 806 may be directly connected to the GPU 808. For direct or point-to-point connections between components, the interconnect system 802 may include a PCIe connection to establish the connection.In these examples, the computing device 800 may not necessarily include a PCI bus.

[0212] The memory 804 may consist of a variety of computer-readable media. The computer-readable media may be any available media accessible by the computing device 800. The computer-readable media may include both volatile and non-volatile media, as well as removable and non-removable media. By way of example and without limitation, the computer-readable media may include computer storage media and communication media.

[0213] The computer storage media may include both volatile and non-volatile media and / or removable and non-removable media, as implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, and / or other types of data. For example, the memory 804 may store computer-readable instructions (e.g., programs and / or program elements such as an operating system). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other storage technologies, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cartridges, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by the computing device 800. As used herein, computer storage media does not per se include signals.

[0214] Computer storage media may embody computer-readable instructions, data structures, program modules, and / or other data types in a modulated data signal, such as a carrier wave or other transport mechanism, and includes any media for transmitting information. The term "modulated data signal" may refer to a signal in which one or more of its properties are adjusted or altered to encode information in the signal. Examples of computer storage media include wired media, such as a wired network or a direct wired connection, and wireless media, such as acoustic, RF, infrared, and other wireless media. Combinations of the above media should also be considered computer-readable media.

[0215] The CPU(s) 806 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 800 and to perform one or more of the methods and / or processes described herein. The CPU(s) 806 may each include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of concurrently executing a plurality of software threads. The CPU(s) 806 may include any type of processor and may include different types of processors depending on the type of computing device 800 implemented (e.g., processors with fewer cores for mobile devices and processors with more cores for servers).Depending on the type of computing device 800, the processor may be, for example, an Advanced RISC Machines (ARM) processor using Reduced Instruction Set Computing (RISC) or an x86 processor using Complex Instruction Set Computing (CISC). Computing device 800 may include one or more CPUs 806 in addition to one or more microprocessors or additional coprocessors, such as math coprocessors.

[0216] In addition to or alternatively to the CPU(s) 806, the GPU(s) 808 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 800 to perform one or more of the methods and / or processes described herein. One or more of the GPU(s) 808 may be an integrated GPU (e.g., with one or more of the CPU(s) 806 and / or one or more of the GPU(s) 808 may be a discrete GPU). In certain embodiments, one or more of the GPU(s) 808 may be a coprocessor of one or more of the CPU(s) 806. The GPU(s) 808 may be used by the computing device 800 to render graphics (e.g., 3D graphics) or to perform general-purpose computations. The GPU(s) 808 may be used, for example, for General-Purpose Computing on GPUs (GPGPU).The GPU(s) 808 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. The GPU(s) 808 may generate pixel data for output images in response to rendering commands (e.g., rendering commands from the CPU(s) 806 received via a host interface). The GPU(s) 808 may include graphics memory, e.g., display memory, for storing pixel data or other suitable data, such as GPGPU data. The display memory may be part of the memory 804. The GPU(s) 808 may include two or more GPUs operating in parallel (e.g., via a link). The link may connect the GPUs directly (e.g., using NVLINK) or connect the GPUs via a switch (e.g., using NVSwitch). When used together, each GPU can generate 808 pixel data or GPGPU data for different parts of an output or for different outputs (e.g.a first GPU for a first image and a second GPU for a second image). Each GPU can have its own memory or share memory with other GPUs.

[0217] In addition to or alternatively to the CPU(s) 806 and / or the GPU(s) 808, the logic unit(s) 820 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 800 to perform one or more of the methods and / or processes described herein. In embodiments, the CPU(s) 806, the GPU(s) 808, and / or the logic unit(s) 820 may discretely or jointly execute any combination of the methods, processes, and / or portions thereof. One or more of the logic units 820 may be part of one or more of the CPU(s) 806 and / or the GPU(s) 808, and / or one or more of the logic units 820 may be discrete components or otherwise external to the CPU(s) 806 and / or the GPU(s) 808.In certain embodiments, one or more of the logic units 820 may be a coprocessor of one or more of the CPU(s) 806 and / or one or more of the GPU(s) 808.

