Emergency response vehicle detection for autonomous driving applications

Abstract

The present invention discloses emergency response vehicle detection for autonomous driving applications. In various examples, audio warnings of emergency response vehicles can be detected and classified using audio captured by microphones of autonomous or semi-autonomous machines to identify the direction of travel, position and / or type of emergency response vehicles in an environment. For example, multiple microphone arrays can be set on an autonomous or semi-autonomous machine and used to generate audio signals corresponding to sounds in the environment. These audio signals can be processed (for example, using triangulation) to determine the position and / or direction of travel of the emergency response vehicle. In addition, in order to identify the type of warning and therefore the type of emergency response vehicle corresponding to it, these audio signals can be used to generate a representation of the frequency spectrum, which can be processed using a deep neural network (DNN) that outputs a probability of the warning type represented by the audio data.

Description

Background Art

[0001] Designing a system that can autonomously and safely drive a vehicle without supervision is extremely difficult. An autonomous vehicle should be able to behave at least as functionally as a dedicated driver—leveraging a perception and action system with an incredible ability to identify and react to moving and stationary obstacles in complex environments—to avoid collisions with other objects or structures along the vehicle's path. Furthermore, to operate fully autonomously, the vehicle should be able to follow the rules or customs of the road, including those related to emergency response vehicles. For example, depending on the geographic location, various rules or customs may be in place, such as pulling over and / or stopping when an emergency response vehicle is detected.

[0002] Some conventional systems attempt to identify emergency response vehicles using the ego vehicle's perception. For example, various sensor types—e.g., LiDAR, RADAR, cameras, etc.—can be used to detect emergency response vehicles when they are perceptible by the ego vehicle. However, due to the many occlusions in the environment, these systems may have difficulty identifying emergency response vehicles, or at least having sufficient reaction time to do so. For example, in the event of an emergency response vehicle approaching an intersection that is obscured by a building or other structure, the ego vehicle may not perceive the emergency response vehicle until the vehicle also enters the intersection. At this point, it may be too late for the ego vehicle to perform a maneuver consistent with local rules, and thus may reduce the overall safety of the situation and / or hinder the movement of the emergency response vehicle. Summary of the Invention

[0003] Embodiments of the present disclosure relate to emergency response vehicle detection for autonomous driving applications. Disclosed are systems and methods for detecting and classifying alerts, such as, for example and without limitation, sirens, alarms, horns, and other patterns of sounds emitted by emergency response vehicles, and audio captured using microphones of autonomous or semi-autonomous vehicles to identify the direction, location, and / or type of emergency response vehicles in an environment. For example, multiple microphones or microphone arrays can be set on the vehicle and used to generate audio signals corresponding to the sounds in the environment. These audio signals can be processed, for example, using background noise suppression, beamforming, and / or other pre-processing operations to determine the location and / or direction of travel of the emergency response vehicle (e.g., using triangulation). To further improve the quality of the audio data, these microphones can be mounted at various locations around the vehicle (e.g., front, rear, left, right, inside, etc.) and / or include physical structures that contribute to audio quality, such as a windshield that prevents wind from acting on the microphones.

[0004] In order to identify the type of alert and, therefore, the type of emergency response vehicle corresponding to it, the audio signal can be converted to the frequency domain by extracting the Mel frequency coefficients to generate a Mel spectrogram. The Mel spectrogram can be processed using a deep neural network (DNN) such as a convolutional recurrent neural network (CRNN), which outputs a confidence or probability that the various types of alerts used by the emergency response vehicle are represented by the audio data. Due to the spatial and temporal nature of the alerts, using convolutional layers for feature extraction and recurrent layers for temporal feature recognition improves the accuracy of the DNN. For example, the DNN can operate on a window of audio data while maintaining state, such as using a gated recurrent unit (GRU), to allow for a more lightweight and accurate DNN that accounts for the spatial and temporal audio distribution of the alert. In addition, attention can be applied to the output of the temporal feature extraction layer to determine the probability of each type of emergency responder alert within the entire window, for example giving a higher weight to detected alert patterns regardless of their temporal position within the window. To train the DNN for accuracy under different driving conditions, such as rain, wind, traffic, tunnels, open air, etc., audio data from different alert types can be enhanced using time stretching, time shifting, pitch shifting, dynamic range compression, noise enhancement at different signal-to-noise ratios (SNRs), etc., to generate a more robust training dataset that accounts for echo, reverberation, attenuation, Doppler effect, and / or other effects of environmental physics. Once deployed, the DNN can accurately predict emergency responder alert types under various operating conditions.

[0005] Finally, using the location, direction of travel, and alert type can allow the vehicle to identify the emergency response vehicle and, in response, make planning and / or control decisions that are consistent with local regulations or practices. Furthermore, by using audio (and, in embodiments, perception) rather than perception alone, emergency response vehicles can be identified at an earlier time regardless of obstruction, thereby allowing the vehicle to make planning decisions in advance that contribute to the overall safety of the situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Hereinafter, the present system and method for emergency response vehicle detection for autonomous driving applications is described in detail with reference to the accompanying drawings, in which:

[0007] Figure 1A A data flow diagram depicting a process for emergency response vehicle detection according to an embodiment of the present disclosure;

[0008] Figure 1B Depicts an example architecture for a deep neural network in accordance with an embodiment of the present disclosure;

[0009] Figure 2 depicts an example arrangement of a microphone array on a vehicle in accordance with an embodiment of the present disclosure;

[0010] Figure 3-4 A flow chart depicting a method for emergency response vehicle detection according to an embodiment of the present disclosure;

[0011] Figure 5A is an illustration of an example autonomous vehicle in accordance with some embodiments of the present disclosure;

[0012] Figure 5B For use in accordance with some embodiments of the present disclosure Figure 5A Examples of camera positions and fields of view for autonomous vehicles;

[0013] Figure 5C For use in accordance with some embodiments of the present disclosure Figure 5A a block diagram of an example system architecture for an example autonomous vehicle;

[0014] Figure 5D For cloud-based servers and Figure 5A System diagram of an example of communication between autonomous vehicles;

[0015] Figure 6 is a block diagram of an example computing device suitable for implementing some embodiments of the present disclosure; and

[0016] Figure 7 is a block diagram of an example data center suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0017] Systems and methods are disclosed relating to emergency response vehicle detection for autonomous driving applications. Although the present disclosure may be directed to an example autonomous vehicle 500 (alternatively referred to herein as "vehicle 500" or "ego vehicle 500"), the examples herein are directed to Figures 5A-5D The present disclosure may be described with respect to autonomous driving, but this is not intended to be limiting. For example, the systems and methods described herein may be used by non-autonomous vehicles, semi-autonomous vehicles (e.g., in one or more advanced driver assistance systems (ADAS)), robots, warehouse vehicles, off-road vehicles, airships, ships, and / or other vehicle types. Furthermore, although the present disclosure may be described with respect to autonomous driving, this is not intended to be limiting. For example, the systems and methods described herein may be used in robotics, aerial systems, boating systems (e.g., emergency vessel identification), simulation environments (e.g., emergency response vehicle detection of virtual vehicles within a virtual simulation environment), and / or other technology areas.

[0018] Reference Figure 1A , Figure 1AA data flow diagram of a process 100 for emergency response vehicle detection in accordance with an embodiment of the present disclosure is depicted. It will be understood that this and other arrangements described herein are set forth merely as examples. Other arrangements and elements (e.g., machines, interfaces, functions, sorting, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Further, many of the elements described herein are functional entities that may be implemented as discrete or distributed components or in conjunction with other components and in any appropriate combination and location. The various functions described herein as being performed by entities may be implemented by hardware, firmware, and / or software. For example, the various functions may be implemented by a processor executing instructions stored in a memory.

[0019] The process 100 includes one or more microphones 102 (which may be similar to Figure 5A and Figure 5C For example, any number of microphones 102 may be used and may be mounted or otherwise provided on the vehicle 500 in any arrangement. Figure 2 , various microphone arrays 202A-202D may be mounted in the vehicle 200 (in an embodiment, which may be similar to Figures 5A-5D Each of the microphone arrays 202A-202D may include multiple microphones, such as two, three, four, or the like, and each microphone may include a unidirectional, omnidirectional, or other type of microphone. Different configurations of the microphone array 202 may be used. For example, a rectangular arrangement of four microphones in each array may be used, such as one microphone placed at each vertex of a square and soldered to a printed circuit board (PCB). As another example, a circular array of seven microphones may be used, including six microphones in the circle and one microphone in the center.

[0020] Depending on the embodiment, various positions of the microphone array 202 can be used. Since the vehicle 200 itself acts as a sound barrier, the microphone array 202 can be placed so that the vehicle 200 has minimal impact on the obstruction warning. In this way, by including a microphone array 202 on each side of the vehicle 200 (front, back, left, right), the vehicle 200 can not act as a barrier to at least one of the microphone arrays 202, thereby resulting in a higher quality audio signal and therefore a more accurate prediction. For example, four microphone arrays 202A-202D can be used, one for the front, back, left, and right of the vehicle 200. As another example, an additional fifth microphone array 202E can be placed on the top of the vehicle. In one embodiment, the top of the vehicle 200 may use only a single microphone array, such as microphone array 202E, which may result in greater exposure to wind, rain, dust, etc. In other examples, there may be more or fewer microphone arrays 202 used in different positions of the vehicle 200. For example, in some embodiments, the microphone and / or microphone array 202 can be positioned inside the vehicle 500 so that the resulting audio signal generated within the cab of the vehicle 500 can be compared with a known or learned audio signal within the cab of the vehicle 500 for an alert or warning of an emergency response vehicle.

[0021] The microphone array 202 can be placed in different locations on the vehicle 200 (e.g., corresponding to different parts of the vehicle 200). For example, on the left or right side of the vehicle 200, the microphone array 202 can be placed below the side mirror or below the door handle, and in the front or back of the vehicle 200, the microphone array 202 can be placed near the rear license plate, near the rearview camera, near the front license plate, above or below the windshield, and / or another location. In some embodiments, placing the microphones in these locations can expose the microphone array 202 to less dust, water, and / or wind during operation. For example, the rearview camera is typically placed in the cavity where the license plate is located, which can prevent water and dust from accumulating on the microphone array 202. Similarly, protection from dust, wind, and / or water can be provided below the rearview mirror or door handle.

[0022] Each microphone array 202 can be disposed in a housing, and the housing can be disposed on the vehicle 200. Thus, the housing is disposed on the exterior of the vehicle 200 and can therefore be designed to withstand the ingress of dust and moisture to reduce or eliminate corrosion or degradation of the microphones within each microphone array 202. In some embodiments, to withstand dust and moisture, the membrane or fabric covering of each housing can be constructed from a material that does not significantly attenuate audio signals while providing adequate protection against rain, snow, dirt, and dust, such as a hydrophobic foam spray, a moisture- and dust-resistant composite synthetic fabric, and the like. In some embodiments, to protect the microphone arrays 202 from wind forces, each of the microphone arrays 202 can include one or more windshields.

[0023] In some embodiments, the microphone may comprise an automotive-grade digital microelectromechanical system (MEMS) microphone connected to an automotive audio bus (A2B) transceiver device using pulse density modulation (PDM), for example, to simplify the microphone topology in vehicle 200. For example, such a configuration may allow for the use of low-cost, lightweight twisted-pair (TP) cabling to capture audio signals from one or more microphones mounted on the exterior of vehicle 200. In such an embodiment, an alert frequency from an emergency response vehicle at any point within a radius of vehicle 200 may be represented by an oversampled 1-bit PDM audio stream connected to each A2B transceiver device in a daisy-chain configuration. The audio data 104 may be converted to multi-channel time division multiplexed (TDM) data by the secondary node A2B transceiver before being passed or transmitted upstream (e.g., toward an electrical control unit (ECU) comprising an A2B master node device) to its nearest neighbor A2B node in the daisy-chain. For example, the A2B auxiliary node transceiver closest to the system aggregates the TDM data from each of the microphones (e.g., each microphone from each microphone array 202) before transmitting the TDM data to the A2B master node. The A2B master node can transmit the audio data 104 via a TDM audio port to a system on a chip (SoC) such as Figure 5C Another component of the system is the SoC 504. Control channel commands can be sent to the A2B nodes using a TP cable via an integrated circuit (I2C) interface between the SoC and the A2B transceiver. Each auxiliary node can include an A2B transceiver and one or more PDM microphones, which can be remotely powered in a manner that sufficient current can be supplied to all connected nodes using the same TP cable.

[0024] In an embodiment, the audio data 104 may be pre-processed using a pre-processor 106. For example, one or more background noise suppression algorithms may be used to suppress ambient noise such as wind, vehicle noise, road noise, and / or the like. For example, one or more beamforming algorithms may be implemented to perform background noise suppression on the audio data 104 generated by each microphone array 202. As used herein, audio data 104 may refer to raw audio data and / or pre-processed audio data.

[0025] In some embodiments, e.g., after pre-processing, the audio data 104 may be used by a position determiner 108 to determine an output 110. For example, the position determiner 108 may use one or more (e.g., passive) acoustic localization algorithms, e.g., acoustic triangulation, to determine a position 112 and / or direction of travel 114 of the emergency response vehicle. The position determiner 108 may analyze the audio data 104 from a plurality of, e.g., three or more, microphone arrays 202 to determine the position 112 of the emergency response vehicle using acoustic triangulation to determine a distance (e.g., from the vehicle 200) and / or a direction (e.g., an angular range defining an area of ​​the environment in which the emergency response vehicle is located) of the microphone arrays 202. For example, acoustic triangulation may be used to determine an estimated distance of the emergency response vehicle from the vehicle 200 and an estimated source direction of an alert from the emergency response vehicle. The more microphones or microphone arrays 202 that are used, the more accurate the triangulation may be. However, using four microphone arrays 202 (e.g., Figure 2202A-202D) can allow for accuracy within 30 degrees or less, which may be suitable for locating an emergency response vehicle and reacting to it. In other embodiments, there may be more or fewer microphone arrays 202 used, and the accuracy may vary accordingly (e.g., within forty degrees, twenty-five degrees, ten degrees, etc.). For non-limiting example, due to the geometry and overall layout of the road system, in addition to the distance metric, accuracy within thirty degrees can allow the vehicle 500 to identify which road the emergency response vehicle is located with an accuracy sufficient to respond in accordance with local rules and practices. In this way, when approaching a four-way intersection, the estimated position and distance can allow the vehicle 500 to determine that the emergency response vehicle is entering the intersection from the left on the road, and the direction of travel 114 described in more detail herein can be used to determine the direction of travel of the emergency response vehicle on the road. Thus, if the emergency response vehicle is traveling away from the intersection, vehicle 500 may determine to continue through the intersection without regard to the emergency response vehicle, whereas if the emergency response vehicle is traveling toward the intersection, vehicle 500 may determine to pull over and stop until the emergency response vehicle has passed the intersection (in examples where pulling over and stopping is a local rule or custom). As a result, in some embodiments, knowledge of the road layout (e.g., determined using GNSS maps, high-definition (HD) maps, vehicle perception, etc.) may additionally be used to determine location 112 and / or direction of travel 114.

