Test method, test system and test device of intelligent helmet

Abstract

The invention provides a test method, a test system and a test device for an intelligent helmet, which are applied to the field of helmet testing, in particular to the test system for the intelligent helmet, and the method comprises the following steps: a noise player obtains original noise data; the original noise data are classified through a first neural network to obtain first classification data, and the first classification data comprise at least one of flat road data, crossroad data, congested road section data and roundabout data; respectively playing the first classification data, and simultaneously instructing a player to play instruction information; and obtaining a test result of the intelligent helmet under the first classification data according to the instruction information and the audio signal. The objective of the invention is to classify original noise data so as to carry out noise testing in various scenes after classification, and the intelligent helmet can be tested more accurately and effectively.

Description

Technical Field

[0001] The present application relates to the field of helmet testing, and more specifically, to a testing method, a testing system and a testing device for a smart helmet. Background Art

[0002] Before smart helmets are put on the market, they need to undergo noise testing. On the one hand, this helps improve the user experience and ensures that users can still communicate clearly and use the helmet comfortably in noisy environments. On the other hand, noise testing can verify the effectiveness of the helmet in protecting hearing from damage. In addition, the test results can provide feedback for product design and material selection, which helps optimize product performance while proving the competitiveness of its products and meeting the needs of specific industries or consumers.

[0003] In existing testing methods, testing is usually carried out in an actual usage environment, such as having a user wear a helmet while riding on the road and using a portable device to collect data. During this testing process, external factors (such as weather and traffic) may affect the test results.

[0004] In view of this, how to provide a testing method that can accurately perform helmet noise testing indoors is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The present application provides a testing method, a testing system and a testing device for a smart helmet, which aims to classify original noise data so as to perform noise tests in various classified scenarios, so as to test the smart helmet more accurately and effectively.

[0006] In a first aspect, a test method for a smart helmet is provided, which is applied to a test system for a smart helmet. The test system includes a simulator, a smart helmet installed on the head of the simulator, and a noise player. An instruction player is provided inside the simulator, and the noise player is installed on the smart helmet. The method includes: the noise player obtains original noise data; the noise player classifies the original noise data through a first neural network to obtain first classified data, and the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data; the noise player plays the first classified data respectively, and the instruction player plays instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play an audio signal corresponding to the instruction information; the smart helmet obtains a test result of the smart helmet under the first classified data according to the instruction information; the simulator obtains a test result of the smart helmet under the first classified data according to the audio signal.

[0007] Based on the above scheme, the noise player in the test system processes and classifies the acquired raw noise data, and classifies it into noise data in various scenarios under real environments, so that the smart helmet can be tested in various scenarios, and the test results obtained are more accurate. For example, in congested roads, the noise is very noisy, and the final test results may be misjudged as quality problems of the smart helmet, but in fact, environmental factors affect the final test results. In addition, in the test system of the present application, there is no real person involved in the whole process, and the relevant operations are performed by simulating people, which saves manpower and improves efficiency.

[0008] In combination with the first aspect, in certain implementations of the first aspect, the first neural network includes a first network and a second network, and the noise player classifies the original noise data through the first neural network to obtain first classified data, including: the noise player inputs the original noise data into the first network, and obtains feature data after feature extraction; the noise player inputs the feature data into the second network to obtain the first classified data.

[0009] Based on the above solution, the first classified data is obtained by mapping the feature data after feature extraction to different noise types, that is, the first classified data is finally obtained through feature extraction and classification.

[0010] In combination with the first aspect, in some implementations of the first aspect, the first network includes a convolution layer, a pooling layer, a multi-layer convolution layer and a recursive layer, wherein the convolution layer and the pooling layer are used to obtain the frequency characteristics and amplitude characteristics of the original noise data to obtain frequency characteristic data and intensity characteristic data; the multi-layer convolution layer is used to obtain the frequency interval characteristics of the original noise data to obtain periodic noise characteristic data; the recursive layer is used to obtain the change characteristics of the original noise data in a time series to obtain vehicle turning feature data, and the feature data includes frequency characteristic data, intensity characteristic data, periodic noise characteristic data and vehicle turning feature data.