[0218] Examples of the logical unit(s) 820 include one or more processing cores and / or components thereof, such as data processing units (DPUs), tensor cores (TCs), tensor processing units (TPUs), pixel visual cores (PVCs), vision processing units (VPUs), graphics processing clusters (GPCs), texture processing clusters (TPCs), streaming multiprocessors (SMs), tree traversal units (TTUs), artificial intelligence accelerators (AIAs), deep learning accelerators (DLAs), arithmetic logic units (ALUs), application-specific integrated circuits (ASICs), floating point units (FPUs), input / output (I / O) elements, peripheral component interconnect (PCI) or peripheral component interconnect express (PCIe) elements, and / or the like.

[0219] In various embodiments, one or more CPU(s) 806, GPU(s) 808, and / or logic unit(s) 820 are configured to run one or more instances of sensor fault detector 122. Output sensor frames 216 generated by sensor fault detector 122 may then be used by other components to perform further processing, such as, for example, scheduling or control functions.

[0220] The communication interface 810 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 800 to communicate with other computing devices over an electronic communication network, including wired and / or wireless communication. The communication interface 810 may include components and functions that enable communication over a variety of networks, such as wireless networks (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), wired networks (e.g., communication over Ethernet or InfiniBand), low-power wide area networks (e.g., LoRaWAN, SigFox, etc.), and / or the Internet.In one or more embodiments, the logic unit(s) 820 and / or the communication interface 810 may include one or more data processing units (DPUs) to transfer the data received over a network and / or via the interconnect system 802 directly to one or more GPU(s) 808 (e.g., a memory).

[0221] Through the I / O ports 812, the computing device 800 can be logically connected to other devices, including the I / O components 814, the presentation component(s) 818, and / or other components, some of which may be built into (e.g., integrated) the computing device 800. Example I / O components 814 include a microphone, a mouse, a keyboard, a joystick, a gamepad, a game controller, a satellite dish, a scanner, a printer, a wireless device, etc. The I / O components 814 can provide a natural user interface (NUI) that processes air gestures, voice input, or other physiological input from a user. In some cases, the inputs can be transmitted to an appropriate network element for further processing.An NUI may implement any combination of speech recognition, pen recognition, facial recognition, biometric recognition, both on-screen and off-screen gesture recognition, air gestures, head and eye tracking, and touch detection (as described in more detail below) in conjunction with a display of computing device 800. Computing device 800 may include depth cameras, such as stereoscopic camera systems, infrared camera systems, RGB camera systems, touchscreen technology, and combinations thereof, for gesture recognition and capture. Additionally, computing device 800 may include accelerometers or gyroscopes (e.g., as part of an inertial measurement unit (IMU)) that enable the detection of motion. In some examples, the output of the accelerometers or gyroscopes from computing device 800 may be used to present immersive augmented reality or virtual reality.

[0222] The power supply 816 may be a hardwired power supply, a battery power supply, or a combination thereof. The power supply 816 may supply power to the computing device 800 so that the components of the computing device 800 can operate.

[0223] The presentation component(s) 818 may include a display (e.g., a monitor, a touchscreen, a television screen, a heads-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The presentation component(s) 818 may receive data from other components (e.g., the GPU(s) 808, the CPU(s) 806, DPUs, etc.) and output the data (e.g., as an image, video, audio, etc.). EXAMPLE OF A DATA CENTER

[0224] Fig. Figure 9 shows an example of a data center 900 that may be used in at least one embodiment of the present disclosure. The data center 900 may include a data center infrastructure layer 910, a framework layer 920, a software layer 930, and / or an application layer 940.

[0225] As in Fig. 9, the infrastructure layer of the data center 910 may include a resource orchestrator 912, clustered computing resources 914, and node computing resources (“KKR”) 916(1)-716(N), where “N” represents any natural number. In at least one embodiment, the KRRs 916(1)-716(N) may be any number of central processing units (CPUs) or other processors (including DPUs, accelerators, field programmable gate arrays (FPGAs), graphics processors or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memory), storage devices (e.g., solid-state or disk drives), network input / output devices (NW I / O), network switches, virtual machines (VMs), power modules and / or cooling modules, etc. In some embodiments, one or more KKRs among the KKRs 916(1)-716(N) may correspond to a server having one or more of the above-mentioned computing resources.Furthermore, in some embodiments, the KRRs 916(1)-716(N) may include one or more virtual components, such as vGPUs, vCPUs, and / or the like, and / or one or more of the KRRs 916(1)-716(N) may correspond to a virtual machine (VM).