[0026] The direction of travel 114 can be determined by the position determiner 108 by tracking the position 112 of the emergency response vehicle over time, such as over multiple frames or time steps. For example, changes (e.g., increases or decreases) in sound pressure, particle velocity, sound or audio frequency, and / or other physical quantities of the sound field, as represented by the audio data 104, can indicate that the emergency response vehicle is approaching (e.g., an increase in sound pressure) or moving away (e.g., a decrease in sound pressure).

[0027] In addition to or in lieu of the location 112 and the direction of travel 114, the process 100 can include determining the emergency vehicle alert type 122 using deep learning. For example, the audio data 104—e.g., before and / or after being preprocessed by the preprocessor 106—can be analyzed by the spectrogram generator 116 to generate a spectrogram 118. In some embodiments, a spectrogram 118 can be generated for each microphone array 202, and N instances of the DNN 120 (where N corresponds to the number of microphone arrays 202) can be used to calculate N different outputs corresponding to the alert type 122. In other embodiments, N input channels corresponding to the spectrograms 118 from respective microphone arrays 202 can be input to the same DNN 120 to calculate the alert type 122. In still other embodiments, the N signals from respective microphone arrays 202 can be preprocessed and / or enhanced to generate a single spectrogram 118, and this single spectrogram 118 can be used as input to the DNN 120.

[0028] To generate the spectrogram 118, the spectrogram generator 116 can convert the audio data 104 from each microphone array 202 into the frequency domain by extracting the Mel frequency coefficients. The Mel frequency coefficients can be used to generate the spectrogram 118, which can be used as input to one or more deep neural networks (DNNs) 120 to calculate an output (e.g., confidence, probability, etc.) indicating the alarm type 122. For example, the audio data 104 can be sampled within a window of a predetermined size (e.g., 1024, 2048, etc.), such as a digital representation of air pressure samples over time, with each jump of a predetermined size (e.g., 256, 512, etc.) performed to sample the next window. A fast Fourier transform (FFT) can be calculated for each window to transform the data from the time domain into the frequency domain. The entire spectrum can then be separated or divided into bins, e.g., 120 bins, and each bin can be transformed into a corresponding Mel bin at the Mel scale. For each window, the amplitude of the signal can be decomposed into its components corresponding to the frequency at the Mel scale. Thus, the y-axis corresponding to frequency can be converted to a logarithmic scale, the color dimension corresponding to amplitude can be converted to decibels to form a spectrogram, and in an embodiment, the y-axis corresponding to frequency can be mapped to a Mel scale to form a Mel spectrogram. Thus, the spectrogram 118 can correspond to a certain spectrogram and / or a Mel spectrogram.

[0029] The spectrogram 118, such as a traditional or Mel spectrogram, can be provided as an input to the DNN 120. In an embodiment, the DNN 120 can include a convolutional recurrent neural network (CRNN); however, this is not intended to be limiting. For example and without limitation, the DNN 120 can include any type of machine learning model, such as a machine learning model using linear regression, logistic regression, decision trees, support vector machines (SVMs), naive Bayes, k-nearest neighbors (KNN), K-means clustering, random forests, dimensionality reduction algorithms, gradient boosting algorithms, neural networks (e.g., autoencoders, convolutional, recurrent, perceptrons, long / short-term memory / LSTMs, Hopfield, Boltzmann, deep belief, deconvolution, generative adversarial, liquid machines, etc.), region of interest detection algorithms, computer vision algorithms, and / or other types of machine learning models.

[0030] A CRNN can be used as a DNN 120 to take into account the spatial and temporal properties of emergency alerts such as alarms, sirens, or vehicle horn patterns. For example, each alert can be generated by an instrument or audio emitting device on an emergency response vehicle to have a fixed pattern. Each emergency response vehicle can also have a different alert type, depending on the current state or situation (e.g., an ambulance may have a first alert when it is driving to a scene of unknown severity, but may have a second alert when it is driving to a hospital to transport a patient with only minor injuries). Furthermore, two different alerts may have the same signal, but one signal may extend over a longer period of time than the other, so identifying such differences is important for accurate classification of the alert type 122. Thus, the CRNN can take into account these fixed patterns by looking not only at the frequency, amplitude, and / or other sound representations, but also at the amount of time each of these sound representations is detected over a certain time window.

[0031] Reference Figure 1B , illustrates an example architecture for a DNN 120A (e.g., a CRNN). T may correspond to time, which may be a time window to which the spectrogram corresponds. In some embodiments, T may be 250 milliseconds (ms), 500 ms, 800 ms, or another time window. F may correspond to frequency. The DNN 120A may include a series of convolutions followed by a series of RNNs. Convolutional layers (or feature detector layers) may typically be used to learn unique spatial features of the spectrogram 118 that are useful for classification. Figure 1BIn the architecture of DNN 120A, convolutions are replaced with gated linear units (GLUs) 126, such as GLU 126A and GLU 126B, to improve accuracy and allow for faster convergence during training with smaller training data sets. The outputs of GLUs 126A and 126B can be multiplied and applied to a max pooling layer 128, a convolutional layer 130, and another max pooling layer 132. The output of max pooling layer 132 can be applied to one or more RNNs or stateful layers of DNN 120A. For example, the use of stateful RNNs can help recognize longer warning sequences. DNN 120A can use gated RNNs such as gated recurrent units 134A and 134B to improve accuracy on sequences (e.g., warnings generally follow a predetermined pattern of sounds) while also being lightweight and converging faster during training with smaller training data sets. Furthermore, to optimize between latency and accuracy, the DNN 120A can operate on each inferred window of audio data (as represented by the Mel spectrogram) while maintaining the state of the GRU 134, allowing subsequent inference to continue on the audio sequence. The outputs of the RNN, such as the GRU 134, can be multiplied and applied to one or more attention layers 140. For example, the output of the GRU 134 can be applied to a dense layer with a sigmoid activation function and a dense layer with a softmax activation function, and the outputs of these two dense layers can undergo a weighted average operation 142 to generate a final output indicating the confidence or probability of the alert type 122. The attention layer 140 can help determine where the alert begins within the time window defined by T. The attention layer 140 is therefore used to determine the probability of each type of alert across the entire time window, giving higher weight to alarms, sirens, horns, or other generated sound patterns, regardless of their temporal position within the time window. The output of independent probabilities for different alert types within a given time window allows multiple alert types 122 to be detected at any given time. For example, rather than using a confidence level equal to 1 for a certain number of alerts, a probability may be calculated for each alert type regardless of the other alert type probabilities or confidence levels. However, in some embodiments, a confidence level may be used.

[0032] As a result, the GLU 126 in the CRNN 100A can help the DNN learn optimal features for different types of alerts quickly and without the need for a large training dataset. Furthermore, the stateful GRU 134 combined with the attention layer 140 helps detect alerts within a certain time window while preserving long-term dependencies. This allows the DNN 120 to run faster, as opposed to detecting alerts sequentially for each time frame, e.g., without a temporal element. Furthermore, by breaking down the inference into smaller time windows, the approximation of the Doppler effect can be more accurate. Furthermore, using the attention layer 140 helps the DNN 120 detect alerts earlier while contributing negligibly to latency.

[0033] To train a DNN 120, such as CRNN 100A, a training dataset can be generated. However, because it is difficult to generate a real-world training dataset that includes sufficient variation in alert types and transformations thereto due to the physical properties of the environment, data augmentation can be used to train the DNN 120. For example, capturing real-world audio in which alerts are present is challenging, but capturing that same audio outdoors, in a tunnel, under different weather conditions, and / or in other different situations is even more challenging. Alert patterns may undergo a wide range of transformations, such as Doppler, attenuation, echo, reverberation, and / or the like, before being captured by microphone 102. In order for the DNN 120 to accurately predict the alert type 122, the DNN 120 can be trained to recognize the alert type 122 after these transformations have occurred.

[0034] Thus, to generate a robust training set, a dataset of audio data including warnings (such as, for example and without limitation, sirens, horns, alarms, or other sound patterns generated by emergency response vehicles) can be generated, for example, from a real-world collection using a data collection tool, from an audio track of a warning, from a video including associated audio of the warning, and / or the like. Instances of the training audio data can be labeled with semantic or class labels corresponding to the different warning types 122 represented therein. Instances of the training audio data can then be subjected to one or more transformations, such as time stretching, time shifting, pitch shifting, dynamic range compression at different signal-to-noise ratios (SNRs), noise compression at different SNRs, and / or other types of transformations. For example, a single instance can be used to generate any number of additional instances using different combinations of transformations. Thus, a training dataset can be augmented to generate an updated training set, including some multiple of the original training dataset size—e.g., 25x, 50x, and so on.

[0035] Refer again Figure 1A, alert type 122 may include the type of emergency response vehicle, for example, such that two or more alert types 122 may include the same name, such as "fire truck," "police car," or "ambulance." In other embodiments, alert type 122 may include different alert types that are not associated with the type of emergency response vehicle, such as, for example, "wailing," "screaming," "warbling," "horn," "piercing," "shouting," "howling," "priority," "dual tone," "rumble," and the like. In some embodiments, a combination of the two may be used as alert type 122, such as, for example, "police car: piercing" or "ambulance: wailing." Thus, depending on the output that DNN 120 is trained to detect, the probability or confidence output by DNN 120 may correspond to alert type 122 for the emergency response vehicle name, the alert name, or a combination thereof.

[0036] In some embodiments, in addition to identifying the alert type 122, the direction of travel 114, and / or the location 112, some or all of this information may be fused with additional types of information. For example, in conjunction with other sensors from the vehicle 500 (e.g., Figures 5A-5C As described, in the case of perception using a sensor, such as LiDAR, RADAR, camera, etc., the results from process 100 can be combined with the perception output of vehicle 500 for redundancy or fusion to improve the robustness and accuracy of the results. Thus, in the case of identifying and classifying an emergency response vehicle using object detection, for example, the perception output can be combined with the location 112, direction of travel 114, alert type 122 (or the type of emergency response vehicle as determined therefrom) to update or verify the prediction. Further, in some embodiments, vehicle-to-vehicle communications can include additional sources of information for improving the robustness or accuracy of the results. For example, other vehicles in the environment can share information with vehicle 500 regarding detections of emergency response vehicles (e.g., location, direction of travel, type, etc.).

[0037] The determined alert type 122—which may include one or more at any instance of the DNN 120—the direction of travel of the emergency response vehicle, the location of the emergency response vehicle, the perception of the vehicle 500, and / or vehicle-to-vehicle communication can be used by the autonomous driving software stack (e.g., the driver stack) 124 of the vehicle 500. For example, the perception layer of the driver stack 124 can use the process 100 to identify and locate the emergency response vehicle in order to update the world model (e.g., using a world model manager) to localize the emergency response vehicle to the world model. The planning layer of the driver stack 124 can use the location 112, direction of travel 114, and / or alert type to determine a route or path plan for the emergency response vehicle to slow down, pull over, stop, and / or perform another maneuver. The control layer of the driver stack 124 can then use the route or path plan to control the vehicle 500 according to the route. In some embodiments, the identification of the emergency response vehicle can trigger a remote or teleoperation request. For non-limiting example, the remote operation can be requested and performed similar to the remote operation described in U.S. Non-Provisional Application No. 16 / 366,506, filed on March 27, 2019, which is incorporated herein by reference in its entirety. In this way, the driver stack 124 can use the alert type 122 (and / or the corresponding emergency response vehicle type or emergency type as indicated by the alert type 122), the location 112, and / or the direction of travel 114 to perform one or more operations that account for the presence of the emergency response vehicle in the environment.

[0038] Now, refer to Figure 3-4 , each block of methods 300 and 400 described herein includes a computational process that can be performed using any combination of hardware, firmware, and / or software. For example, the various functions can be implemented by a processor executing instructions stored in a memory. Methods 300 and 400 can also be implemented as computer-usable instructions stored on a computer storage medium. To name a few examples, methods 300 and 400 can be provided by a stand-alone application, a service or a hosted service (stand-alone or in conjunction with another hosted service), or a plug-in to another product. In addition, methods 300 and 400 are provided by way of example to Figure 1A The process 100 and Figures 5A-5D However, these methods may additionally or alternatively be performed by any one process and / or any one system or any combination of processes and systems, including but not limited to those described herein.

[0039] Reference Figure 3 , Figure 3Flowchart illustrating a method 300 for emergency response vehicle detection according to some embodiments of the present disclosure. At block B302, method 300 includes receiving audio data generated using a plurality of microphones of an autonomous machine. For example, audio data 104 may be generated using microphone 102, such as microphone 102 of microphone array 202.

[0040] At block B304, method 300 includes executing an acoustic triangulation algorithm to determine at least one of a position or a direction of travel of the emergency response vehicle. For example, audio data 104 (before or after preprocessing by preprocessor 106) may be used by position determiner 108 to determine position 112 and / or direction of travel 114.

[0041] At block B306, process 300 includes generating a Mel spectrogram. For example, spectrogram generator 116 may generate a (Mel) spectrogram 118 (or another representation of a spectrum of one or more frequencies corresponding to one or more audio signals from the audio data) using audio data 104.

[0042] At block B308, process 300 includes applying first data representing the Mel spectrogram to the CRNN. ​​For example, data representing the (Mel) spectrogram 118 may be applied to the DNN 120 (eg, CRNN 110A).

[0043] At block B310 , the process 300 includes calculating second data representing probabilities of the plurality of alert types using the CRNN. ​​For example, the CRNN may calculate outputs corresponding to the probabilities of the plurality of alert types 122 .

[0044] At block B312, method 300 includes determining the type of emergency response vehicle based, at least in part, on these probabilities. For example, the alert type 122 with the highest probability (or a probability above a threshold) may be determined to be present, and the type of emergency response vehicle associated with the alert type 122 may be determined. In some embodiments, the threshold may be adjusted based on the location of the vehicle 500 and / or other information. For example, in addition to or in lieu of the location of the vehicle 500 near a hospital, fire station, police station, etc. (as determined using GNSS, HD maps, vehicle awareness, etc.), vehicle-to-vehicle communications indicating nearby accidents, SigAlerts, and / or types of structure, forest, roadside, or other fires may be used to lower the confidence or probability threshold due to an increased likelihood that an emergency response may be in the area.