[0011] Based on the above solution, the feature data after feature extraction is mapped to different noise types, for example, the frequency feature after feature extraction is mapped to the congested road section data in the first classification data.

[0012] In combination with the first aspect, in certain implementations of the first aspect, the method also includes: the noise player obtains sensor data, the sensor data includes speed data and status data of the moving vehicle; the noise player classifies the original noise data through a first neural network to obtain first classified data, including: the noise player classifies the original noise data and sensor data through the first neural network to obtain the first classified data.

[0013] In combination with the first aspect, in certain implementations of the first aspect, the noise player classifies the original noise data and sensor data through a first neural network to obtain first classified data, including: when the intersection data and congested road section data cannot be classified according to the frequency characteristic data, classification is performed according to the sensor data.

[0014] Based on the above scheme, when it is not possible to perform good feature extraction relying solely on the original noise data, the sensor data is sent to the first neural network to assist the original noise signal for better feature extraction.

[0015] In combination with the first aspect, in certain implementations of the first aspect, when the intersection data and the congested road section data cannot be classified according to the frequency characteristic data, classification is performed according to the sensor data, including: when the speed of the vehicle is greater than a preset speed and the state of the vehicle is turning, classifying the original noise data as intersection data; when the speed of the vehicle is less than or equal to the preset speed and the state of the vehicle is going straight, classifying the original noise data as congested road section data.

[0016] In combination with the first aspect, in certain implementations of the first aspect, the original noise data is obtained by collecting actual noise of vehicles on the road through a sound level meter, and the original noise data includes at least one of the following: vehicle horn sounds, turning sounds, deceleration sounds, and periodic sound noises generated by vehicles.

[0017] In combination with the first aspect, in some implementations of the first aspect, the noise player is installed on the left side of the smart helmet.

[0018] In combination with the first aspect, in certain implementations of the first aspect, the instruction information includes control instructions, navigation instructions, safety reminder instructions, status query instructions, and emergency instructions.

[0019] In combination with the first aspect, in certain implementations of the first aspect, obtaining a test result of the smart helmet based on the instruction information and the audio signal includes: the smart helmet compares the instruction information with preset instruction information to obtain the test result of the smart helmet.

[0020] In combination with the first aspect, in certain implementations of the first aspect, the smart helmet compares the instruction information with the preset instruction information to obtain a test result of the smart helmet, including: when the instruction information is the same as the preset instruction information, determining that the quality of the smart helmet is qualified.

[0021] In combination with the first aspect, in certain implementations of the first aspect, a test result of the smart helmet is obtained based on the instruction information and the audio signal, including: passing the audio signal through an audio quality perceptual evaluation algorithm to obtain a quality score, and determining whether the quality of the smart helmet is qualified based on the quality score.

[0022] In combination with the first aspect, in certain implementations of the first aspect, the test system further includes a fan, and the fan is located directly in front of the smart helmet. The method further includes: while playing the first classification data and the instruction information, turning on the fan to generate wind noise.

[0023] In combination with the first aspect, in certain implementations of the first aspect, the first neural network is trained by: obtaining first sample data and label data; training an initial network based on the first sample data and the label data to obtain the first neural network; wherein the first sample data includes multiple groups of original noise data and sensor data, and the label data includes multiple groups of first classification data.

[0024] In combination with the first aspect, in certain implementations of the first aspect, the initial network is trained to obtain the first neural network based on the first sample data and the label data, including: inputting the first sample data and the label into the initial network to obtain second classification data; and adjusting the parameters of the initial network based on the second classification data to train the first neural network.

[0025] In combination with the first aspect, in certain implementations of the first aspect, adjusting the parameters of the initial network according to the second classified data includes: performing a first function operation on the second classified data to adjust the parameters of the initial network, and the first function operation includes a cross entropy function operation.

