WO2022150999A1 - 一种检测轮胎异常的方法和装置 - Google Patents
一种检测轮胎异常的方法和装置 Download PDFInfo
- Publication number
- WO2022150999A1 WO2022150999A1 PCT/CN2021/071370 CN2021071370W WO2022150999A1 WO 2022150999 A1 WO2022150999 A1 WO 2022150999A1 CN 2021071370 W CN2021071370 W CN 2021071370W WO 2022150999 A1 WO2022150999 A1 WO 2022150999A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- tire
- tire pressure
- data
- pressure data
- audio
- Prior art date
Links
- 230000005856 abnormality Effects 0.000 title claims abstract description 76
- 238000001514 detection method Methods 0.000 title abstract description 37
- 230000002159 abnormal effect Effects 0.000 claims abstract description 128
- 238000000034 method Methods 0.000 claims abstract description 51
- 230000004807 localization Effects 0.000 claims abstract description 29
- 230000007423 decrease Effects 0.000 claims abstract description 24
- 238000004458 analytical method Methods 0.000 claims abstract description 22
- 230000008569 process Effects 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 11
- 230000003213 activating effect Effects 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 4
- 238000004891 communication Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 238000013480 data collection Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000005236 sound signal Effects 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 206010039203 Road traffic accident Diseases 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 206010021703 Indifference Diseases 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000037433 frameshift Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60C—VEHICLE TYRES; TYRE INFLATION; TYRE CHANGING; CONNECTING VALVES TO INFLATABLE ELASTIC BODIES IN GENERAL; DEVICES OR ARRANGEMENTS RELATED TO TYRES
- B60C23/00—Devices for measuring, signalling, controlling, or distributing tyre pressure or temperature, specially adapted for mounting on vehicles; Arrangement of tyre inflating devices on vehicles, e.g. of pumps or of tanks; Tyre cooling arrangements
- B60C23/02—Signalling devices actuated by tyre pressure
- B60C23/04—Signalling devices actuated by tyre pressure mounted on the wheel or tyre
- B60C23/0401—Signalling devices actuated by tyre pressure mounted on the wheel or tyre characterised by the type of alarm
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60C—VEHICLE TYRES; TYRE INFLATION; TYRE CHANGING; CONNECTING VALVES TO INFLATABLE ELASTIC BODIES IN GENERAL; DEVICES OR ARRANGEMENTS RELATED TO TYRES
- B60C23/00—Devices for measuring, signalling, controlling, or distributing tyre pressure or temperature, specially adapted for mounting on vehicles; Arrangement of tyre inflating devices on vehicles, e.g. of pumps or of tanks; Tyre cooling arrangements
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60C—VEHICLE TYRES; TYRE INFLATION; TYRE CHANGING; CONNECTING VALVES TO INFLATABLE ELASTIC BODIES IN GENERAL; DEVICES OR ARRANGEMENTS RELATED TO TYRES
- B60C23/00—Devices for measuring, signalling, controlling, or distributing tyre pressure or temperature, specially adapted for mounting on vehicles; Arrangement of tyre inflating devices on vehicles, e.g. of pumps or of tanks; Tyre cooling arrangements
- B60C23/02—Signalling devices actuated by tyre pressure
- B60C23/04—Signalling devices actuated by tyre pressure mounted on the wheel or tyre
- B60C23/0408—Signalling devices actuated by tyre pressure mounted on the wheel or tyre transmitting the signals by non-mechanical means from the wheel or tyre to a vehicle body mounted receiver
- B60C23/0481—System diagnostic, e.g. monitoring battery voltage, detecting hardware detachments or identifying wireless transmission failures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01L—MEASURING FORCE, STRESS, TORQUE, WORK, MECHANICAL POWER, MECHANICAL EFFICIENCY, OR FLUID PRESSURE
- G01L17/00—Devices or apparatus for measuring tyre pressure or the pressure in other inflated bodies
Definitions
- the present application relates to the field of automobiles, and in particular, to a method for detecting abnormality of tires.
- the Tire Pressure Monitoring System used in automobiles can detect tire pressure in real time, and can issue an alarm or even advance warning in time when abnormal tire pressure is detected, so as to ensure the safety of vehicles and passengers.
- the basic principle of TPMS is to first set a reasonable range of vehicle tire pressure to ensure that the set value is consistent with the vehicle tire parameters. When the tire pressure deviates from the set value and is not within the reasonable range of the set vehicle tire pressure, TPMS will issue an alarm. . TPMS can alarm in time when the tire pressure is too low or there is gas leakage.
- TPMS on the market can be roughly divided into two categories: indirect TPMS and direct TPMS.
- Indirect TPMS refers to a system that measures other parameters of the vehicle, such as the four-wheel speed, rather than directly measuring the tire pressure of the car.
- the indirect TPMS cannot measure the instantaneous air pressure of the vehicle tires and has low sensitivity.