[0045] At block B314, method 300 includes performing, by the autonomous machine, one or more operations based at least in part on the type, location, and / or direction of travel of the emergency response vehicle. For example, the type (and / or alert type 122), location 112, and / or direction of travel 114 of the emergency response vehicle may be used to perform one or more operations to comply with local regulations or practices regarding emergency response vehicles.

[0046] Reference Figure 4 , Figure 4 Flowchart illustrating a method 400 for emergency response vehicle detection according to some embodiments of the present disclosure. At block B402, method 400 includes receiving audio data generated using a plurality of microphones. For example, audio data 104 may be generated using microphone 102, such as microphone 102 of microphone array 202.

[0047] At block B404, process 400 includes generating a spectrogram.For example, spectrogram generator 116 may use audio data 104 to generate spectrogram 118 (or another representation of a spectrum of one or more frequencies corresponding to one or more audio signals from the audio data).

[0048] At block B406, process 400 includes applying first data representing the spectrogram to the DNN. For example, data representing spectrogram 118 may be applied to DNN 120 (e.g., CRNN 110A).

[0049] At block B408 , process 400 includes computing second data based at least in part on the first data using one or more feature extraction layers of the DNN. For example, GLU 126 may be used to compute a feature map or feature vector using data representing spectrogram 118 .

[0050] At block B410, process 400 includes computing third data based at least in part on the second data using one or more stateful layers of the DNN. For example, GRU 134 may compute output from the output of GLU 126, for example, before or after processing by one or more additional layers such as layers 128, 130, and / or 132.

[0051] At block B412, process 400 includes computing, using one or more attention layers of the DNN and based at least in part on the third data, fourth data representing probabilities of the plurality of alert types. For example, dense layer 140 of DNN 120 may be used to compute outputs indicating probabilities of the plurality of alert types 122.

[0052] At block B414, method 400 includes performing one or more actions based at least in part on the probability. For example, the type of emergency response vehicle (and / or alert type 122), location 112, and / or direction of travel 114 may be used to perform one or more actions to comply with local regulations or practices regarding emergency response vehicles.

[0053] Example autonomous vehicle

[0054] Figure 5A FIG2 is an illustration of an example autonomous vehicle 500 in accordance with some embodiments of the present disclosure. Autonomous vehicle 500 (alternatively referred to herein as “vehicle 500”) may include, but is not limited to, a passenger vehicle, such as a car, truck, bus, first responder vehicle, shuttle, electric or motorized bicycle, motorcycle, fire truck, police car, ambulance, boat, construction vehicle, underwater vehicle, drone, and / or another type of vehicle (e.g., unmanned and / or accommodating one or more passengers). Autonomous vehicles are generally described according to 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, issued June 15, 2018, Standard No. J3016-201609, issued September 30, 2016, and prior and future versions of such standards). The vehicle 500 may be capable of implementing functionality consistent with one or more of autonomous driving levels 3 to 5. For example, depending on the embodiment, the vehicle 500 may be capable of conditional automation (Level 3), high automation (Level 4), and / or full automation (Level 5).

[0055] Vehicle 500 may include components such as a chassis, a body, wheels (e.g., 2, 4, 6, 8, 18, etc.), tires, axles, and other vehicle components. Vehicle 500 may include a propulsion system 550, such as an internal combustion engine, a hybrid power plant, an all-electric engine, and / or another type of propulsion system. Propulsion system 550 may be connected to a drivetrain of vehicle 500, which may include a transmission, to achieve propulsion of vehicle 500. Propulsion system 550 may be controlled in response to receiving a signal from throttle / accelerator 552.

[0056] A steering system 554, which may include a steering wheel, may be used to steer vehicle 500 (e.g., along a desired path or route) when propulsion system 550 is operating (e.g., when the vehicle is in motion). Steering system 554 may receive signals from steering actuator 556. For fully automated (Level 5) functionality, a steering wheel may be optional.

[0057] Brake sensor system 546 may be used to operate vehicle brakes in response to receiving signals from brake actuator 548 and / or brake sensors.

[0058] May include one or more system on chip (SoC) 504 ( Figure 5C ) and / or one or more GPUs can provide signals (e.g., representing commands) to one or more components and / or systems of the vehicle 500. For example, the one or more controllers can send signals to operate the vehicle brakes via one or more brake actuators 548, to operate the steering system 554 via one or more steering actuators 556, and to operate the propulsion system 550 via one or more throttles / accelerators 552. The one or more controllers 536 can include one or more onboard (e.g., integrated) computing devices (e.g., supercomputers) that process sensor signals and output operational commands (e.g., signals representing commands) to enable autonomous driving and / or assist a human driver in driving the vehicle 500. The one or more controllers 536 can include a first controller 536 for autonomous driving functions, a second controller 536 for functional safety functions, a third controller 536 for artificial intelligence functions (e.g., computer vision), a fourth controller 536 for infotainment functions, a fifth controller 536 for redundancy in emergency situations, and / or other controllers. In some examples, a single controller 536 may handle two or more of the above functions, two or more controllers 536 may handle a single function, and / or any combination thereof.

[0059] The one or more controllers 536 may provide signals for controlling one or more components and / or systems of the vehicle 500 in response to sensor data (e.g., sensor inputs) received from one or more sensors. The sensor data may be received from, for example and without limitation, a global navigation satellite system sensor 558 (e.g., a global positioning system sensor), a RADAR sensor 560, an ultrasonic sensor 562, a LIDAR sensor 564, an inertial measurement unit (IMU) sensor 566 (e.g., an accelerometer, a gyroscope, a magnetic compass, a magnetometer, etc.), a microphone 596, a stereo camera 568, a wide-angle camera 570 (e.g., a fisheye camera), an infrared camera 572, a surround camera 574 (e.g., a 360-degree camera), a long-range and / or mid-range camera 598, a speed sensor 544 (e.g., for measuring the velocity of the vehicle 500), a vibration sensor 542, a steering sensor 540, a brake sensor (e.g., as part of a brake sensor system 546), and / or other sensor types.

[0060] One or more of the controllers 536 may receive input (e.g., represented by input data) from the instrument cluster 532 of the vehicle 500 and provide output (e.g., represented by output data, display data, etc.) via a human machine interface (HMI) display 534, an audible annunciator, a speaker, and / or via other components of the vehicle 500. These outputs may include information such as vehicle speed, velocity, time, map data (e.g., Figure 5C The HMI display 534 may include information such as the HD map 522 of the vehicle 500 , location data (e.g., the location of the vehicle 500 on the map), directions, the locations of other vehicles (e.g., an occupancy grid), information about objects and object states as sensed by the controller 536 , and the like. For example, the HMI display 534 may display information about the presence of one or more objects (e.g., street signs, warning signs, traffic light changes, etc.) and / or information about driving maneuvers that the vehicle has made, is making, or will make (e.g., changing lanes now, leaving 34B in two miles, etc.).

[0061] The vehicle 500 also includes a network interface 524 that can communicate over one or more networks using one or more wireless antennas 526 and / or a modem. For example, the network interface 524 can be capable of communicating over LTE, WCDMA, UMTS, GSM, CDMA2000, etc. The one or more wireless antennas 526 can also enable communication between objects in the environment (e.g., vehicles, mobile devices, etc.) using one or more local area networks such as Bluetooth, Bluetooth LE, Z-wave, ZigBee, etc. and / or one or more low power wide area networks (LPWANs) such as LoRaWAN, SigFox, etc.

[0062] Figure 5BFor use in accordance with some embodiments of the present disclosure Figure 5A An example of camera positions and fields of view for an autonomous vehicle 500 is shown. The cameras and respective fields of view are an example embodiment and are not intended to be limiting. For example, additional and / or alternative cameras may be included and / or located at different locations on the vehicle 500.

[0063] The camera type used for the camera may include, but is not limited to, a digital camera that may be suitable for use with components and / or systems of the vehicle 500. The camera may operate under Automotive Safety Integrity Level (ASIL) B and / or under another ASIL. The camera type may have any image capture rate, such as 60 frames per second (fps), 120fps, 240fps, and the like, depending on the embodiment. The camera may be capable of using a rolling shutter, a global shutter, another type of shutter, or a combination thereof. In some examples, the color filter array may include a red-white-white-white (RCCC) color filter array, a red-white-white-blue (RCCB) color filter array, a red-blue-green-white (RBGC) color filter array, a Foveon X3 color filter array, a Bayer sensor (RGGB) color filter array, a monochrome sensor color filter array, and / or another type of color filter array. In some embodiments, a clear pixel camera such as a camera with an RCCC, RCCB, and / or RBGC color filter array may be used in an effort to improve light sensitivity.

[0064] In some examples, one or more of the cameras can be used to perform advanced driver assistance system (ADAS) functions (e.g., as part of a redundant or fail-safe design). For example, a multi-function monocular camera can be installed to provide functions including lane departure warning, traffic sign assistance, and intelligent headlight control. One or more of the cameras (e.g., all of the cameras) can simultaneously record and provide image data (e.g., video).

[0065] One or more of the cameras can be mounted in a mounting assembly, such as a custom-designed (3-D printed) assembly, to cut off stray light and reflections from within the car (e.g., reflections from the dashboard reflected in the windshield mirror) that could interfere with the camera's ability to capture image data. With respect to the wing mirror mounting assembly, the wing mirror assembly can be custom 3-D printed so that the camera mounting plate matches the shape of the wing mirror. In some examples, one or more cameras can be integrated into the wing mirror. For side-view cameras, one or more cameras can also be integrated into the four pillars at each corner of the cab.

[0066] A camera (e.g., a front-facing camera) having a field of view that includes a portion of the environment in front of the vehicle 500 can be used for surround view to help identify the forward path and obstacles, as well as assist in providing information critical to generating an occupancy grid and / or determining a preferred vehicle path with the help of one or more controllers 536 and / or control SoCs. The front-facing camera can be used to perform many of the same ADAS functions as LIDAR, including emergency braking, pedestrian detection, and collision avoidance. The front-facing camera can also be used for ADAS functions and systems, including lane departure warning ("LDW"), autonomous cruise control ("ACC"), and / or other functions such as traffic sign recognition.

[0067] A variety of cameras can be used in the front-facing configuration, including, for example, a monocular camera platform including a CMOS (complementary metal oxide semiconductor) color imager. Another example could be a wide-angle camera 570, which can be used to sense objects entering the field of view from the periphery (e.g., pedestrians, intersection traffic, or bicycles). Although Figure 5B The figure shows only one wide-angle camera, but any number of wide-angle cameras 570 can be present on the vehicle 500. In addition, long-range cameras 598 (e.g., a pair of long-view stereo cameras) can be used for depth-based object detection, especially for objects for which neural networks have not yet been trained. Long-range cameras 598 can also be used for object detection and classification and basic object tracking.

[0068] One or more stereo cameras 568 may also be included in the front configuration. The stereo camera 568 may include an integrated control unit including a scalable processing unit that may provide a multi-core microprocessor and programmable logic (FPGA) with an integrated CAN or Ethernet interface on a single chip. Such a unit may be used to generate a 3-D map of the vehicle environment, including distance estimates for all points in the image. An alternative stereo camera 568 may include a compact stereo vision sensor that may include two camera lenses (one on the left and one on the right) and an image processing chip that may measure the distance from the vehicle to the target object and use the generated information (e.g., metadata) to activate autonomous emergency braking and lane departure warning functions. Other types of stereo cameras 568 may be used in addition to or alternatively to those described herein.

[0069] Cameras with a field of view that includes a portion of the environment to the sides of the vehicle 500 (e.g., side view cameras) can be used for surround view, providing information used to create and update the occupancy grid and generate side impact collision warnings. For example, surround camera 574 (e.g., Figure 5BFour surround cameras 574 (shown in FIG) can be placed on the vehicle 500. The surround cameras 574 can include a wide-angle camera 570, a fisheye camera, a 360-degree camera, and / or the like. For example, the four fisheye cameras can be placed on the front, rear, and sides of the vehicle. In an alternative arrangement, the vehicle can use three surround cameras 574 (e.g., left, right, and rear), and can utilize one or more other cameras (e.g., a forward-facing camera) as a fourth surround-view camera.

[0070] A camera having a field of view that includes a portion of the environment behind the vehicle 500 (e.g., a rearview camera) can be used to assist with parking, surround view, rear collision warning, and creating and updating an occupancy grid. A variety of cameras can be used, including but not limited to cameras that are also suitable as front-facing cameras as described herein (e.g., long-range and / or mid-range cameras 598, stereo cameras 568, infrared cameras 572, etc.).

[0071] Figure 5C For use in accordance with some embodiments of the present disclosure Figure 5A 5. Block diagram of an example system architecture for an example autonomous vehicle 500. It should be understood that this arrangement and other arrangements described herein are set forth merely as examples. Other arrangements and elements (e.g., machines, interfaces, functions, sequences, functional groupings, etc.) may be used in addition to or in place of those shown, and some elements may be omitted entirely. Further, many of the elements described herein are functional entities that may be implemented as discrete or distributed components or in conjunction with other components, and in any appropriate combination and location. The various functions described herein as being performed by entities may be implemented by hardware, firmware, and / or software. For example, the various functions may be implemented by a processor executing instructions stored in memory.

[0072] Figure 5C Each of the components, features, and systems of the vehicle 500 is illustrated as being connected via a bus 502. The bus 502 may include a controller area network (CAN) data interface (alternatively, referred to herein as a "CAN bus"). The CAN may be a network internal to the vehicle 500 that assists in controlling various features and functions of the vehicle 500, such as actuation of brakes, acceleration, braking, steering, windshield wipers, and the like. The CAN bus may be configured to have tens or even hundreds of nodes, each with its own unique identifier (e.g., a CAN ID). The CAN bus may be read to find steering wheel angle, ground speed, engine revolutions per minute (RPM), button positions, and / or other vehicle status indicators. The CAN bus may be ASIL B compliant.

[0073] Although bus 502 is described here as a CAN bus, this is not intended to be limiting. For example, FlexRay and / or Ethernet may be used in addition to or in lieu of a CAN bus. Furthermore, although bus 502 is represented by a single line, this is not intended to be limiting. For example, there may be any number of buses 502, including one or more CAN buses, one or more FlexRay buses, one or more Ethernet buses, and / or one or more other types of buses using different protocols. In some examples, two or more buses 502 may be used to perform different functions and / or may be used for redundancy. For example, a first bus 502 may be used for collision avoidance functionality, and a second bus 502 may be used for drive control. In any example, each bus 502 may communicate with any component of vehicle 500, and two or more buses 502 may communicate with the same component. In some examples, each SoC 504, each controller 536, and / or each computer within the vehicle may have access to the same input data (e.g., input from sensors on vehicle 500) and may be connected to a common bus such as a CAN bus.