[0026] In a second aspect, a test system for a smart helmet is provided, comprising a simulator, a smart helmet installed on the simulator's head, and a noise player, wherein an instruction player is provided inside the simulator, the noise player is installed on the smart helmet, and the noise player obtains original noise data; the noise player classifies the original noise data through a first neural network to obtain first classified data, and the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data; the noise player plays the first classified data respectively, and the instruction player plays instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play an audio signal corresponding to the instruction information; the smart helmet obtains a test result of the smart helmet under the first classified data according to the instruction information; the simulator obtains a test result of the smart helmet under the first classified data according to the audio signal.

[0027] In a third aspect, a testing device for a smart helmet is provided, comprising: a processor and a memory, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to execute: the testing method in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic structural block diagram of a communication system of a smart helmet provided in an embodiment of the present application.

[0029] Figure 2 It is a schematic structural block diagram of a smart helmet provided in an embodiment of the present application.

[0030] Figure 3 It is a schematic flowchart of a testing method for a smart helmet provided in an embodiment of the present application.

[0031] Figure 4 It is a schematic structural block diagram of a test system for a smart helmet provided in an embodiment of the present application.

[0032] Figure 5 It is a schematic structural block diagram of a testing device for a smart helmet provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] The technical solution in this application will be described below in conjunction with the accompanying drawings.

[0034] This application involves testing of helmets in noisy environments, that is, before helmets are put on the market, they need to be tested for noise. On the one hand, this helps improve the user experience and ensures that users can still communicate clearly and use the helmets comfortably in noisy environments. On the other hand, noise testing can verify the effectiveness of helmets in protecting hearing from damage. In addition, test results can provide feedback for product design and material selection, which helps to optimize product performance while proving the competitiveness of its products and meeting the needs of specific industries or consumers. Optionally, the helmet can be a smart helmet, that is, some functions are added on the basis of the safety protection of ordinary helmets, as follows.

[0035] Figure 1 A schematic structural block diagram of a communication system 100 for a smart helmet provided in an embodiment of the present application is shown.

[0036] like Figure 1 As shown, the communication system 100 of the smart helmet includes a helmet 110 and a user terminal 120 .

[0037] The smart helmet 110 is usually equipped with a Bluetooth or Wi-Fi module, which can be connected to the user terminal 120 to achieve information exchange and remote control. The user terminal 120 can be, for example, a mobile phone, a computer, etc., which communicates with the smart helmet 110. The user can send relevant control instructions to the smart helmet 110 through the user terminal 120.

[0038] It should be understood that in this application, the smart helmet is taken as an example to illustrate the relevant technical solutions. In the case where the helmet is other types of helmets, the implementation details can be found in the specific description below, and this article will not go into details.

[0039] Figure 2A schematic structural block diagram of a smart helmet provided in an embodiment of the present application is shown.

[0040] like Figure 2 As shown, the smart helmet includes a navigation unit 210 , a safety assistance unit 220 , a data recording unit 230 and an emergency assistance unit 240 .

[0041] Among them, the navigation unit 210 in the smart helmet is connected to the user terminal to provide navigation information, call reminders, text message reading and other functions, and can also support voice calls.

[0042] The safety assistance unit 220 is equipped with a light or motion sensor in the smart helmet, which can automatically turn on the warning light at night or in conditions of poor visibility, or alert the user when abnormal movements are detected.

[0043] The data recording unit 230 can be used to record the user's riding data, such as speed, distance, calorie consumption, etc., and analyze it through the user terminal.

[0044] The emergency assistance unit 240 can send a distress signal to a connected user terminal when encountering an emergency.

[0045] It should be understood that the functions of the smart helmet are not limited to the parts described above, there are other functional units, such as camera photography function, etc., which will not be elaborated in this article.

[0046] In the existing testing methods of smart helmets, the tests are usually carried out in actual environments, such as when a user wears a helmet while riding on the road, and uses a portable device to collect data. During this test, external factors (such as weather and traffic) may affect the test results, and there may be safety hazards during the test.

[0047] In view of this, the present application proposes a testing method for a smart helmet, which aims to test the smart helmet indoors and process the collected outdoor noise so as to accurately obtain the test results during the test. In addition, through data processing, the detection process can be divided into multiple scenarios for testing.