[0226] In at least one embodiment, the grouped computing resources 914 may include separate groupings of KRRs 916 housed in one or more racks (not shown) or in many racks in data centers in different geographical locations (also not shown). Separate groupings of KRRs 916 within the grouped computing resources 914 may include grouped computing, networking, and / or memory resources that may be configured or allocated to support one or more workloads. In at least one embodiment, multiple KRRs 916 with CPUs, GPUs, DPUs, and / or other processors may be grouped in one or more racks to provide computing resources to support one or more workloads. The one or more racks may also include any number of power modules, cooling modules, and / or network switches in any combination.

[0227] Resource orchestrator 912 may configure or otherwise control one or more KRRs 916(1)-716(N) and / or grouped computing resources 914. In at least one embodiment, resource orchestrator 912 may be a management entity for the software design infrastructure (SDI) of data center 900. Resource orchestrator 912 may be comprised of hardware, software, or a combination thereof.

[0228] In at least one embodiment, as in Fig. 9, the framework layer 920 may include a job scheduler 933, a configuration manager 934, a resource manager 936, and / or a distributed file system 938. The framework layer 920 may include a framework to support the software 932 of the software layer 930 and / or one or more applications 942 of the application layer 940. The software 932 or the application(s) 942 may include web-based service software or applications such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. The framework layer 920 may be some type of free and open source software web application framework such as, but not limited to, Apache Spark™ (hereinafter "Spark"), which may utilize the distributed file system 938 for processing large amounts of data (e.g., "Big Data").In at least one embodiment, the job scheduler 933 may include a Spark driver to facilitate the scheduling of workloads supported by various layers of the data center 900. The configuration manager 934 may be capable of configuring various layers such as the software layer 930 and the framework layer 920, including Spark and the distributed file system 938, to support the processing of large amounts of data. The resource manager 936 may manage clustered or grouped computing resources allocated to support the distributed file system 938 and the job scheduler 933. In at least one embodiment, the clustered or grouped computing resources may include the clustered computing resources 914 at the infrastructure layer 910 of the data center.The resource manager 936 may coordinate with the resource orchestrator 912 to manage these allocated or assigned computing resources.

[0229] In at least one embodiment, the software 932 included in software layer 930 may include software used by at least portions of KRRs 916(1)-716(N), clustered computing resources 914, and / or distributed file system 938 of framework layer 920. One or more types of software may include, but are not limited to, web searching software, email virus scanning software, database software, and video content streaming software.

[0230] In at least one embodiment, the application(s) 942 included in the application layer 940 may include one or more types of applications used by at least portions of the KKRs 916(1)-716(N), the clustered compute resources 914, and / or the distributed file system 938 of the framework layer 920. One or more types of applications may include any number of genomic applications, cognitive computation, and machine learning applications, including, but not limited to, training or inference software, machine learning framework software (e.g., PyTorch, TensorFlow, Caffe, etc.), and / or other machine learning applications used in connection with one or more embodiments.

[0231] In at least one embodiment, the configuration manager 934, the resource manager 936, and the resource orchestrator 912 may perform any number and type of modifying actions based on any amount and type of data collected in any technically feasible manner. Self-modifying actions may relieve the operator of a data center 900 from potentially making poor configuration decisions and potentially avoid underutilized and / or malfunctioning parts of a data center.

[0232] Data center 900 may include tools, services, software, or other resources to train one or more machine learning models or to predict or infer information using one or more machine learning models according to one or more embodiments described herein. For example, one or more machine learning models may be trained by calculating weighting parameters according to a neural network architecture using software and / or computing resources described above with respect to data center 900.In at least one embodiment, trained or deployed machine learning models corresponding to one or more neural networks may be used to infer or predict information using the resources described above with respect to data center 900 by using weighting parameters calculated by one or more training techniques, such as, but not limited to, those described herein.