[0074] The vehicle 500 may include one or more controllers 536, such as those described herein. Figure 5A Controller 536 may be used for a variety of functions. Controller 536 may be coupled to any of the various other components and systems of vehicle 500 and may be used for control of vehicle 500, artificial intelligence of vehicle 500, infotainment for vehicle 500, and / or the like.

[0075] The vehicle 500 may include one or more system on a chip (SoC) 504. The SoC 504 may include a CPU 506, a GPU 508, a processor 510, a cache 512, an accelerator 514, a data store 516, and / or other components and features not shown. The SoC 504 may be used to control the vehicle 500 in a variety of platforms and systems. For example, the one or more SoCs 504 may be combined with an HD map 522 in a system (e.g., a system of the vehicle 500), and the HD map may be downloaded from one or more servers (e.g., a server) via a network interface 524. Figure 5D one or more servers 578) to obtain map refreshes and / or updates.

[0076] The CPU 506 may include a CPU cluster or CPU complex (alternatively, referred to herein as a "CCPLEX"). The CPU 506 may include multiple cores and / or L2 caches. For example, in some embodiments, the CPU 506 may include eight cores in a coherent multiprocessor configuration. In some embodiments, the CPU 506 may include four dual-core clusters, each with a dedicated L2 cache (e.g., a 2MB L2 cache). The CPU 506 (e.g., CCPLEX) may be configured to support simultaneous cluster operations such that any combination of CPU 506 clusters can be active at any given time.

[0077] The CPU 506 may implement power management capabilities including one or more of the following features: each hardware block may be automatically clock gated when idle to conserve dynamic power; each core clock may be gated when the core is not actively executing instructions due to the execution of WFI / WFE instructions; each core may be independently power gated; each core cluster may be independently clock gated when all cores are clock gated or power gated; and / or each core cluster may be independently power gated when all cores are power gated. The CPU 506 may further implement an enhanced algorithm for managing power states, wherein allowed power states and expected wakeup times are specified, and hardware / microcode determines the optimal power state to enter for the core, cluster, and CCPLEX. The processing core may support a simplified power state entry sequence in software, with this work being offloaded to the microcode.

[0078] The GPU 508 may include an integrated GPU (alternatively referred to herein as an "iGPU"). The GPU 508 may be programmable and efficient for parallel workloads. In some examples, the GPU 508 may use an enhanced tensor instruction set. The GPU 508 may include one or more streaming microprocessors, each of which may include an L1 cache (e.g., an L1 cache with at least 96KB of storage capacity), and two or more of these streaming microprocessors may share an L2 cache (e.g., an L2 cache with 512KB of storage capacity). In some embodiments, the GPU 508 may include at least eight streaming microprocessors. The GPU 508 may use a computing application programming interface (API). In addition, the GPU 508 may use one or more parallel computing platforms and / or programming models (e.g., NVIDIA's CUDA).

[0079] In the case of automotive and embedded use, GPU 508 can be power optimized to achieve optimal performance. For example, GPU 508 can be manufactured on fin field effect transistors (FinFETs). However, this is not intended to be limiting, and GPU 508 can be manufactured using other semiconductor manufacturing processes. Each streaming microprocessor can merge several mixed precision processing cores divided into multiple blocks. For example and without limitation, 64 PF32 cores and 32 PF64 cores can be divided into four processing blocks. In such an example, each processing block can be allocated 16 FP32 cores, 8 FP64 cores, 16 INT32 cores, two mixed precision NVIDIA tensor cores for deep learning matrix arithmetic, L0 instruction cache, warp scheduler, dispatch unit and / or 64KB register file. In addition, the streaming microprocessor may include independent parallel integer and floating point data paths to provide efficient execution of workloads using a mix of computation and addressing calculations. The streaming microprocessor may include independent thread scheduling capabilities to allow for finer-grained synchronization and collaboration between parallel threads. Streaming microprocessors may include a combined L1 data cache and shared memory unit to increase performance while simplifying programming.

[0080] The GPU 508 can include high bandwidth memory (HBM) and / or a 16GB HBM2 memory subsystem that provides a peak memory bandwidth of approximately 900 GB / s in some examples. In some examples, synchronous graphics random access memory (SGRAM), such as fifth generation graphics double data rate synchronous random access memory (GDDR5), can be used in addition to or in lieu of HBM memory.

[0081] The GPU 508 may include unified memory technology that includes access counters to allow memory pages to be more accurately migrated to the processor that accesses them most frequently, thereby improving the efficiency of memory ranges shared between processors. In some examples, address translation services (ATS) support may be used to allow the GPU 508 to directly access the CPU 506 page tables. In such an example, when the GPU 508 memory management unit (MMU) experiences a miss, an address translation request may be transmitted to the CPU 506. In response, the CPU 506 may look up the virtual-to-physical mapping for the address in its page table and transmit the translation back to the GPU 508. In this way, unified memory technology may allow a single unified virtual address space to be used for memory of both the CPU 506 and the GPU 508, thereby simplifying GPU 508 programming and porting applications to the GPU 508.

[0082] In addition, GPU 508 can include access counters that can track how often GPU 508 accesses the memory of other processors. Access counters can help ensure that memory pages are moved to the physical memory of the processor that accesses them most frequently.

[0083] SoC 504 may include any number of caches 512, including those described herein. For example, cache 512 may include an L3 cache available to both CPU 506 and GPU 508 (e.g., connected to both CPU 506 and GPU 508). Cache 512 may include a write-back cache that can track the state of lines, for example, using a cache coherence protocol (e.g., MEI, MESI, MSI, etc.). Depending on the embodiment, the L3 cache may include 4MB or more, although smaller cache sizes may also be used.

[0084] The SoC 504 may include an arithmetic logic unit (ALU) that can be utilized to perform processing for any of the various tasks or operations described with respect to the vehicle 500, such as processing a DNN. Furthermore, the SoC 504 may include a floating point unit (FPU) or other type of math coprocessor or digital coprocessor for performing mathematical operations within the system. For example, the SoC 104 may include one or more FPUs integrated within the CPU 506 and / or the GPU 508 as execution units.

[0085] SoC 504 may include one or more accelerators 514 (e.g., hardware accelerators, software accelerators, or a combination thereof). For example, SoC 504 may include a hardware acceleration cluster that may include optimized hardware accelerators and / or large on-chip memory. The large on-chip memory (e.g., 4MB SRAM) may enable the hardware acceleration cluster to accelerate neural networks and other calculations. The hardware acceleration cluster may be used to supplement GPU 508 and offload some tasks of GPU 508 (e.g., freeing up more cycles of GPU 508 to perform other tasks). As an example, accelerator 514 may be used for targeted workloads (e.g., perception, convolutional neural networks (CNNs), etc.) that are stable enough to easily control acceleration. When used herein, the term "CNN" may include all types of CNNs, including region-based or regional convolutional neural networks (RCNNs) and fast RCNNs (e.g., for object detection).

[0086] The accelerator 514 (e.g., a hardware acceleration cluster) may include a deep learning accelerator (DLA). The DLA may include one or more tensor processing units (TPUs) that can be configured to provide an additional 10 trillion operations per second for deep learning applications and reasoning. The TPU may be an accelerator configured to perform image processing functions (e.g., for CNN, RCNN, etc.) and optimized for performing image processing functions. The DLA may be further optimized for a specific set of neural network types and floating-point operations and reasoning. The design of the DLA may provide higher performance per millimeter than a general-purpose GPU and far exceed the performance of the CPU. The TPU may perform several functions, including a single-instance convolution function, support for INT8, INT16, and FP16 data types for both features and weights, and post-processor functions.

[0087] The DLA can quickly and efficiently execute neural networks, particularly CNNs, on processed or unprocessed data for any of a wide variety of functions, such as, but not limited to: a CNN for object recognition and detection using data from a camera sensor; a CNN for distance estimation using data from a camera sensor; a CNN for emergency response vehicle detection and identification and detection using data from a microphone; a CNN for facial recognition and vehicle owner identification using data from a camera sensor; and / or a CNN for safety and / or security-related events.

[0088] The DLA can perform any function of the GPU 508, and by using an inference accelerator, for example, the designer can target any function to either the DLA or the GPU 508. For example, the designer can focus the processing of CNNs and floating-point operations on the DLA and leave other functions to the GPU 508 and / or other accelerators 514.

[0089] The accelerator 514 (e.g., a hardware acceleration cluster) may include a programmable vision accelerator (PVA), which may be alternatively referred to herein as a computer vision accelerator. The PVA may be designed and configured to accelerate computer vision algorithms for advanced driver assistance systems (ADAS), autonomous driving, and / or augmented reality (AR) and / or virtual reality (VR) applications. The PVA may provide a balance between performance and flexibility. For example, each PVA may include, for example and without limitation, any number of reduced instruction set computer (RISC) cores, direct memory access (DMA), and / or any number of vector processors.

[0090] The RISC core can interact with an image sensor (e.g., an image sensor of any camera described herein), an image signal processor, and / or the like. Each of these RISC cores can include any amount of memory. Depending on the embodiment, the RISC core can use any of a number of protocols. In some examples, the RISC core can execute a real-time operating system (RTOS). The RISC core can be implemented using one or more integrated circuit devices, application specific integrated circuits (ASICs), and / or memory devices. For example, the RISC core can include an instruction cache and / or tightly coupled RAM.

[0091] The DMA can enable components of the PVA to access system memory independently of the CPU 506. The DMA can support any number of features used to provide optimizations for the PVA, including but not limited to support for multi-dimensional addressing and / or circular addressing. In some examples, the DMA can support addressing in up to six or more dimensions, which can include block width, block height, block depth, horizontal block stride, vertical block stride, and / or depth stride.

[0092] A vector processor can be a programmable processor that can be designed to efficiently and flexibly execute programming for computer vision algorithms and provide signal processing capabilities. In some examples, the PVA can include a PVA core and two vector processing subsystem partitions. The PVA core can include a processor subsystem, one or more DMA engines (e.g., two DMA engines), and / or other peripherals. The vector processing subsystem can operate as the main processing engine of the PVA and can include a vector processing unit (VPU), an instruction cache, and / or a vector memory (e.g., VMEM). The VPU core can include a digital signal processor, such as, for example, a single instruction multiple data (SIMD), a very long instruction word (VLIW) digital signal processor. The combination of SIMD and VLIW can enhance throughput and speed.

[0093] Each of the vector processors can include an instruction cache and can be coupled to dedicated memory. As a result, in some examples, each of the vector processors can be configured to execute independently of the other vector processors. In other examples, the vector processors included in a particular PVA can be configured to employ data parallelism. For example, in some embodiments, multiple vector processors included in a single PVA can execute the same computer vision algorithm, but on different regions of an image. In other examples, the vector processors included in a particular PVA can execute different computer vision algorithms simultaneously on the same image, or even execute different algorithms on sequential images or portions of images. Among other things, any number of PVAs can be included in a hardware acceleration cluster, and any number of vector processors can be included in each of these PVAs. In addition, the PVAs can include additional error correction code (ECC) memory to enhance overall system security.

[0094] The accelerator 514 (e.g., a hardware acceleration cluster) may include an on-chip computer vision network and SRAM to provide high bandwidth, low latency SRAM for the accelerator 514. In some examples, the on-chip memory may include at least 4MB of SRAM consisting of, for example and without limitation, eight field-configurable memory blocks that can be accessed by both the PVA and the DLA. Each pair of memory blocks may include an advanced peripheral bus (APB) interface, configuration circuitry, a controller, and a multiplexer. Any type of memory may be used. The PVA and DLA may access the memory via a backbone that provides high-speed memory access to the PVA and DLA. The backbone may include an on-chip computer vision network that interconnects the PVA and DLA to the memory (e.g., using APB).

[0095] The on-chip computer vision network can include an interface that ensures that both the PVA and DLA provide ready and valid signals before transmitting any control signals / addresses / data. Such an interface can provide separate phases and separate channels for transmitting control signals / addresses / data, as well as burst-based communication for continuous data transmission. This type of interface can comply with ISO 26262 or IEC 61508 standards, but other standards and protocols can also be used.

[0096] In some examples, the SoC 504 may include a real-time ray tracing hardware accelerator such as that described in U.S. patent application Ser. No. 16 / 101,232, filed on Aug. 10, 2018. The real-time ray tracing hardware accelerator may be used to quickly and efficiently determine the position and extent of objects (e.g., within a world model) to generate real-time visualization simulations for use in RADAR signal interpretation, for sound propagation synthesis and / or analysis, for SONAR system simulations, for general wave propagation simulations, for comparison with LIDAR data for positioning and / or other functions, and / or for other uses. In some embodiments, one or more tree traversal units (TTUs) may be used to perform one or more ray tracing related operations.

[0097] The accelerator 514 (e.g., a hardware accelerator cluster) has a wide range of uses in autonomous driving. The PVA can be a programmable vision accelerator that can be used in key processing stages in ADAS and autonomous vehicles. The capabilities of the PVA are a good match for algorithmic domains that require predictable processing, low power, and low latency. In other words, the PVA performs well on semi-intensive or intensive rule computations, and even on small data sets that require predictable runtimes with low latency and low power. Therefore, in the context of a platform for autonomous vehicles, the PVA is designed to run classic computer vision algorithms because they are efficient at object detection and integer math operations.

[0098] For example, according to one embodiment of the technology, PVA is used to perform computer stereo vision. In some examples, a semi-global matching-based algorithm can be used, but this is not intended to be limiting. Many applications for Level 3-5 autonomous driving require on-the-fly motion estimation / stereo matching (e.g., structure from motion, pedestrian recognition, lane detection, etc.). PVA can perform computer stereo vision functions on input from two monocular cameras.

[0099] In some examples, PVA can be used to perform dense optical flow, by processing raw RADAR data (e.g., using a 4D Fast Fourier Transform) to provide processed RADAR. In other examples, PVA is used for time-of-flight depth processing, by processing raw time-of-flight data to provide processed time-of-flight data.