[0048] The testing method of the smart helmet in the present application is applied to a testing system of the smart helmet, which includes a simulated person, a smart helmet installed on the head of the simulated person, a noise player and a fan.

[0049] The simulator is internally provided with a command player, an audio receiving device and an audio processing device.

[0050] The smart helmet includes a control unit for receiving instruction information and playing corresponding audio signals according to the instruction information.

[0051] The fan can be placed in front of the simulated person because the direction of the wind is opposite to the riding direction during the riding of the electric vehicle, so the fan is placed in front of the simulated person to more realistically simulate the actual riding scene.

[0052] The noise player can be placed on the left side of the helmet. This is because electric vehicles usually travel on the non-motorized vehicle lane on the right during riding, and the noise comes from the left side of the electric vehicle. Therefore, in order to more realistically simulate the real scene, the noise player can be placed on the left side of the helmet.

[0053] It should be understood that the placement of the fan and the noise player may also be adjusted accordingly based on actual test conditions, and this application does not impose any limitation on this.

[0054] Figure 3 A schematic flowchart of a smart helmet testing method 300 provided in an embodiment of the present application is shown.

[0055] like Figure 3 As shown, the testing method 300 of the smart helmet includes the following steps.

[0056] S310: The noise player obtains original noise data.

[0057] S320: The noise player classifies the original noise data through a first neural network to obtain first classified data, where the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data.

[0058] S330: The noise player plays the first classification data respectively, and the instruction player plays the instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play the audio signal corresponding to the instruction information.

[0059] S340: The smart helmet obtains a test result of the smart helmet under the first classification data according to the instruction information.

[0060] S350: The simulated person obtains a test result of the smart helmet under the first classification data according to the audio signal.

[0061] In the above step S310, the original noise data is obtained by collecting actual noise data on the road through a collection device, including whistle sounds, vehicle driving sounds, etc., wherein the collection device can be a sound level meter, a microphone array, etc.

[0062] Optionally, the noise player obtains sensor data, wherein the sensor data includes speed data and status data of a moving vehicle. The sensor data is information such as the type of vehicle, the speed of the vehicle, and the status of the vehicle obtained by the sensor in the process of collecting actual noise on the road.

[0063] In the above step S320, the noise player includes a first neural network. Based on the first neural network, the original noise data is classified to finally obtain noise data in various road scenes, namely, first classified data, which can be specifically divided into flat road data, intersection data, congested road section data, roundabout data, etc.

[0064] Optionally, before the original noise data is input into the first neural network, the original noise data may be first input into the information refining network to remove irrelevant static noise and interference, wherein the static noise and interference include equipment noise, electromagnetic interference, static noise, and the like.

[0065] More specifically, device noise includes noise generated by the sensor itself, such as thermal noise, circuit noise, etc. These noises are determined by the physical characteristics of the sensor hardware and affect the quality of the data. Electromagnetic interference is generated by electronic devices on the road, radio signals, etc. Static noise is a type of noise that does not change over time. It may be a fixed pattern noise generated by the sensor being exposed to the external environment for a long time.

[0066] Optionally, the first neural network includes a first network and a second network, and the noise player inputs the original noise data into the first network to obtain feature data after feature extraction; the noise player inputs the feature data into the second network to obtain first classification data.

[0067] Specifically, the first network includes a convolutional layer, a pooling layer, a multi-layer convolutional layer and a recursive layer, and the process of feature extraction is as follows:

[0068] The convolution layer and pooling layer are used to obtain the frequency characteristics and amplitude characteristics of the original noise data, and obtain the frequency characteristic data and intensity characteristic data based on the frequency characteristics and amplitude characteristics. For example, the first convolution layer in the convolution layer uses a small-sized convolution kernel, such as 3x3, to capture the frequency component characteristics of the whistle.

[0069] Multi-layer convolutional layers are used to obtain the frequency interval characteristics of the original noise data to obtain periodic noise feature data.