[0233] In at least one embodiment, the data center 900 may utilize CPUs, application-specific integrated circuits (ASICs), GPUs, FPGAs, and / or other hardware (or corresponding virtual computing resources) to perform training and / or inference with the resources described above. Furthermore, one or more of the software and / or hardware resources described above may be configured as a service that enables the user to train or infer information, such as image recognition, speech recognition, or other artificial intelligence services. EXAMPLE OF NETWORK ENVIRONMENTS

[0234] Network environments suitable for implementing embodiments of the disclosure may include one or more client devices, servers, network-attached storage (NAS), other backend devices, and / or other device types. The client devices, servers, and / or other device types (e.g., each individual device) may reside on one or more instances of the computing device(s) 800 of Fig. 8 - e.g., each device may include similar components, features, and / or functions of the computing device(s) 800. If backend devices (e.g., servers, NAS, etc.) are implemented, the backend devices may also be part of a data center 900, an example of which is shown in Fig. 9 is described in more detail.

[0235] The components of a network environment can communicate with each other over one or more networks, which can be wired, wireless, or both. The network can comprise multiple networks or a network of networks. For example, the network can comprise one or more wide area networks (WANs), one or more local area networks (LANs), one or more public networks such as the Internet and / or a public switched telephone network (PSTN), and / or one or more private networks. If the network comprises a wireless telecommunications network, components such as a base station, a communications tower, or even access points (as well as other components) can provide wireless connectivity.

[0236] Compatible network environments include one or more peer-to-peer network environments—in which case, a server may not be included in a network environment—and one or more client-server network environments—in which case, one or more servers may be included in a network environment. In peer-to-peer network environments, the functionality described here can be implemented with respect to one or more servers on any number of client devices.

[0237] In at least one embodiment, a network environment may include one or more cloud-based network environments, a distributed computing environment, a combination thereof, etc. A cloud-based environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more servers, which may include one or more core network servers and / or edge servers. A framework layer may include a framework for supporting software of a software layer and / or one or more applications of an application layer. The software or application(s) may each include web-based service software or applications. In certain embodiments, one or more of the client devices may utilize the web-based service software or applications (e.g.,by accessing the service software and / or applications via one or more application programming interfaces (APIs). The framework layer can be, among other things, a type of free and open-source software framework for web applications that, for example, uses a distributed file system for processing large amounts of data (e.g., "Big Data").

[0238] A cloud-based network environment may provide cloud computing and / or cloud storage performing any combination of the computing and / or data storage functions described herein (or one or more portions thereof). Each of these various functions may be distributed across multiple locations from central servers or core servers (e.g., from one or more data centers that may be located across a state, region, country, globe, etc.). When a connection to a user (e.g., a client device) is relatively close to one or more edge servers, one or more core servers may delegate at least some functionality to the edge server(s). A cloud-based network environment may be private (e.g., restricted to a single organization), public (e.g., available to many organizations), and / or a combination thereof (e.g., a hybrid cloud environment).

[0239] The client device(s) may include at least some of the components, features, and functions of the Fig.8. A client device may be, for example, a personal computer (PC), a laptop, a mobile device, a smartphone, a tablet computer, a smartwatch, a wearable computer, a personal digital assistant (PDA), an MP3 player, a virtual reality headset, a global positioning system (GPS) or device, a video player, a video camera, a surveillance device or system, a vehicle, a boat, an aircraft, a virtual machine, a drone, a robot, a handheld communication device, a hospital device, a gaming device or system, an entertainment system, a vehicle computing system, an embedded system controller, a remote control, an appliance, a consumer electronics device, a workstation, an edge device, any combination of these described devices, or any other suitable device.