[0100] The DLA can be used to run any type of network to enhance control and driving safety, including, for example, a neural network that outputs a confidence measure for each object detection. Such confidence values ​​can be interpreted as probabilities, or as providing a relative "weight" of each detection compared to other detections. This confidence value enables the system to make further decisions about which detections should be considered true positives versus false positives. For example, the system can set a threshold for confidence and only consider detections that exceed the threshold as true positives. In an automatic emergency braking (AEB) system, a false positive detection could cause the vehicle to automatically apply emergency braking, which is clearly undesirable. Therefore, only the most confident detections should be considered triggers for AEB. The DLA can run a neural network to regress the confidence value. This neural network can take as its input at least some subset of parameters, such as bounding box dimensions, a ground plane estimate obtained (e.g., from another subsystem), inertial measurement unit (IMU) sensor 566 output related to the vehicle 500's orientation and distance, and 3D position estimates of objects obtained from the neural network and / or other sensors (e.g., LIDAR sensor 564 or RADAR sensor 560).

[0101] SoC 504 may include one or more data stores 516 (e.g., memory). Data stores 516 may be on-chip memory of SoC 504 that may store neural networks to be executed on the GPU and / or DLA. In some examples, data stores 516 may be large enough to store multiple instances of the neural network for redundancy and safety. Data stores 512 may include L2 or L3 cache 512. References to data stores 516 may include references to memory associated with the PVA, DLA, and / or other accelerators 514 as described herein.

[0102] SoC 504 may include one or more processors 510 (e.g., embedded processors). Processors 510 may include a boot and power management processor, which may be a dedicated processor and subsystem for handling boot power and management functions and related safety implementations. The boot and power management processor may be part of the SoC 504 boot sequence and may provide runtime power management services. The boot power and management processor may provide clock and voltage programming, auxiliary system low power state transitions, SoC 504 thermal and temperature sensor management, and / or SoC 504 power state management. Each temperature sensor may be implemented as a ring oscillator whose output frequency is proportional to temperature, and SoC 504 may use the ring oscillator to detect the temperature of CPU 506, GPU 508, and / or accelerator 514. If it is determined that the temperature exceeds a threshold, the boot and power management processor may enter a temperature fault routine and place SoC 504 in a lower power state and / or place vehicle 500 in a driver safety parking mode (e.g., to safely park vehicle 500).

[0103] The processor 510 may also include a set of embedded processors that can be used as an audio processing engine. The audio processing engine can be an audio subsystem that allows for full hardware support for multi-channel audio through multiple interfaces and a wide range of flexible audio I / O interfaces. In some examples, the audio processing engine is a dedicated processor core having a digital signal processor with dedicated RAM.

[0104] The processor 510 may also include an always-on processor engine that can provide the necessary hardware features to support low-power sensor management and wake-up use cases. The always-on processor engine may include a processor core, tightly coupled RAM, supporting peripherals (such as timers and interrupt controllers), various I / O controller peripherals, and routing logic.

[0105] The processor 510 may also include a safety cluster engine, which includes a dedicated processor subsystem that handles safety management of automotive applications. The safety cluster engine may include two or more processor cores, tightly coupled RAM, supporting peripherals (such as timers, interrupt controllers, etc.), and / or routing logic. In safety mode, the two or more cores can operate in lockstep mode and act as a single core with comparison logic to detect any differences between their operations.

[0106] Processor 510 may also include a real-time camera engine, which may include a dedicated processor subsystem for handling real-time camera management.

[0107] Processor 510 may also include a high dynamic range signal processor, which may include an image signal processor, which is a hardware engine that is part of the camera processing pipeline.

[0108] The processor 510 may include a video image compositer, which may be a processing block (e.g., implemented on a microprocessor) that implements the video post-processing functions required by the video playback application to produce the final image for the player window. The video image compositer may perform lens distortion correction for the wide-angle camera 570, the surround camera 574, and / or for the in-cab monitoring camera sensor. The in-cab monitoring camera sensor is preferably monitored by a neural network running on another instance of the advanced SoC, configured to recognize in-cab events and respond accordingly. The in-cab system may perform lip reading to activate mobile phone service and place calls, dictate emails, change vehicle destinations, activate or change the vehicle's infotainment system and settings, or provide voice-activated web surfing. Certain functions are only available to the driver when the vehicle is operating in autonomous mode and are disabled in other circumstances.

[0109] The video image compositer can include enhanced temporal noise reduction for both spatial and temporal noise reduction. For example, in the presence of motion in the video, the noise reduction appropriately weights spatial information and downweights information provided by neighboring frames. In the case where an image or portion of an image does not include motion, the temporal noise reduction performed by the video image compositer can use information from previous images to reduce noise in the current image.

[0110] The video image compositor can also be configured to perform stereo rectification on the input stereo footage frames. The video image compositor can further be used for user interface composition when the operating system desktop is in use and the GPU 508 does not need to continuously render new surfaces. Even when the GPU 508 is powered on and active for 3D rendering, the video image compositor can be used to offload the GPU 508 to improve performance and responsiveness.

[0111] The SoC 504 may also include a Mobile Industry Processor Interface (MIPI) camera serial interface, a high-speed interface for receiving video and input from a camera, and / or a video input block that may be used for camera and related pixel input functions. The SoC 504 may also include an input / output controller that may be controlled by software and may be used to receive I / O signals that are not assigned to a specific role.

[0112] SoC 504 may also include a wide range of peripheral interfaces to enable communication with peripherals, audio codecs, power management, and / or other devices. SoC 504 may be used to process data from cameras (connected via Gigabit multimedia serial links and Ethernet), sensors (e.g., LIDAR sensor 564, RADAR sensor 560, etc., which may be connected via Ethernet), data from bus 502 (e.g., vehicle 500 speed, steering wheel position, etc.), and data from GNSS sensor 558 (connected via Ethernet or CAN bus). SoC 504 may also include dedicated high-performance mass storage controllers, which may include their own DMA engines, and which may be used to free up CPU 506 from routine data management tasks.

[0113] SoC 504 can be an end-to-end platform with a flexible architecture that spans levels 3-5 of automation, providing a comprehensive functional safety architecture that leverages and efficiently uses computer vision and ADAS technologies for diversity and redundancy, along with deep learning tools to provide a platform for a flexible and reliable driving software stack. SoC 504 can be faster, more reliable, and even more energy- and space-efficient than conventional systems. For example, when combined with CPU 506, GPU 508, and data storage 516, accelerator 514 can provide a fast and efficient platform for Level 3-5 autonomous vehicles.

[0114] This technology therefore provides capabilities and functionality 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 wide variety of processing algorithms across a wide variety of visual data. However, CPUs often fail to meet the performance requirements of many computer vision applications, such as those related to execution time and power consumption. In particular, many CPUs are unable to execute complex object detection algorithms in real time, a requirement for in-vehicle ADAS applications and practical Level 3-5 autonomous vehicles.

[0115] In contrast to conventional systems, by providing a CPU complex, a GPU complex, and a hardware acceleration cluster, the technology described herein allows multiple neural networks to be executed simultaneously and / or sequentially, and the results to be combined to achieve Level 3-5 autonomous driving capabilities. For example, a CNN executed on a DLA or dGPU (e.g., GPU 520) can include text and word recognition, allowing the supercomputer to read and understand traffic signs, including signs for which a neural network has not been specifically trained. The DLA can also include a neural network that can recognize, interpret, and provide semantic understanding of the signs, and pass that semantic understanding to a path planning module running on the CPU complex.

[0116] As another example, as required for Level 3, 4, or 5 driving, multiple neural networks can be run simultaneously. For example, a warning sign consisting of "Caution: Flashing lights indicate icing conditions" along with a light can be interpreted by several neural networks, either independently or collectively. The sign itself can be identified as a traffic sign by a first neural network deployed (e.g., a trained neural network), and the text "Flashing lights indicate icing conditions" can be interpreted by a second neural network deployed, which informs the vehicle's path planning software (preferably executing on a CPU complex) that icing conditions exist when the flashing lights are detected. The flashing lights can be identified by operating a third neural network deployed over multiple frames, which informs the vehicle's path planning software of the presence (or absence) of the flashing lights. All three neural networks can run simultaneously, for example, within the DLA and / or on GPU 508.

[0117] In some examples, a CNN for facial recognition and owner recognition can use data from a camera sensor to identify the presence of an authorized driver and / or owner of the vehicle 500. The always-on sensor processing engine can be used to unlock the vehicle and turn on the lights when the owner approaches the driver's door, and in security mode, disable the vehicle when the owner leaves the vehicle. In this way, the SoC 504 provides security against theft and / or carjacking.

[0118] In another example, a CNN for emergency response vehicle detection and identification can use data from microphone 596 to detect and identify emergency response vehicle alerts. In contrast to conventional systems that use general classifiers to detect alerts and manually extract features, SoC504 uses a CNN to classify environmental and urban sounds as well as visual data. In a preferred embodiment, the CNN running on the DLA is trained to identify the relative closing speed of emergency response vehicles (for example, by using the Doppler effect). The CNN can also be trained to identify emergency response vehicles specific to the local area in which the vehicle is operating, as identified by the GNSS sensor 558. Thus, for example, when operating in Europe, the CNN will seek to detect European alerts, and when in the United States, the CNN will seek to identify only North American alerts. Once an emergency response vehicle is detected, with the assistance of the ultrasonic sensor 562, the control program can be used to execute emergency response vehicle safety routines, slowing the vehicle, pulling to the side of the road, stopping the vehicle, and / or idling the vehicle until the emergency response vehicle passes.

[0119] The vehicle may include a CPU 518 (e.g., a discrete CPU or dCPU) that may be coupled to the SoC 504 via a high-speed interconnect (e.g., PCIe). The CPU 518 may include, for example, an X86 processor. The CPU 518 may be used to perform any of a variety of functions, including, for example, arbitrating potentially inconsistent results between ADAS sensors and the SoC 504, and / or monitoring the status and health of the controller 536 and / or the infotainment SoC 530.

[0120] The vehicle 500 may include a GPU 520 (e.g., a discrete GPU or dGPU) that may be coupled to the SoC 504 via a high-speed interconnect (e.g., NVIDIA's NVLINK). The GPU 520 may provide additional artificial intelligence functionality, for example, by executing redundant and / or different neural networks, and may be used to train and / or update the neural network based at least in part on input from sensors of the vehicle 500 (e.g., sensor data).

[0121] The vehicle 500 may also include a network interface 524, which may include one or more wireless antennas 526 (e.g., one or more wireless antennas for different communication protocols, such as a cellular antenna, a Bluetooth antenna, etc.). The network interface 524 can be used to enable wireless connections to the cloud (e.g., to a server 578 and / or other network devices), to other vehicles, and / or to computing devices (e.g., a passenger's client device) via the Internet. In order to communicate with other vehicles, a direct link can be established between the two vehicles, and / or an indirect link can be established (e.g., across a network and through the Internet). The direct link can be provided using a vehicle-to-vehicle communication link. The vehicle-to-vehicle communication link can provide the vehicle 500 with information about vehicles approaching the vehicle 500 (e.g., vehicles in front of, to the side of, and / or behind the vehicle 500). This functionality can be part of the cooperative adaptive cruise control functionality of the vehicle 500.

[0122] The network interface 524 may include a SoC that provides modulation and demodulation functions and enables the controller 536 to communicate over a wireless network. The network interface 524 may include an RF front-end for up-conversion from baseband to radio frequency and down-conversion from radio frequency to baseband. The frequency conversion may be performed by a well-known process and / or may be performed using a super-heterodyne process. In some examples, the RF front-end function may be provided by a separate chip. The network interface may include wireless functions for communicating over LTE, WCDMA, UMTS, GSM, CDMA2000, Bluetooth, Bluetooth LE, Wi-Fi, Z-wave, ZigBee, LoRaWAN, and / or other wireless protocols.

[0123] The vehicle 500 may also include a data storage area 528, which may include off-chip storage (e.g., outside the SoC 504). The data storage area 528 may include one or more storage elements, including RAM, SRAM, DRAM, VRAM, flash memory, a hard disk, and / or other components and / or devices that can store at least one bit of data.

[0124] The vehicle 500 may also include a GNSS sensor 558. The GNSS sensor 558 (e.g., GPS, assisted GPS sensor, differential GPS (DGPS) sensor, etc.) is used to assist with mapping, perception, occupancy grid generation, and / or path planning functions. Any number of GNSS sensors 558 may be used, including, for example and without limitation, GPS using a USB connector with an Ethernet to serial (RS-232) bridge.

[0125] The vehicle 500 may also include a RADAR sensor 560. The RADAR sensor 560 can be used by the vehicle 500 for remote vehicle detection even in darkness and / or in adverse weather conditions. The RADAR functional safety level can be ASIL B. The RADAR sensor 560 can use CAN and / or bus 502 (e.g., to transmit data generated by the RADAR sensor 560) for control and access to object tracking data, and in some examples access Ethernet to access raw data. A variety of RADAR sensor types can be used. For example and without limitation, the RADAR sensor 560 can be suitable for front, rear, and side RADAR use. In some examples, a pulsed Doppler RADAR sensor is used.

[0126] The RADAR sensor 560 can include different configurations, such as long-range with a narrow field of view, short-range with a wide field of view, short-range side coverage, and so on. In some examples, long-range RADAR can be used for adaptive cruise control functions. The long-range RADAR system can provide a wide field of view (e.g., within 250m) achieved by two or more independent scans. The RADAR sensor 560 can help distinguish between static objects and moving objects and can be used by the ADAS system for emergency braking assistance and forward collision warning. The long-range RADAR sensor may include a single-station multimode RADAR with multiple (e.g., six or more) fixed RADAR antennas and high-speed CAN, Ethernet and / or FlexRay interfaces. In the example with six antennas, the central four antennas can create a focused beam pattern that is designed to record the surroundings of the vehicle 500 at a higher rate with minimal traffic interference from adjacent lanes. The other two antennas can expand the field of view, making it possible to quickly detect vehicles entering or leaving the lane of the vehicle 500.

[0127] As an example, a medium-range RADAR system may include a range of up to 560m (front) or 80m (rear) and a field of view of up to 42 degrees (front) or 550 degrees (rear). A short-range RADAR system may include, but is not limited to, a RADAR sensor designed to be mounted on both ends of the rear bumper. When mounted on both ends of the rear bumper, such a RADAR sensor system can create two beams that continuously monitor the blind spots behind and beside the vehicle.

[0128] Short-range RADAR systems can be used in ADAS systems for blind spot detection and / or lane change assistance.

[0129] Vehicle 500 may also include ultrasonic sensors 562. Ultrasonic sensors 562, which may be located on the front, rear, and / or sides of vehicle 500, may be used for parking assistance and / or for creating and updating an occupancy grid. A variety of ultrasonic sensors 562 may be used, and different ultrasonic sensors 562 may have different detection ranges (e.g., 2.5 m, 4 m). Ultrasonic sensors 562 may operate at functional safety level ASIL B.