[0070] The recursive layer is used to obtain the changing characteristics of the original noise data in time series to obtain the vehicle turning feature data. For example, the recursive layer uses the Long Short-Term Memory (LSTM) or Gated Recurrent Unit (GRU) layer to process time series information and capture the dynamic changes of sound, such as turning and deceleration noise.

[0071] The characteristic data include frequency characteristic data, intensity characteristic data, periodic noise characteristic data, vehicle turning characteristic data and the like.

[0072] Optionally, in feature extraction, frequency feature extraction uses multiple convolutional layers and pooling layers, and the data processing network can identify the frequency features of the whistle from the time-frequency graph of the audio signal. In the sound intensity feature, the amplitude of the audio signal can represent the sound intensity, and the data processing network identifies the intensity feature by learning the amplitude information in the time-frequency graph. Turning noise and deceleration noise can capture the time dependency of the sound signal through the LSTM layer, and identify turning and deceleration noise by learning the changes in the time series. Periodic noise By using filters in the convolution layer, the data processing network can identify periodic patterns, such as periodic noise caused by traffic flow.

[0073] Specifically, the second network includes a fully connected layer and an output layer.

[0074] More specifically, the fully connected layer maps the extracted feature data to different noise types. The output layer uses the softmax activation function to output the probability of each category.

[0075] Optionally, the noise player obtains sensor data, the sensor data including speed data and state data of the moving vehicle. The noise player classifies the original noise data and the sensor data through a first neural network to obtain first classified data.

[0076] For example, when the intersection data and the congested road section data cannot be classified according to the frequency characteristic data, the classification is performed according to the sensor data. For example, when the speed of the vehicle is greater than the preset speed and the vehicle is in a turning state, the raw noise data is classified as intersection data; when the speed of the vehicle is less than or equal to the preset speed and the vehicle is in a straight state, the raw noise data is classified as congested road section data.

[0077] For another example, when the intersection data and the roundabout data cannot be classified according to the periodic noise feature data, the classification is performed according to the sensor data. For example, when the vehicle is in a continuous turning state, the original noise data is classified as the roundabout data.

[0078] In the above step S330, the noise player plays the first classification data respectively, and the instruction player plays the instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play the audio signal corresponding to the instruction information. The instruction information may include control instructions, navigation instructions, safety reminders, status inquiries, system settings, emergency instructions, etc.

[0079] For example, the instruction player plays "Helmet, play music", and the smart helmet connects to the user terminal and starts playing music.

[0080] For another example, if the command player plays "Helmet, navigate to the nearest coffee shop", the smart helmet will start the navigation system and plan the route.

[0081] For another example, instruct the player to play "Helmet, please check the weather conditions", and the smart helmet will broadcast the current weather conditions by voice.

[0082] For another example, the instruction player plays "Helmet, connect to my phone", and the smart helmet will start searching and connect to the user's phone.

[0083] Among them, the smart helmet and the user terminal can be connected via Bluetooth or WIFI and other modes, and this application does not impose any restrictions on this.

[0084] Optionally, during the test, a fan located in front of the smart helmet is used to simulate the wind noise during the user's riding process. The wind noise can be controlled indoors by adjusting the fan gear and the distance between the fan and the smart helmet.

[0085] It should be understood that the audio signal in the present application may be played music, planned route information, forecasted weather information, Bluetooth connection signal, etc., and the present application does not impose any limitation on this.

[0086] In the above step S340, the smart helmet obtains the test result of the smart helmet under the first classification data according to the instruction information.

[0087] In one possible implementation, the smart helmet compares the instruction information with the preset instruction information to obtain the test result of the smart helmet. If the instruction information is the same as the preset instruction information, the quality of the smart helmet is determined to be qualified. It can also be expressed as comparing the instruction information received by the smart helmet in a noisy noise environment with the instruction information in the instruction library. If the received instruction information can be found in the instruction library, the quality of the smart helmet is determined to be qualified.

[0088] In the above step S350, the simulated person obtains the test result of the smart helmet under the first classification data according to the audio signal.