[0240] The disclosure of this application also includes the following numbered clauses: Clause 1. In some embodiments, a method comprises receiving, from a sensor, sensor output data and at least one output sensor setting associated with the sensor output data, determining whether a queue of pending input sensor settings includes a matching input sensor setting that matches the output sensor setting, identifying, in response to determining that the queue of pending input sensor settings includes the matching input sensor setting, a pending input sensor setting included in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue, determining, at least based on a pending request counter associated with the pending input setting, whether a sensor setting fault has occurred, in response to determining,that the sensor setting error has occurred, generating, at least based on the counter of outstanding requests, an error message, and modifying, at least based on the error message, one or more downstream operations based on the sensor output data. Clause 2. The method of Clause 1, comprising, in response to identifying the matching input sensor setting, decrementing a pending request counter associated with the matching input sensor setting, the pending request counter being based at least on a difference between a number of consecutive requests to apply the pending input sensor setting to the sensor and a number of times a corresponding matching output sensor setting has been received from the sensor. Clause 3. The method of clause 1 or 2, wherein determining whether the queue of pending input sensor settings contains the matching input sensor setting comprises determining whether the matching input sensor setting matches the output sensor setting within a threshold tolerance. Clause 4. The method of any of clauses 1-3, wherein the threshold tolerance is based on at least one or more of the matching input sensor settings or the output sensor settings. Clause 5. The method of any of clauses 1-4, comprising determining that the queue of pending input sensor settings does not contain the appropriate input sensor setting, generating an error message indicating that a data corruption error has occurred. Clause 6. The method of any of clauses 1-5, wherein determining whether a sensor setting fault has occurred is based at least on a pending request counter associated with the pending input setting and comprises comparing the pending request counter to zero. Clause 7. The method of any of clauses 1-6, wherein generating the error message based at least on the outstanding request counter comprises determining a sensor setting fault type based at least on the outstanding request counter. Clause 8. A method according to any one of clauses 1-7, wherein the sensor setting fault type corresponds to no fault if the outstanding request counter is zero. Clause 9. The method of any of clauses 1-8, wherein the sensor setting failure type corresponds to a sensor setting update delay or a sensor setting loss when the outstanding request counter associated with the outstanding input setting is less than zero. Clause 10. The method of any of clauses 1-9, wherein the sensor setting fault type corresponds to a sensor setting update delay if the outstanding request counter associated with the outstanding input setting is less than zero and a next outstanding request counter associated with a next queue element after the outstanding input setting is zero. Clause 11. The method of any of clauses 1-10, wherein the sensor setting failure type corresponds to a sensor setting loss if the counter of outstanding requests associated with the outstanding input setting is less than zero and the next counter of outstanding requests associated with a next queue element after the outstanding input setting is greater than zero. Clause 12. The method of any of clauses 1-11, wherein the sensor adjustment failure type corresponds to a frame loss or a sensor adjustment loss if the outstanding request counter associated with the outstanding input adjustment is greater than zero. Clause 13. The method of any of clauses 1-12, wherein the sensor setting failure type corresponds to a frame loss if the outstanding request counter associated with the outstanding input setting is greater than zero and a previous outstanding request counter associated with a previous queue element preceding the outstanding input setting is zero. Clause 14. The method of any of clauses 1-13, wherein the sensor setting fault type corresponds to a sensor setting abort if the outstanding request counter associated with the outstanding input setting is greater than zero and the next outstanding request counter associated with a preceding queue element preceding the outstanding input setting is less than zero. Clause 15. The method of any of clauses 1-14, further comprising, in response to identifying the pending input settings in the queue of pending input sensor settings, removing the pending input sensor setting from the queue. Clause 16. The method of any of clauses 1-15, further comprising receiving an input setting, determining whether the input setting matches a previous setting in the queue of pending settings, and in response to determining that the input setting matches the previous setting, incrementing a counter of pending requests associated with the previous setting. Clause 17. The method of any of clauses 1-16, further comprising, in response to determining that the input setting does not match the previous setting, adding the counter of outstanding requests to the end of the queue. Clause 18. In some embodiments, a processor comprises one or more processing units to perform operations including: receiving sensor output data and at least one output sensor setting associated with the sensor output data; determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting in response to determining that the queue of pending input sensor settings contains the matching input sensor setting; identifying a pending input sensor setting in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue; determining, based at least on a count of pending requests associated with the pending input setting, whether a sensor setting fault has occurred;and in response to determining that the sensor setting fault has occurred, generating an error message based at least on the counter of outstanding requests. Clause 19. The processor according to Clause 18, wherein the processor is included in at least one of the following systems: a control system for an autonomous or semi-autonomous machine, a perception system for an autonomous or semi-autonomous machine, a system for performing one or more simulation operations, a system for performing one or more digital twin operations, a system for performing light transport simulation, a system for performing collaborative content creation for 3D assets, a system for performing one or more deep learning operations, a system implemented using an edge device, a system for generating or presenting virtual reality content, augmented reality content, or mixed reality content, a system implemented using a robot,a system for performing one or more conversational AI operations, a system for performing one or more generative AI operations, a system implementing one or more large language models (LLMs), a system for generating synthetic data, a system including one or more virtual machines (VMs), a system implemented at least partially in a data center, or a system implemented at least partially using cloud computing resources. Clause 20. In some embodiments, a system comprises one or more processing units to perform operations including receiving sensor output data and at least one output sensor setting associated with the sensor output data from a sensor, determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting in response to determining that the queue of pending input sensor settings contains the matching input sensor setting, identifying a pending input sensor setting in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue, determining, at least based on a count of pending requests associated with the pending input setting, whether a sensor setting fault has occurred,and in response to determining that the sensor setting fault has occurred, generating an error message based at least on the counter of outstanding requests.