[0130] Vehicle 500 may include a LIDAR sensor 564. LIDAR sensor 564 may be used for object and pedestrian detection, emergency braking, collision avoidance, and / or other functions. LIDAR sensor 564 may be ASIL B functional safety level. In some examples, vehicle 500 may include multiple LIDAR sensors 564 (e.g., two, four, six, etc.) that may use Ethernet (e.g., to provide data to a Gigabit Ethernet switch).

[0131] In some examples, the LIDAR sensor 564 may be capable of providing a list of objects and their distances for a 360-degree field of view. Commercially available LIDAR sensors 564 may have, for example, an advertised range of approximately 500 meters, an accuracy of 2-3 cm, and support for 500 Mbps Ethernet connections. In some examples, one or more non-obtrusive LIDAR sensors 564 may be used. In such examples, the LIDAR sensor 564 may be implemented as a small device that can be embedded in the front, back, sides, and / or corners of the vehicle 500. In such an example, the LIDAR sensor 564 may provide a field of view of up to 120 degrees horizontally and 35 degrees vertically, even for low-reflectivity objects, with a range of 200 meters. The front-mounted LIDAR sensor 564 may be configured for a horizontal field of view between 45 and 135 degrees.

[0132] In some examples, LIDAR technologies such as 3D flash LIDAR may also be used. 3D flash LIDAR uses flashes of laser as an emission source to illuminate the vehicle's surroundings up to about 200 m. The flash LIDAR unit includes a receiver that records the laser pulse transmission time and reflected light on each pixel, which in turn corresponds to the range from the vehicle to the object. Flash LIDAR can allow a highly accurate and distortion-free image of the surrounding environment to be generated with each laser flash. In some examples, four flash LIDAR sensors can be deployed, one on each side of the vehicle 500. Available 3D flash LIDAR systems include solid-state 3D staring array LIDAR cameras (e.g., non-scanning LIDAR devices) with no moving parts other than a fan. The flash LIDAR device can use 5 nanosecond Class I (eye-safe) laser pulses per frame and can capture the reflected laser light in the form of a 3D range point cloud and co-registered intensity data. By using flash LIDAR, and because flash LIDAR is a solid-state device with no moving parts, the LIDAR sensor 564 may be less susceptible to motion blur, vibration, and / or shock.

[0133] The vehicle may also include an IMU sensor 566. In some examples, the IMU sensor 566 may be located at the center of the rear axle of the vehicle 500. The IMU sensor 566 may include, for example and without limitation, an accelerometer, a magnetometer, a gyroscope, a magnetic compass, and / or other sensor types. In some examples, such as in a six-axis application, the IMU sensor 566 may include an accelerometer and a gyroscope, while in a nine-axis application, the IMU sensor 566 may include an accelerometer, a gyroscope, and a magnetometer.

[0134] In some embodiments, the IMU sensor 566 can be implemented as a miniature, high-performance GPS-assisted inertial navigation system (GPS / INS) that combines micro-electromechanical systems (MEMS) inertial sensors, a high-sensitivity GPS receiver, and advanced Kalman filtering algorithms to provide estimates of position, velocity, and attitude. Thus, in some examples, the IMU sensor 566 can enable the vehicle 500 to estimate heading without input from a magnetic sensor by directly observing and correlating velocity changes from the GPS to the IMU sensor 566. In some examples, the IMU sensor 566 and the GNSS sensor 558 can be combined into a single integrated unit.

[0135] The vehicle may include microphones 596 positioned in and / or around the vehicle 500. The microphones 596 may be used for, among other things, emergency response vehicle detection and identification.

[0136] The vehicle may also include any number of camera types, including stereo cameras 568, wide angle cameras 570, infrared cameras 572, surround cameras 574, long and / or medium range cameras 598, and / or other camera types. These cameras may be used to capture image data around the entire periphery of the vehicle 500. The type of camera used depends on the embodiment and the requirements of the vehicle 500, and any combination of camera types may be used to provide the necessary coverage around the vehicle 500. Additionally, the number of cameras may vary depending on the embodiment. For example, the vehicle may include six cameras, seven cameras, ten cameras, twelve cameras, and / or another number of cameras. As an example and not limitation, the cameras may support Gigabit Multimedia Serial Link (GMSL) and / or Gigabit Ethernet. Each of the cameras described herein may include a GMSL and / or Gigabit Ethernet network. Figure 5A and Figure 5B Described in more detail.

[0137] Vehicle 500 may also include a vibration sensor 542. Vibration sensor 542 can measure the vibration of a component of the vehicle, such as an axle. For example, changes in vibration can indicate changes in the road surface. In another example, when two or more vibration sensors 542 are used, the difference between the vibrations can be used to determine friction or slip on the road surface (e.g., when there is a vibration difference between a powered axle and a freely rotating axle).

[0138] The vehicle 500 may include an ADAS system 538. In some examples, the ADAS system 538 may include a SoC. The ADAS system 538 may include autonomous / adaptive / automatic cruise control (ACC), cooperative adaptive cruise control (CACC), forward collision warning (FCW), automatic emergency braking (AEB), lane departure warning (LDW), lane keeping assist (LKA), blind spot warning (BSW), rear cross traffic warning (RCTW), collision warning system (CWS), lane centering (LC), and / or other features and functions.

[0139] The ACC system may utilize RADAR sensor 560, LIDAR sensor 564, and / or a camera. The ACC system may include longitudinal ACC and / or lateral ACC. Longitudinal ACC monitors and controls the distance to the vehicle immediately in front of vehicle 500, automatically adjusting the vehicle speed to maintain a safe distance from the vehicle in front. Lateral ACC maintains distance and, when necessary, recommends lane changes for vehicle 500. Lateral ACC is related to other ADAS applications such as LCA and CWS.

[0140] CACC uses information from other vehicles, which can be received from other vehicles indirectly via a wireless link or through a network connection (e.g., through the Internet) via the network interface 524 and / or wireless antenna 526. A direct link can be provided by a vehicle-to-vehicle (V2V) communication link, while an indirect link can be an infrastructure-to-vehicle (I2V) communication link. Typically, the V2V communication concept provides information about the vehicle immediately ahead (e.g., the vehicle immediately ahead of the vehicle 500 and in the same lane as it), while the I2V communication concept provides information about traffic further ahead. A CACC system can include either or both I2V and V2V information sources. Given information about the vehicle ahead of the vehicle 500, CACC can be more reliable, and it has the potential to improve the smoothness of traffic flow and reduce road congestion.

[0141] The FCW system is designed to alert the driver to hazards so that the driver can take corrective action. The FCW system uses a front-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC, which is electrically coupled to driver feedback such as a display, speaker, and / or vibrating components. The FCW system can provide warnings in the form of, for example, audible, visual warnings, vibrations, and / or rapid brake pulses.

[0142] The AEB system detects an impending forward collision with another vehicle or other object and can automatically apply the brakes if the driver does not take corrective action within specified time or distance parameters. The AEB system may use a front-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA and / or ASIC. When the AEB system detects a hazard, it typically first alerts the driver to take corrective action to avoid the collision, and if the driver does not take corrective action, the AEB system may automatically apply the brakes in an effort to prevent or at least mitigate the effects of the predicted collision. The AEB system may include technologies such as dynamic brake support and / or collision approach braking.

[0143] The LDW system provides visual, audible, and / or tactile warnings, such as steering wheel or seat vibrations, to alert the driver when the vehicle 500 crosses a lane marking. When the driver indicates an intention to leave the lane by activating a turn signal, the LDW system is deactivated. The LDW system may utilize a front-facing camera coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration components.

[0144] The LKA system is a variation of the LDW system. If the vehicle 500 begins to leave the lane, the LKA system provides steering input or braking to correct the vehicle 500.

[0145] The BSW system detects and warns the driver of vehicles in the car's blind spot. The BSW system can provide visual, audible, and / or tactile alerts to indicate that merging or changing lanes is unsafe. The system can provide additional warnings when the driver uses a turn signal. The BSW system can use a rear-facing camera and / or RADAR sensor 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration component.

[0146] The RCTW system can provide visual, audible, and / or tactile notifications when an object is detected outside the range of the rear-mounted camera while the vehicle 500 is in reverse. Some RCTW systems include AEB to ensure that the vehicle brakes are applied to avoid a collision. The RCTW system can use one or more rear-mounted RADAR sensors 560 coupled to a dedicated processor, DSP, FPGA, and / or ASIC that is electrically coupled to driver feedback such as a display, speaker, and / or vibration component.

[0147] Conventional ADAS systems may be prone to false positive results, which may be annoying and distracting to the driver, but are typically not catastrophic because the ADAS system alerts the driver and allows the driver to decide whether a safety condition actually exists and take action accordingly. However, in the autonomous vehicle 500, in the event of conflicting results, the vehicle 500 itself must decide whether to pay attention to the results from the main computer or the auxiliary computer (e.g., the first controller 536 or the second controller 536). For example, in some embodiments, the ADAS system 538 can be a backup and / or auxiliary computer for providing perception information to the backup computer rationality module. The backup computer rationality monitor can run redundant and diverse software on hardware components to detect failures in perception and dynamic driving tasks. The output from the ADAS system 538 can be provided to the supervisory MCU. If the outputs from the main computer and the auxiliary computer conflict, the supervisory MCU must determine how to reconcile the conflict to ensure safe operation.

[0148] In some examples, the primary computer can be configured to provide a confidence score to the supervisory MCU, indicating the primary computer's confidence in the selected result. If the confidence score exceeds a threshold, the supervisory MCU can follow the primary computer's direction, regardless of whether the secondary computer provides conflicting or inconsistent results. In the event that the confidence score does not meet the threshold and the primary and secondary computers indicate different results (e.g., a conflict), the supervisory MCU can arbitrate between these computers to determine the appropriate result.

[0149] The supervisory MCU can be configured to run a neural network that is trained and configured to determine conditions under which the secondary computer provides a false alarm based, at least in part, on outputs from the primary computer and the secondary computer. Thus, the neural network in the supervisory MCU can learn when the output of the secondary computer can be trusted and when it cannot. For example, when the secondary computer is a RADAR-based FCW system, the neural network in the supervisory MCU can learn when the FCW system is identifying a metal object that is not actually a danger, such as a drain grate or manhole cover, which triggers an alarm. Similarly, when the secondary computer is a camera-based LDW system, the neural network in the supervisory MCU can learn to disregard the LDW when a cyclist or pedestrian is present and lane departure is actually the safest strategy. In embodiments that include a neural network running on the supervisory MCU, the supervisory MCU can include at least one of a DLA or a GPU suitable for running the neural network with associated memory. In preferred embodiments, the supervisory MCU can include and / or be included as a component of the SoC 504.

[0150] In other examples, the ADAS system 538 may include an auxiliary computer that uses traditional computer vision rules to perform ADAS functions. In this way, the auxiliary computer can use classic computer vision rules (if-then), and the presence of a neural network in the supervisory MCU can improve reliability, safety, and performance. For example, diverse implementations and intentional non-identity make the entire system more fault-tolerant, especially with respect to failures caused by software (or software-hardware interface) functions. For example, if there is a software vulnerability or bug in the software running on the main computer and the non-identical software code running on the auxiliary computer provides the same overall result, the supervisory MCU can be more confident that the overall result is correct and that the vulnerability in the software or hardware on the main computer did not cause a substantial error.

[0151] In some examples, the output of ADAS system 538 can be fed into the primary computer's perception block and / or the primary computer's dynamic driving task block. For example, if ADAS system 538 indicates a forward collision warning due to an object immediately ahead, the perception block can use this information when identifying the object. In other examples, the secondary computer can have its own neural network that is trained and thus reduces the risk of false positives as described herein.

[0152] The vehicle 500 may also include an infotainment SoC 530 (e.g., an in-vehicle infotainment system (IVI)). Although illustrated and described as an SoC, the infotainment system may not be an SoC and may include two or more separate components. The infotainment SoC 530 may include a combination of hardware and software that can be used to provide audio (e.g., music, personal digital assistant, navigation instructions, news, radio, etc.), video (e.g., TV, movies, streaming, etc.), phone (e.g., hands-free calling), network connectivity (e.g., LTE, WiFi, etc.), and / or information services (e.g., a navigation system, rear parking assistance, radio data system, vehicle-related information such as fuel level, total distance covered, brake fuel level, oil level, door open / closed, air filter information, etc.) to the vehicle 500. For example, the infotainment SoC 530 may include a radio, a disc player, a navigation system, a video player, USB and Bluetooth connectivity, an onboard computer, in-vehicle entertainment, WiFi, steering wheel audio controls, hands-free voice controls, a head-up display (HUD), an HMI display 534, a telematics device, a control panel (e.g., for controlling and / or interacting with various components, features, and / or systems), and / or other components. The infotainment SoC 530 may further be used to provide information (e.g., visual and / or auditory) to a user of the vehicle, such as information from an ADAS system 538, autonomous driving information such as planned vehicle maneuvers, trajectories, surrounding environment information (e.g., intersection information, vehicle information, road information, etc.), and / or other information.

[0153] The infotainment SoC 530 may include GPU functionality. The infotainment SoC 530 may communicate with other devices, systems, and / or components of the vehicle 500 via a bus 502 (e.g., a CAN bus, Ethernet, etc.). In some examples, the infotainment SoC 530 may be coupled to a supervisory MCU so that in the event of a failure of a primary controller 536 (e.g., a primary and / or backup computer of the vehicle 500), the infotainment system's GPU may perform some self-driving functions. In such an example, the infotainment SoC 530 may place the vehicle 500 in a driver-safe parking mode as described herein.

[0154] The vehicle 500 may also include an instrument cluster 532 (e.g., a digital instrument panel, an electronic instrument cluster, a digital instrument panel, etc.). The instrument cluster 532 may include a controller and / or a supercomputer (e.g., a separate controller or a supercomputer). The instrument cluster 532 may include a set of instruments, such as a speedometer, fuel level, oil pressure, a tachometer, an odometer, a turn indicator, a shift position indicator, a seat belt warning light, a parking brake warning light, an engine check light, airbag (SRS) system information, lighting controls, safety system controls, navigation information, etc. In some examples, information may be displayed and / or shared between the infotainment SoC 530 and the instrument cluster 532. In other words, the instrument cluster 532 may be included as part of the infotainment SoC 530, or vice versa.