[0089] One possible implementation method is to simulate a person passing an audio signal through a perceptual evaluation of audio quality (PEAQ) algorithm to obtain a quality score, and then determine whether the quality of the smart helmet is qualified based on the quality score.

[0090] The audio receiving device in the simulated person receives the audio signal, and the audio processing device in the simulated person processes the received audio signal.

[0091] The specific test result analysis is carried out by simulating the human ear's perception of audio quality through PEAQ or ITU-T P.862 (PESQ-Perceptual Evaluation of Speech Quality) to obtain the test results.

[0092] For example, in this application, PEAQ is used for result judgment. The specific process is: sampling rate matching and synchronization processing are performed on the reference audio and the test audio, and then a filter group is used to perform time-frequency conversion on the audio signal to obtain a time-frequency signal. The characteristics of the audio signal are extracted from the obtained time-frequency representation signal, such as the spectrum envelope, time domain waveform, harmonic structure, etc.

[0093] The features of the extracted audio signal are passed through the auditory masking model and the perceptual model respectively. The auditory masking model is used to evaluate the impact of different frequency components on auditory perception and determine which differences are significant and which may be masked. The perceptual model is used to simulate the human ear's perception of audio signals, including sensitivity to signal distortion and interference. The perceptual model takes into account the transience and continuity of audio signals and their impact on perceptual quality.

[0094] Based on the feature differences and the results of the auditory masking model, the feature differences between the reference audio and the test audio are calculated, and the feature differences are converted into quality scores.

[0095] The quality score can be a value between -4 and 0, where 0 indicates imperceptible distortion and -4 indicates very poor audio quality.

[0096] Optionally, the audio processing device obtains the accuracy of the audio signal received by the simulated person in the noisy environment according to the final quality score. If the accuracy of the audio signal received by the simulated person in the environment is not high, it can be indicated that there may be a problem with the playback device of the smart helmet, and then corresponding repairs are performed.

[0097] It should be understood that by classifying the original noise data into data for various scenarios, the smart helmet can be tested more accurately. For example, testing can be performed on congested roads and on flat roads. The horns of vehicles on congested roads are louder. In this environment, the tester issues instructions, but the smart helmet may not hear the instructions. The final test result may mistakenly be due to quality problems with the helmet. Therefore, more detailed classification of the smart helmet can make the test results of the smart helmet more accurate.

[0098] Optionally, the first neural network is trained in the following manner: obtaining first sample data and label data; training an initial network according to the first sample data and label data to obtain the first neural network; wherein the first sample data includes multiple groups of original noise data and sensor data, and the label data includes multiple groups of first classification data.

[0099] It should be understood that the training of the first neural network can be performed offline on the server, and this application does not impose any limitation on this.

[0100] Specifically, the first sample data and the label are input into the initial network to obtain the second classification data; according to the second classification data, the parameters of the initial network are adjusted to train the first neural network. The second classification data is processed by a first function operation to adjust the parameters of the initial network, and the first function operation includes a cross entropy function operation.

[0101] It should be understood that the loss function of the initial network is set to the cross entropy loss function, which is suitable for multi-classification problems. The optimizer uses Adam or SGD to adjust the network weights.

[0102] Through the technical solution of the embodiment of the present application, the noise player in the test system processes and classifies the acquired raw noise data, and classifies it into noise data in various scenarios under real environments, so that the smart helmet can be tested in various scenarios, and the obtained test results are more accurate. In addition, in the test system of the present application, there is no real person involved in the whole process, and related operations are performed by simulating people, which saves manpower and improves efficiency.

[0103] Combination of the above Figure 3 The method embodiment provided by the present application is described below. Figure 4 and Figure 5 The device embodiments provided in the present application are described. It is understandable that the device embodiments below may correspond to the method embodiments above, and the specific solutions may refer to the above description, and the repeated parts will not be described in detail.

[0104] Figure 4 A schematic structural block diagram of a smart helmet testing system 400 provided in an embodiment of the present application is shown. Optionally, the smart helmet testing system 400 may be the smart helmet testing system in the above method embodiment.