[0241] The disclosure may be described in the general context of computer code or machine-usable instructions, including computer-executable instructions such as program modules executed by a computer or other machine, such as a personal data assistant or other handheld device. In general, program modules, including routines, programs, objects, components, data structures, etc., refer to code that performs specific tasks or implements specific abstract data types. The disclosure may be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. The disclosure may also be applied in distributed computing environments where tasks are performed by remotely controlled devices interconnected via a communications network.

[0242] As used herein, any reference to "and / or" in reference to two or more elements should be interpreted to mean only one element or a combination of elements. For example, "Element A, Element B, and / or Element C" may include only Element A, only Element B, only Element C, Element A and Element B, Element A and Element C, Element B and Element C, or both Elements A, B, and C. Furthermore, "at least one of Element A or Element B" may include at least one of Element A, at least one of Element B, or at least one of Element A and at least one of Element B.

[0243] The subject matter of the present disclosure is described herein with a certain degree of particularity to satisfy legal requirements. However, the description itself is not intended to limit the scope of this disclosure. Rather, the inventors contemplated that the claimed subject matter may be embodied in other ways to incorporate various steps or combinations of steps similar to those described herein in connection with other present or future technologies. Although the terms "step" and / or "block" are used herein to refer to various elements of the methods employed, the terms should not be interpreted as implying any particular ordering among or between the various steps described herein unless the order of each step is expressly described.

[0244] It is to be understood that the aspects and embodiments described above are only exemplary and that changes in detail may be made within the scope of the claims.

[0245] Each apparatus, method, and feature described in the specification and (where appropriate) in the claims and drawings may be provided independently or in any suitable combination.

[0246] The reference numbers contained in the claims are for illustrative purposes only and do not limit the scope of the claims. QUOTES CONTAINED IN THE DESCRIPTION

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

[0000] US 16 / 101,232

[0144] Cited non-patent literature

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

[0102]