[0155] Figure 5D For cloud-based servers and Figure 5A 5. System diagram of communication between an example autonomous vehicle 500. System 576 may include a server 578, a network 590, and a vehicle including vehicle 500. Server 578 may include multiple GPUs 584(A)-584(H) (collectively referred to herein as GPUs 584), PCIe switches 582(A)-582(H) (collectively referred to herein as PCIe switches 582), and / or CPUs 580(A)-580(B) (collectively referred to herein as CPUs 580). GPUs 584, CPUs 580, and PCIe switches may be interconnected with a high-speed interconnect and / or PCIe connection 586, such as, for example and without limitation, an NVLink interface 588 developed by NVIDIA. In some examples, GPUs 584 are connected via NVLink and / or NVSwitch SoCs, and GPUs 584 and PCIe switches 582 are connected via a PCIe interconnect. Although eight GPUs 584, two CPUs 580, and two PCIe switches are illustrated, this is not intended to be limiting. Depending on the embodiment, each of the servers 578 can include any number of GPUs 584, CPUs 580, and / or PCIe switches. For example, each of the servers 578 can include eight, sixteen, thirty-two, and / or more GPUs 584.

[0156] Server 578 can receive image data from a vehicle via network 590, the image data representing images showing unexpected or changed road conditions, such as recently begun road construction. Server 578 can transmit neural network 592, updated neural network 592, and / or map information 594, including information about traffic and road conditions, via network 590 and to the vehicle. Updates to map information 594 can include updates to HD map 522, such as information about construction sites, potholes, curves, flooding, or other obstacles. In some examples, neural network 592, updated neural network 592, and / or map information 594 can be generated from new training and / or data received from any number of vehicles in the environment and / or based on experience with training performed at a data center (e.g., using server 578 and / or other servers).

[0157] Server 578 can be used to train a machine learning model (e.g., a neural network) based on training data. The training data can be generated by the vehicle and / or can be generated in simulation (e.g., using a game engine). In some examples, the training data is labeled (e.g., in cases where the neural network benefits from supervised learning) and / or undergoes other preprocessing, while in other examples, the training data is not labeled and / or preprocessed (e.g., in cases where the neural network does not require supervised learning). Training can be performed according to any one or more categories of machine learning techniques, including but not limited to categories such as: supervised training, semi-supervised training, unsupervised training, self-learning, reinforcement learning, federated learning, transfer learning, feature learning (including principal component and cluster analysis), multilinear subspace learning, manifold learning, representation learning (including alternative dictionary learning), rule-based machine learning, anomaly detection, and any variants or combinations thereof. Once the machine learning model is trained, the machine learning model can be used by the vehicle (e.g., transmitted to the vehicle via network 590), and / or the machine learning model can be used by server 578 to remotely monitor the vehicle.

[0158] In some examples, server 578 can receive data from the vehicle and apply the data to the latest real-time neural network for real-time intelligent reasoning. Server 578 can include a deep learning supercomputer and / or a dedicated AI computer powered by GPU 584, such as the DGX and DGX Station machines developed by NVIDIA. However, in some examples, server 578 can include the deep learning infrastructure of a data center using only CPU power.

[0159] The deep learning infrastructure of server 578 may be capable of rapid real-time inference and may use this capability to assess and verify the health of the processors, software, and / or associated hardware in vehicle 500. For example, the deep learning infrastructure may receive periodic updates from vehicle 500, such as an image sequence and / or objects that vehicle 500 has located in the image sequence (e.g., via computer vision and / or other machine learning object classification techniques). The deep learning infrastructure may run its own neural network to identify objects and compare them to the objects identified by vehicle 500, and if the results do not match and the infrastructure concludes that the AI ​​in vehicle 500 has malfunctioned, server 578 may transmit a signal to vehicle 500 instructing the vehicle's 500 fail-safe computer to take control, notify passengers, and complete a safe parking maneuver.

[0160] For inference, the server 578 may include a GPU 584 and one or more programmable inference accelerators (e.g., NVIDIA's TensorRT 3). The combination of GPU-powered servers and inference acceleration can enable real-time responses. In other examples, such as where performance is less important, CPU, FPGA, and other processor-powered servers can be used for inference.

[0161] Example computing device

[0162] Figure 6 FIG6 is a block diagram of an example computing device 600 suitable for implementing some embodiments of the present disclosure. Computing device 600 may include an interconnect system 602 that directly or indirectly couples the following devices: memory 604, one or more central processing units (CPUs) 606, one or more graphics processing units (GPUs) 608, a communication interface 610, input / output (I / O) ports 612, an I / O component 614, a power supply 616, one or more presentation components 618 (e.g., a display), and one or more logic units 620. In at least one embodiment, computing device 600 may include one or more virtual machines (VMs), and / or any of its components may include virtual components (e.g., virtual hardware components). By way of non-limiting example, one or more of GPUs 608 may include one or more vGPUs, one or more of CPUs 606 may include one or more vCPUs, and / or one or more of logic units 620 may include one or more virtual logic units. As such, computing device 600 may include discrete components (eg, a complete GPU dedicated to computing device 600 ), virtual components (eg, portions of a GPU dedicated to computing device 600 ), or a combination thereof.

[0163] although Figure 6The various blocks of are shown as being connected via an interconnect system 602 having wires, but this is not intended to be limiting and is provided for clarity only. For example, in some embodiments, a presentation component 618 such as a display device may be considered an I / O component 614 (e.g., if the display is a touch screen). As another example, the CPU 606 and / or the GPU 608 may include memory (e.g., the memory 604 may represent a storage device in addition to the memory of the GPU 608, the CPU 606, and / or the other components). In other words, Figure 6 The term computing device is illustrative only. No distinction is made between categories such as "workstation," "server," "laptop," "desktop," "tablet," "client device," "mobile device," "handheld device," "game console," "electronic control unit (ECU)," "virtual reality system," and / or other device or system types, as all are considered within the Figure 6 within the range of computing devices.

[0164] Interconnect system 602 can represent one or more links or buses, such as an address bus, a data bus, a control bus, or a combination thereof. Interconnect system 602 can include one or more bus or link types, such as an industry standard architecture (ISA) bus, an extended industry standard architecture (EISA) bus, a video electronics standard association (VESA) bus, a peripheral component interconnect (PCI) bus, a peripheral component interconnect express (PCIe) bus, and / or another type of bus or link. In some embodiments, there is a direct connection between components. For example, CPU 606 can be directly connected to memory 604. Further, CPU 606 can be directly connected to GPU 608. In the case where there is a direct or point-to-point connection between components, interconnect system 602 can include a PCIe link that implements the connection. In these examples, the PCI bus need not be included in computing device 600.

[0165] Memory 604 may include any of a variety of computer-readable media. Computer-readable media can be any available media that can be accessed by computing device 600. Computer-readable media can include volatile and non-volatile media and removable and non-removable media. By way of example and not limitation, computer-readable media can include computer storage media and communication media.

[0166] Computer storage media may include volatile and non-volatile media and / or removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, and / or other data types. For example, memory 604 may store computer-readable instructions (e.g., representing programs and / or program elements, such as an operating system). Computer storage media may include, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage devices, magnetic cassettes, magnetic tape, magnetic disk storage devices 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 600. As used herein, computer storage media does not include signals themselves.

[0167] 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 transmission mechanism, and include any information delivery media. The term "modulated data signal" may refer to a signal that has one or more of its characteristics set or changed in such a manner as to encode information into the signal. By way of example and not limitation, computer storage media may 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. Any combination of the above should also be included within the scope of computer-readable media.

[0168] The CPU 606 can be configured to execute at least some of the computer-readable instructions in order to control one or more components of the computing device 600 to perform one or more of the methods and / or processes described herein. Each of the CPUs 606 can include one or more cores (e.g., one, two, four, eight, twenty-eight, seventy-two, etc.) capable of processing a large number of software threads simultaneously. The CPU 606 can include any type of processor and can include different types of processors, depending on the type of computing device 600 implemented (e.g., a processor with fewer cores for mobile devices and a processor with more cores for servers). For example, depending on the type of computing device 600, the processor can be an Advanced RISC Machine (ARM) processor implemented using Reduced Instruction Set Computing (RISC) or an x86 processor implemented using Complex Instruction Set Computing (CISC). The computing device 600 can also include one or more CPUs 606 in addition to one or more microprocessors or supplementary coprocessors such as math coprocessors.

[0169] In addition to or in lieu of the CPU 606, the GPU 608 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 600 to perform one or more of the methods and / or processes described herein. One or more of the GPUs 608 may be integrated GPUs (e.g., with one or more of the CPUs 606) and / or one or more of the GPUs 608 may be discrete GPUs. In embodiments, one or more of the GPUs 608 may be coprocessors for one or more of the CPUs 606. The GPU 608 may be used by the computing device 600 to render graphics (e.g., 3D graphics) or perform general-purpose computations. For example, the GPU 608 may be used for general-purpose computing on a GPU (GPGPU). The GPU 608 may include hundreds or thousands of cores capable of processing hundreds or thousands of software threads simultaneously. The GPU 608 may generate pixel data for outputting an image in response to rendering commands (e.g., rendering commands received from the CPU 606 via a host interface). The GPU 608 may include graphics memory, such as display memory, for storing pixel data or any other suitable data, such as GPGPU data. Display memory may be included as part of memory 604. GPU 608 may include two or more GPUs operating in parallel (e.g., via a link). The link may connect the GPUs directly (e.g., using NVLINK) or may connect the GPUs through a switch (e.g., using NVSwitch). When combined, each GPU 608 may generate pixel data or GPGPU data for different portions of output or for different outputs (e.g., a first GPU for a first image and a second GPU for a second image). Each GPU may include its own memory or may share memory with other GPUs.

[0170] In addition to or in lieu of the CPU 606 and / or GPU 608, the logic unit 620 may be configured to execute at least some of the computer-readable instructions to control one or more components of the computing device 600 to perform one or more of the methods and / or processes described herein. In embodiments, the CPU 606, GPU 608, and / or the logic unit 620 may independently or jointly perform any combination of the methods, processes, and / or portions thereof. One or more of the logic units 620 may be part of and / or integrated into one or more of the CPU 606 and / or GPU 608, and / or one or more of the logic units 620 may be discrete components or otherwise external to the CPU 606 and / or GPU 608. In embodiments, one or more of the logic units 620 may be coprocessors to one or more of the CPU 606 and / or GPU 608.

[0171] Examples of logic unit 620 include one or more processing cores and / or components thereof, such as a tensor core (TC), a tensor processing unit (TPU), a pixel visual core (PVC), a visual processing unit (VPU), a graphics processing cluster (GPC), a texture processing cluster (TPC), a streaming multiprocessor (SM), a tree traversal unit (TTU), an artificial intelligence accelerator (AIA), a deep learning accelerator (DLA), an arithmetic logic unit (ALU), an application specific integrated circuit (ASIC), a floating point unit (FPU), an input / output (I / O) element, a peripheral component interconnect (PCI) or a peripheral component interconnect express (PCIe) element, and / or the like.

[0172] The communication interface 610 may include one or more receivers, transmitters, and / or transceivers that enable the computing device 600 to communicate with other computing devices via an electronic communication network, including wired and / or wireless communications. The communication interface 610 may include components and functionality that enable communication over any of a number of different networks, such as wireless networks (e.g., Wi-Fi, Z-Wave, Bluetooth, Bluetooth LE, ZigBee, etc.), wired networks (e.g., via Ethernet or InfiniBand communications), low-power wide-area networks (e.g., LoRaWAN, SigFox, etc.), and / or the Internet.

[0173] The I / O ports 612 can enable the computing device 600 to be logically coupled to other devices including an I / O component 614, a presentation component 618, and / or other components, some of which can be built into (e.g., integrated into) the computing device 600. Illustrative I / O components 614 include a microphone, a mouse, a keyboard, a joystick, a game pad, a game controller, a satellite dish, a scanner, a printer, a wireless device, and the like. The I / O component 614 can provide a natural user interface (NUI) that processes user-generated air gestures, voice, or other physiological input. In some examples, the input can be transmitted to an appropriate network element for further processing. The NUI can implement any combination of voice recognition, stylus recognition, facial recognition, biometric recognition, gesture recognition on and adjacent to the screen, air gestures, head and eye tracking, and touch recognition associated with the display of the computing device 600 (as described in more detail below). The computing device 600 can include a depth camera such as a stereo camera system, an infrared camera system, an RGB camera system, touch screen technology, and combinations thereof for gesture detection and recognition. Additionally, computing device 600 may include an accelerometer or gyroscope to enable motion detection (e.g., as part of an inertial measurement unit (IMU)). In some examples, the output of the accelerometer or gyroscope may be used by computing device 600 to render immersive augmented or virtual reality.

[0174] The power supply 616 may include a hardwired power supply, a battery power supply, or a combination thereof. The power supply 616 may provide power to the computing device 600 to enable the components of the computing device 600 to operate.

[0175] The presentation component 618 may include a display (e.g., a monitor, a touch screen, a television screen, a head-up display (HUD), other display types, or a combination thereof), speakers, and / or other presentation components. The presentation component 618 may receive data from other components (e.g., the GPU 608, the CPU 606, etc.) and output the data (e.g., as images, video, sound, etc.).

[0176] Sample Data Center

[0177] Figure 7 An example data center 700 is shown that can be used in at least one embodiment of the present disclosure. The data center 700 can include a data center infrastructure layer 710, a framework layer 720, a software layer 730, and / or an application layer 740.

[0178] like Figure 7As shown, the data center infrastructure layer 710 may include a resource coordinator 712, grouped computing resources 714, and node computing resources ("node CRs") 716(1)-716(N), where "N" represents any complete positive integer. In at least one embodiment, the node CRs 716(1)-716(N) may include, but are not limited to, any number of central processing units ("CPUs") or other processors (including accelerators, field programmable gate arrays (FPGAs), graphics processors or graphics processing units (GPUs), etc.), memory devices (e.g., dynamic read-only memory), storage devices (e.g., solid-state drives or disk drives), network input / output ("NW I / O") devices, network switches, virtual machines ("VMs"), power modules and / or cooling modules, etc. In some embodiments, one or more of the node CRs 716(1)-716(N) may correspond to a server having one or more of the above-mentioned computing resources. Furthermore, in some embodiments, nodes CR 716 ( 1 )- 716 (N) may include one or more virtual components, such as vGPUs, vCPUs, etc., and / or one or more of nodes CR 716 ( 1 )- 716 (N) may correspond to virtual machines (VMs).

[0179] In at least one embodiment, the computing resources 714 of grouping can include a separate grouping (not shown) of the node CR 716 housed in one or more racks, or many racks (also not shown) housed in the data center of each geographical location. The separate grouping of the node CR in the computing resources 714 of grouping can include computing, network, memory or storage resources that can be configured or assigned to support the grouping of one or more workloads. In at least one embodiment, several node CRs 716 comprising CPU, GPU and / or other processors can be grouped in one or more racks to provide computing resources to support one or more workloads. One or more racks can also include any number of power modules, cooling modules and / or network switches, in any combination.