[0105] like Figure 4 As shown, the test system 400 includes: a simulator 410, a smart helmet 420 installed on the head of the simulator 410, and a noise player 430. The simulator 410 is provided with an instruction player inside, and the noise player 430 is installed on the smart helmet 420.

[0106] The noise player 430 acquires raw noise data.

[0107] The noise player 430 classifies the original noise data through a first neural network to obtain first classified data, where the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data.

[0108] The noise player 430 plays the first classification data respectively, and at the same time instructs the player to play the instruction information, wherein the instruction information is used to instruct the smart helmet 420 to play the audio signal corresponding to the instruction information.

[0109] The smart helmet 420 obtains the test result of the smart helmet 420 under the first classification data according to the instruction information.

[0110] The simulated person 410 obtains the test result of the smart helmet 420 under the first classification data according to the audio signal.

[0111] Figure 5 A schematic structural block diagram of a smart helmet testing device 500 provided in an embodiment of the present application is shown.

[0112] like Figure 5 As shown, the test device 500 of the smart helmet includes: a processor 510 and a memory 520, the memory 520 is used to store a computer program, and the processor 510 is used to call and run the computer program stored in the memory 520 to execute the above Figure 3 The testing method of the smart helmet provided in any embodiment.

[0113] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0115] It should be understood that the specific examples in this article are only intended to help those skilled in the art better understand the embodiments of the present application, rather than to limit the scope of the embodiments of the present application.

[0116] It should also be understood that in the various embodiments of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0117] It should also be understood that the various implementation methods described in this specification can be implemented individually or in combination, and the embodiments of the present application are not limited to this.

[0118] Unless otherwise stated, all technical and scientific terms used in the embodiments of the present application are the same as the meanings generally understood by those skilled in the art of the technical field of the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the scope of the present application. The term "and / or" used in the present application includes any and all combinations of one or more related listed items. The singular forms of "a kind of", "above" and "the" used in the embodiments of the present application and the appended claims are also intended to include majority forms, unless the context clearly indicates other meanings. In addition, the terms "first", "second", "third" etc. are only used for description purposes and cannot be understood as indicating or suggesting relative importance.

[0119] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0120] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0121] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0122] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for testing a smart helmet, characterized in that: A test system applied to the smart helmet, the test system comprising a simulated person, the smart helmet mounted on the head of the simulated person, and a noise player, wherein an instruction player is arranged inside the simulated person, and the noise player is mounted on the smart helmet, and the method comprises: The noise player obtains original noise data; The noise player classifies the original noise data through a first neural network to obtain first classified data, wherein the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data; The noise player plays the first classification data respectively, and the instruction player plays instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play the audio signal corresponding to the instruction information; The smart helmet obtains a test result of the smart helmet under the first classification data according to the instruction information; The simulated person obtains the test result of the smart helmet under the first classification data according to the audio signal.

2. The testing method according to claim 1, characterized in that: The first neural network includes a first network and a second network, The noise player classifies the original noise data through a first neural network to obtain first classified data, including: The noise player inputs the original noise data into the first network, and obtains feature data after feature extraction; The noise player inputs the feature data into a second network to obtain the first classification data.

3. The testing method according to claim 2, characterized in that: The first network includes a convolutional layer, a pooling layer, a multi-layer convolutional layer and a recursive layer. The convolution layer and the pooling layer are used to obtain the frequency characteristics and amplitude characteristics of the original noise data to obtain frequency characteristic data and intensity characteristic data; The multi-layer convolution layer is used to obtain the frequency interval characteristics of the original noise data to obtain periodic noise characteristic data; The recursive layer is used to obtain the change characteristics of the original noise data in time series to obtain vehicle turning feature data. The characteristic data includes the frequency characteristic data, the intensity characteristic data, the periodic noise characteristic data and the vehicle turning characteristic data.

4. The testing method according to claim 3, characterized in that: The method further comprises: The noise player acquires sensor data, wherein the sensor data includes speed data and state data of the moving vehicle; The noise player classifies the original noise data through a first neural network to obtain first classified data, including: The noise player classifies the original noise data and the sensor data through the first neural network to obtain the first classified data.