Claims

[1] Method comprising: Receiving, from a sensor, sensor output data and at least one output sensor setting associated with the sensor output data; Determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting; in response to determining that the queue of pending input sensor settings contains the matching input sensor setting, identifying a pending input sensor setting in the queue of pending input sensor settings that precedes the matching input sensor setting in the queue; Determining, at least based on a counter of outstanding requests associated with the outstanding input setting, whether a sensor setting failure has occurred; in response to determining that the sensor setting fault has occurred, generating an error message based on at least the counter of outstanding requests; and Modifying one or more downstream operations based on the sensor output data based on at least the error message. [2] A method according to claim 1, comprising: in response to identifying the matching input sensor setting, decrementing a counter of outstanding requests associated with the matching input sensor setting, wherein the counter for outstanding requests is based at least on a difference between a number of consecutive requests to apply the pending input sensor setting to the sensor and a number of times a corresponding matching output sensor setting has been received from the sensor. [3] The method of any preceding claim, wherein determining whether the queue of pending input sensor settings includes the matching input sensor setting comprises determining whether the matching input sensor setting matches the output sensor setting within a threshold tolerance. [4] The method of claim 3, wherein the threshold tolerance is based on at least one or more of the matching input sensor setting or the output sensor setting. [5] A method according to any one of the preceding claims, comprising: in response to determining that the queue of pending input sensor settings does not contain the matching input sensor setting, generate an error message indicating that a data corruption error has occurred. [6] The method of any preceding claim, wherein determining whether a sensor adjustment failure has occurred based on at least one outstanding request counter associated with the outstanding input adjustment comprises comparing the outstanding request counter to zero. [7] The method of any preceding claim, wherein generating the error message based at least on the outstanding request counter comprises determining a sensor setting failure type based at least on the outstanding request counter. [8] The method of claim 7, wherein the sensor setting fault type corresponds to no fault being present when the outstanding request counter is zero. [9] A method according to any preceding claim, wherein the sensor setting failure type corresponds to a sensor setting update delay or a sensor setting loss when the outstanding request counter associated with the outstanding input setting is less than zero. [10] A method according to any preceding claim, wherein the sensor setting failure type corresponds to a sensor setting update delay if the outstanding request counter associated with the outstanding input setting is less than zero and the next outstanding request counter associated with a next queue element after the outstanding input setting is zero. [11] A method according to any preceding claim, wherein the sensor tuning failure type corresponds to a frame loss or sensor tuning loss if the outstanding request counter associated with the outstanding input tuning is less than zero and the next outstanding request counter associated with a next queue element after the outstanding input tuning is greater than zero. [12] A method according to any preceding claim, wherein the sensor adjustment failure type corresponds to a frame loss or a sensor adjustment loss if the outstanding request counter associated with the outstanding input adjustment is greater than zero. [13] A method according to any preceding claim, wherein the sensor setting failure type corresponds to a frame loss if the outstanding request counter associated with the outstanding input setting is greater than zero and the previous outstanding request counter associated with a previous queue element preceding the outstanding input setting is zero. [14] A method according to any preceding claim, wherein the sensor setting failure type corresponds to a sensor setting loss if the outstanding request counter associated with the outstanding input setting is greater than zero and the next outstanding request counter associated with a previous queue element preceding the outstanding input setting is less than zero. [15] A method according to any one of the preceding claims, further comprising: in response to identifying the pending input settings in the pending input sensor settings queue, removing the pending input sensor setting from the queue. [16] A method according to any one of the preceding claims, the method further comprising: Receiving an input setting; Determining whether the input setting matches a previous setting in the queue of pending settings; and in response to determining that the input setting matches the previous setting, incrementing a counter of outstanding requests associated with the previous setting. [17] A method according to claim 16, comprising: in response to determining that the input setting does not match the previous setting, adding the counter of outstanding requests to the end of the queue. [18] Processor comprising: one or more processing units for performing comprehensive operations: Receiving sensor output data and at least one output sensor setting associated with the sensor output data from a sensor; Determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting; in response to determining that the queue of pending input sensor settings contains the matching input sensor setting, identifying, in the queue of pending input sensor settings, a pending input sensor setting that precedes the matching input sensor setting in the queue; Determining, at least based on a counter of outstanding requests associated with the outstanding input setting, whether a sensor setting failure has occurred; and in response to determining that the sensor setting fault has occurred, generating an error message based on at least the counter of outstanding requests. [19] The processor of claim 18, wherein the processor is included in at least one of the following systems: Control system for an autonomous or semi-autonomous machine; Perception system for an autonomous or semi-autonomous machine; System for performing one or more simulation operations; System for performing one or more digital twin operations; System for performing light transport simulations; System for collaborative content creation for 3D assets; System for performing one or more deep learning operations; System implemented with an edge device; System for generating or displaying virtual reality, augmented reality or mixed reality content; System implemented with the help of a robot; System for performing one or more conversational AI operations; System for performing one or more generative AI operations; System that implements one or more large language models (LLMs); System for generating synthetic data; System that contains one or more virtual machines (VMs); System that is at least partially implemented in a data center; or System implemented at least in part using cloud computing resources. [20] System comprising: one or more processing units for performing operations, comprising: Receiving sensor output data and at least one output sensor setting associated with the sensor output data from a sensor; Determining whether a queue of pending input sensor settings contains a matching input sensor setting that matches the output sensor setting; in response to determining that the queue of pending input sensor settings contains the matching input sensor setting, identifying, in the queue of pending input sensor settings, a pending input sensor setting that precedes the matching input sensor setting in the queue; Determining, at least based on a counter of outstanding requests associated with the outstanding input setting, whether a sensor setting fault has occurred; and in response to determining that the sensor setting fault has occurred, generating an error message based on at least the counter of outstanding requests.

Images