[0180] Resource coordinator 722 may configure or otherwise control one or more nodes CR 716(1)-716(N) and / or grouped computing resources 714. In at least one embodiment, resource coordinator 722 may comprise a software design infrastructure ("SDI") management entity for data center 700. Resource coordinator 722 may comprise hardware, software, or some combination thereof.

[0181] In at least one embodiment, Figure 7As shown, the framework layer 720 may include a job scheduler 732, a configuration manager 734, a resource manager 736, and / or a distributed file system 738. The framework layer 720 may include a framework that supports the software 732 of the software layer 730 and / or one or more applications 742 of the application layer 740. The software 732 or the application 742 may include web-based service software or applications, such as those provided by Amazon Web Services, Google Cloud, and Microsoft Azure. The framework layer 720 may be, but is not limited to, a free and open source software web application framework, such as Apache Spark, which may utilize the distributed file system 738 for large-scale data processing (e.g., "big data"). TM (hereinafter referred to as "Spark"). In at least one embodiment, the job scheduler 732 may include a Spark driver to facilitate scheduling of workloads supported by the various layers of the data center 700. The configuration manager 734 may be capable of configuring different layers, such as the software layer 730 and the framework layer 720 including Spark and a distributed file system 738 for supporting large-scale data processing. The resource manager 736 may be capable of managing the mapping or allocation of clustered or grouped computing resources used to support the distributed file system 738 and the job scheduler 732. In at least one embodiment, the clustered or grouped computing resources may include grouped computing resources 714 on the data center infrastructure layer 710. The resource manager 736 may coordinate with the resource coordinator 712 to manage these mapped or allocated computing resources.

[0182] In at least one embodiment, the software 732 included in the software layer 730 may include software used by at least a portion of the node CRs 716(1)-716(N), the grouped computing resources 714, and / or the distributed file system 738 of the framework layer 720. The one or more types of software may include, but are not limited to, Internet web page search software, email virus scanning software, database software, and streaming video content software.

[0183] In at least one embodiment, the one or more applications 742 included in the application layer 740 may include one or more types of applications used by at least a portion of the node CRs 716(1)-716(N), the grouped computing resources 714, and / or the distributed file system 738 of the framework layer 720. The one or more types of applications may include, but are not limited to, any number of genomics applications, cognitive computing, and machine learning applications, including training or inference software, machine learning framework software (e.g., PyTorch, TensorFlow, Caffe, etc.), and / or other machine learning applications used in conjunction with one or more embodiments.

[0184] In at least one embodiment, any of configuration manager 734, resource manager 736, and resource coordinator 712 can implement any number and type of self-modification actions based on any number and type of data obtained in any technically feasible manner. The self-modification actions can relieve the data center operator of data center 700 from making potentially poor configuration decisions and can avoid underutilized and / or poorly performing portions of the data center.

[0185] The data center 700 may include tools, services, software, or other resources to train one or more machine learning models or use one or more machine learning models to predict or infer information in accordance with one or more embodiments described herein. For example, a machine learning model may be trained by calculating weight parameters in accordance with a neural network architecture using the software and / or computing resources described above with respect to the data center 700. In at least one embodiment, information may be inferred or predicted using trained or deployed machine learning models corresponding to one or more neural networks using the resources described above with respect to the data center 700 using weight parameters calculated using one or more training techniques such as, but not limited to, those described herein.

[0186] In at least one embodiment, the data center 700 can use a CPU, an application-specific integrated circuit (ASIC), a GPU, an FPGA, and / or other hardware (or corresponding virtual computing resources) to use the above resources to perform training and / or reasoning. In addition, one or more of the above software and / or hardware resources can be configured as a service to allow users to train or perform information reasoning, such as image recognition, speech recognition, or other artificial intelligence services.

[0187] Sample network environment

[0188] A network environment suitable for implementing embodiments of the present disclosure may include one or more client devices, servers, network attached storage (NAS), other backend devices, and / or other device types. The client devices, servers, and / or other device types (e.g., each device) may be configured to: Figure 6 The data center 700 may be implemented on one or more instances of one or more computing devices 600 - for example, each device may include similar components, features and / or functions of one or more computing devices 600. In addition, in the case of implementing a backend device (e.g., a server, NAS, etc.), the backend device may be included as part of the data center 700, an example of which is described herein with respect to Figure 7 Describe in more detail.

[0189] The components of the network environment can communicate with each other via one or more networks, which can be wired, wireless, or both. The network can include multiple networks or a network among multiple networks. For example, the network can include one or more wide area networks (WANs), one or more local area networks (LANs), one or more public networks such as the Internet and / or a public switched telephone network (PSTN), and / or one or more private networks. In the case where the network includes a wireless telecommunications network, components such as base stations, communication towers, or even access points (and other components) can provide wireless connections.

[0190] Compatible network environments may include one or more peer-to-peer network environments (in which case the server may not be included in the network environment) and one or more client-server network environments (in which case one or more servers may be included in the network environment). In a peer-to-peer network environment, the functionality described herein with respect to the server may be implemented on any number of client devices.

[0191] In at least one embodiment, the network environment may include one or more cloud-based network environments, distributed computing environments, combinations thereof, and the like. The cloud-based network environment may include a framework layer, a job scheduler, a resource manager, and a distributed file system implemented on one or more servers, which may include one or more core network servers and / or edge servers. The framework layer may include software supporting the software layer and / or a framework for one or more applications at the application layer. The software or applications may include network-based service software or applications, respectively. In an embodiment, one or more client devices may use web-based service software or applications (e.g., by accessing the service software and / or applications via one or more application programming interfaces (APIs)). The framework layer may be, but is not limited to, a type of free and open source software web application framework, such as one that may use a distributed file system for large-scale data processing (e.g., "big data").

[0192] A cloud-based network environment can provide cloud computing and / or cloud storage that performs any combination of the computing and / or data storage functions described herein (or one or more portions thereof). Any of these different functions can be distributed across multiple locations from a central or core server (e.g., one or more data centers that can be distributed across a state, region, country, global, etc.). If the connection to the user (e.g., client device) is relatively close to an edge server, the core server can assign at least a portion of the functionality to the edge server. A cloud-based network environment can be private (e.g., limited to a single organization), public (e.g., available to many organizations), and / or a combination thereof (e.g., a hybrid cloud environment).

[0193] One or more client devices may include Figure 6 At least some of the components, features, and functionality of one or more example computing devices 600 are described. By way of example and not limitation, the client device may be embodied as a personal computer (PC), a laptop computer, 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 camera, a surveillance device or system, a vehicle, a boat, a spacecraft, a virtual machine, a drone, a robot, a handheld communication device, a hospital device, a gaming device or system, an entertainment system, a vehicle computer system, an embedded system controller, a remote control, an appliance, a consumer electronic device, a workstation, an edge device, any combination of these depicted devices, or any other suitable device.

[0194] The present disclosure can be described in the general context of machine-usable instructions or computer code executed by a computer or other machine such as a personal digital assistant or other handheld device, including computer-executable instructions such as program modules. Generally, program modules including routines, programs, objects, components, data structures, etc. refer to code that performs a specific task or implements a specific abstract data type. The present disclosure can be practiced in a variety of system configurations, including handheld devices, consumer electronics, general-purpose computers, more specialized computing devices, etc. The present disclosure can also be practiced in a distributed computing environment where tasks are performed by remote processing devices linked through a communication network.

[0195] As used herein, the statement "and / or" with respect to two or more elements should be interpreted as referring to only one element or combination of elements. For example, "element A, element B and / or element C" may include only element A, only element B, only element C, element A and element B, element A and element C, element B and element C, or elements A, B and C. In addition, "at least one of element A or element B" may include at least one of element A, at least one of element B, or at least one of element A and at least one of element B. Further, "at least one of element A and element B" may include at least one of element A, at least one of element B, or at least one of element A and at least one of element B.

[0196] The subject matter of the present disclosure is described in detail herein to meet statutory requirements. However, the description itself is not intended to limit the scope of the present disclosure. On the contrary, the inventors have contemplated that the claimed subject matter may also be embodied in other ways to include steps that are different from the steps described herein in conjunction with other current or future technologies, or combinations of similar steps. Moreover, although the terms "step" and / or "block" may be used herein to imply different elements of the method employed, these terms should not be interpreted as implying any particular order among or between the various steps disclosed herein, unless the order of the steps is explicitly described.

Claims

1. A method for emergency response vehicle detection, comprising: generating, based at least on first audio data generated using one or more microphones of a machine, a first representation of a first spectrum of one or more first frequencies corresponding to one or more first audio signals from the first audio data, the first audio data being associated with a first time period; computing second data using one or more neural networks and based on at least first data representing the first representation; generating, based at least on second audio data generated using the one or more microphones of the machine, a second representation of a second spectrum of one or more second frequencies corresponding to one or more second audio signals from the second audio data, the second audio data being associated with a second time period different from the first time period; calculating, using the one or more neural networks and based at least on the second data and third data representing the second representation, fourth data representing one or more probabilities of one or more alert types; determining a type of emergency response vehicle based at least on the one or more probabilities; as well as One or more operations are performed by the machine based at least on the type of the emergency response vehicle. 2 . The method of claim 1 , wherein the first representation of the first spectrum of the one or more first frequencies or the second representation of the second spectrum of the one or more second frequencies comprises a Mel spectrogram.

3. The method of claim 1 , wherein the one or more neural networks comprise a convolutional recurrent neural network (CRNN).

4. The method of claim 1 , wherein the one or more alert types correspond to at least one of: Alarm; A sequence or pattern of one or more alarms; alarm; a sequence or pattern of one or more alarms; One or more emissions from a vehicle horn; or A sequence or pattern of emissions from at least one vehicle horn. 5 . The method of claim 1 , wherein the one or more microphones are disposed within a plurality of microphone arrays of the machine, individual microphone arrays of the plurality of microphone arrays comprising a plurality of microphones.

6. The method of claim 5, wherein the plurality of microphone arrays comprises two or more of a first microphone array disposed at the front of the machine, a second microphone array disposed at the rear of the machine, a third microphone array disposed at the left of the machine, a fourth microphone array disposed at the right of the machine, or a fifth microphone array disposed at the top of the machine. The method of claim 5 , wherein the individual microphone arrays of the plurality of microphone arrays include a windshield disposed thereon.

8. The method according to claim 1, further comprising: Background noise is filtered out from the first audio data to generate processed audio data, wherein generating the first representation is based at least on the processed audio data.

9. The method according to claim 1, wherein calculating the second data comprises: processing the first data representing the first representation using one or more gated linear units (GLUs) of the one or more neural networks; as well as Fifth data is processed using one or more gated recurrent units (GRUs), the fifth data being generated based at least on outputs of the one or more GRUs. 10 . The method according to claim 9 , wherein attention is applied to sixth data generated based on at least the output of the GRU.

11. The method of claim 1 , wherein the one or more neural networks are trained using recorded audio data and enhanced audio data, the enhanced audio data being generated using the recorded audio data and one or more of enhancement techniques, the one or more enhancement techniques comprising at least one of time stretching, time shifting, pitch shifting, dynamic range compression, and noise enhancement at different signal-to-noise ratios (SNRs).

12. The method of claim 1 , wherein the second data is associated with one or more second probabilities of one or more alert types; and the method further comprises: A type of the emergency response vehicle is determined based at least on the one or more second probabilities.

13. The method according to claim 1, wherein the second data represents at least one of the following: one or more second probabilities of the one or more alert types; and One or more values ​​stored by one or more layers of the one or more neural networks.

14. A method for emergency response vehicle detection, comprising: calculating, using one or more neural networks and based at least on first audio data generated using one or more microphone arrays of the machine, first data representing at least a first probability associated with a first alert type and a second probability associated with a second alert type; calculating, using the one or more neural networks and based on at least the first audio data and a portion of second audio data generated using the one or more microphone arrays of the machine, second data representing at least a third probability associated with the first alert type and a fourth probability associated with the second alert type, wherein the first audio data is associated with a first time window and the portion of the first audio data and the second audio data are associated with a second time window; calculating third data representing a type of emergency response vehicle based at least on the first data and the second data; and One or more operations are performed based at least on the third data.

15. The method according to claim 14, further comprising: A spectrogram is generated based on at least the first audio data, wherein the first data is calculated based on at least the spectrogram.

16. The method according to claim 15, further comprising: preprocessing the first audio data using an ambient noise suppression algorithm to generate processed audio data, Wherein generating the spectrogram is based at least on the processed audio data.

17. The method of claim 14, wherein the one or more neural networks comprise one or more feature extraction layers, and wherein the one or more feature extraction layers comprise one or more gated linear units (GLUs).

18. The method of claim 14, wherein the one or more neural networks comprise one or more stateful layers, and wherein the one or more stateful layers comprise one or more gated recurrent units (GRUs).

19. The method of claim 14, wherein the one or more neural networks comprise one or more attention layers, the one or more attention layers comprising one or more dense layers using at least one of a SoftMax function or a sigmoid function.

20. The method of claim 14, wherein determining the type of emergency response vehicle comprises: determining that the first probability is greater than the second probability; determining a type of the emergency response vehicle based at least on the first probability being greater than the second probability; determining that the third probability is greater than the fourth probability; and The type of the emergency response vehicle is verified based at least on the third probability being greater than the fourth probability.

21. A system for emergency response vehicle detection, comprising: One or more processing units for: calculating, using one or more neural networks and based at least on first audio data generated using one or more microphone arrays of the machine, first data representing at least a first probability associated with a first alert type and a second probability associated with a second alert type; calculating, using the one or more neural networks and based on at least the first audio data and a portion of second audio data generated using the one or more microphone arrays of the machine, second data representing at least a third probability associated with the first alert type and a fourth probability associated with the second alert type, wherein the first audio data is associated with a first time window and the portion of the first audio data and the second audio data are associated with a second time window; determining a type of emergency response vehicle based on at least the first probability, the second probability, the third probability, and the fourth probability; and One or more operations are performed with respect to the machine based at least on the type of emergency response vehicle.

22. The system of claim 21 , wherein the one or more microphone arrays comprise two or more of a first microphone array disposed on the front of the machine, a second microphone array disposed on the rear of the machine, a third microphone array disposed on the left of the machine, a fourth microphone array disposed on the right of the machine, or a fifth microphone array disposed on the top of the machine.

23. The system of claim 21 , wherein determining the type of the emergency response vehicle comprises: determining that the first probability is greater than the second probability and the third probability is greater than the fourth probability; as well as The type of the emergency response vehicle is determined based at least on the first probability being greater than the second probability and the third probability being greater than the fourth probability.

Images