5. The testing method according to claim 4, characterized in that: The noise player classifies the original noise data and the sensor data through the first neural network to obtain the first classified data, including: When the intersection data and the congested road section data cannot be classified according to the frequency characteristic data, classification is performed according to the sensor data.

6. The testing method according to claim 5, characterized in that: In the case that the intersection data and the congested road section data cannot be classified according to the frequency characteristic data, classifying according to the sensor data includes: When the speed of the vehicle is greater than a preset speed and the vehicle is in a turning state, classifying the original noise data as the intersection data; When the speed of the vehicle is less than or equal to the preset speed and the vehicle is in a straight-moving state, the original noise data is classified as the congested road section data.

7. The testing method according to any one of claims 1 to 6, characterized in that: The original noise data is obtained by collecting actual noise of vehicles on the road through a sound level meter, and the original noise data includes at least one of the following: a horn sound, a turning sound, a deceleration sound of the vehicle, and a periodic sound noise generated by the vehicle.

8. The testing method according to any one of claims 1 to 6, characterized in that: The noise player is installed on the left side of the smart helmet.

9. The testing method according to any one of claims 1 to 6, characterized in that: The command information includes control commands, navigation commands, safety reminder commands, status query commands and emergency commands.

10. The testing method according to any one of claims 1 to 6, characterized in that: The smart helmet obtains a test result of the smart helmet under the first classification data according to the instruction information, including: The smart helmet compares the instruction information with preset instruction information to obtain a test result of the smart helmet.

11. The testing method according to claim 10, characterized in that: The smart helmet compares the instruction information with preset instruction information to obtain a test result of the smart helmet, including: When the instruction information is the same as the preset instruction information, it is determined that the quality of the smart helmet is qualified.

12. The testing method according to any one of claims 1 to 10, characterized in that: The simulated person obtains, according to the audio signal, a test result of the smart helmet under the first classification data, including: The audio signal is passed through an audio quality perceptual evaluation algorithm to obtain a quality score, According to the quality score, it is determined whether the quality of the smart helmet is qualified.

13. The testing method according to any one of claims 1 to 10, characterized in that: The testing system further includes a fan, and the fan is located directly in front of the smart helmet. The method further includes: While playing the first classification data and the instruction information, the fan is turned on to generate wind noise.

14. The testing method according to any one of claims 1 to 10, characterized in that: The first neural network is trained in the following way: Obtaining first sample data and label data; According to the first sample data and the label data, training the initial network to obtain the first neural network; The first sample data includes multiple groups of original noise data and sensor data, and the label data includes multiple groups of first classification data.

15. The testing method according to claim 14, characterized in that: The step of training the initial network according to the first sample data and the label data to obtain the first neural network includes: Inputting the first sample data and the label into the initial network to obtain second classification data; According to the second classification data, the parameters of the initial network are adjusted to train the first neural network.

16. The testing method according to claim 15, characterized in that: The adjusting the parameters of the initial network according to the second classification data includes: The second classification data is processed by a first function operation to adjust the parameters of the initial network, wherein the first function operation includes a cross entropy function operation.

17. A testing system for a smart helmet, characterized in that: The invention comprises a manikin, the smart helmet installed on the manikin's head and a noise player, wherein the manikin is provided with an instruction player inside, and the noise player is installed on the smart helmet. The noise player obtains original noise data; The noise player classifies the original noise data through a first neural network to obtain first classified data, wherein the first classified data includes at least one of the following: flat road data, intersection data, congested road section data, and roundabout data; The noise player plays the first classification data respectively, and the instruction player plays instruction information at the same time, wherein the instruction information is used to instruct the smart helmet to play the audio signal corresponding to the instruction information; The smart helmet obtains a test result of the smart helmet under the first classification data according to the instruction information; The simulated person obtains the test result of the smart helmet under the first classification data according to the audio signal.

18. A testing device for a smart helmet, characterized in that: include: A processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute: the test method according to any one of claims 1 to 16.

Images