- Direct TPMS refers to a system that directly measures the tire pressure through a pressure sensor installed inside the vehicle tire, and sends the data directly to the receiver.
- the direct TMPS has high monitoring accuracy and can be Monitoring is performed with high sensitivity and no false alarms, but this direct TPMS has higher installation costs and faster battery drain.
- an alarm is usually issued when the tire pressure value is lower than a certain preset value, and an alarm cannot be issued in time to remind the driver that there is potential driving safety.
- the embodiments of the present application provide a tire abnormality detection method and device.
- an embodiment of the present application provides a method for detecting abnormality of a tire, which can be used to detect whether the tire is abnormal, including whether the tire is leaking air or not.
- the specific steps include: periodically collecting the tire pressure data of the target tire through the built-in pressure sensor on the vehicle wheel; analyzing the periodically collected tire pressure data, and identifying the situation that the tire pressure of the target tire continuously decreases; Collect audio data through the sound sensitive sensor configured on the vehicle; perform abnormal sound recognition on the audio data to identify the presence of abnormal sounds in the audio data; perform sound source localization analysis on the audio data to identify the abnormal sound source; Matching the pointing position of the abnormal sound source with the target tire, and matching the pointing position of the abnormal sound source with the target tire indicates that the target tire is abnormal.
- the continuous decrease of the tire pressure of the target tire includes: the first tire pressure data is greater than the second tire pressure data, and the difference between the first tire pressure data and the second tire pressure data is subtracted from the first tire pressure data. The value is not greater than the preset value;
- the first tire pressure data is the tire pressure data of the target tire collected in the first cycle
- the second tire pressure data of the target tire collected in the second cycle the first cycle
- the first cycle and the The second cycle is any two adjacent cycles in a plurality of consecutive cycles, and the first cycle is located before the second cycle.
- the method for detecting tire abnormality further includes: analyzing periodically collected tire pressure data to identify that the target tire has air leakage; and issuing an alarm to prompt the user that the target tire has occurred Air leak.
- the occurrence of air leakage in the target tire includes: the third tire pressure data is greater than the fourth tire pressure data, and the difference between the third tire pressure data and the fourth tire pressure data is greater than a predetermined value. set value, the third tire pressure data is the tire pressure data of the target tire collected in the third cycle, the fourth tire pressure data of the target tire collected in the fourth cycle, the first cycle and the second period are two adjacent periods, and the third period is located before the fourth period.
- the performing abnormal sound identification on the audio data to identify the presence of abnormal sounds in the audio data includes: processing the audio data to obtain sound feature parameters corresponding to the audio data; The sound feature parameters corresponding to the audio data are compared with the audio template library obtained by pre-training, and it is output that the audio data has abnormal sounds.
- the method for detecting tire abnormality further includes: issuing a tire abnormality alarm to prompt a user that the target tire is abnormal; receiving a feedback result from the user on the tire abnormality alarm; based on the feedback
- the audio template library is updated with the result and the sound feature parameters corresponding to the audio data.
- a possible implementation manner before the acquisition of audio data through the sound-sensitive sensor, further includes: activating the sound-sensitive sensor associated with the target tire.
- an embodiment of the present application provides a device for detecting tire abnormality, which is used to implement the method for detecting tire abnormality provided in the first aspect.
- the sensor periodically collects the tire pressure data of the target tire; the tire pressure analysis module is used to analyze the tire pressure data collected periodically, and identify the continuous decrease of the tire pressure of the target tire; the audio data collection module is used for The audio data is collected by the sound-sensitive sensor; the abnormal sound recognition module is used to identify the abnormal sound of the audio data, and identify that there is abnormal sound in the audio data; the sound source localization module is used to identify the sound source of the audio data.
- Positioning analysis to identify the abnormal sound source; a matching module for matching the pointing position of the abnormal sound source with the target tire, and matching the pointing position of the abnormal sound source with the target tire indicates that the target tire Abnormal.
- the continuous decrease of the tire pressure of the target tire includes: the first tire pressure data is greater than the second tire pressure data, and the difference between the first tire pressure data and the second tire pressure data is subtracted from the first tire pressure data. The value is not greater than the preset value;
- the first tire pressure data is the tire pressure data of the target tire collected in the first cycle
- the second tire pressure data of the target tire collected in the second cycle the first cycle
- the first cycle and the The second cycle is any two adjacent cycles in a plurality of consecutive cycles, and the first cycle is located before the second cycle.
- the tire pressure analysis module is further configured to: analyze the tire pressure data collected periodically, and identify that the target tire has air leakage; the device further includes: a tire air leakage alarm module , which is used to issue an alarm to prompt the user that the target tire is leaking.
- the occurrence of air leakage in the target tire includes: the third tire pressure data is greater than the fourth tire pressure data, and the difference between the third tire pressure data and the fourth tire pressure data is greater than a predetermined value. set value, the third tire pressure data is the tire pressure data of the target tire collected in the third cycle, the fourth tire pressure data of the target tire collected in the fourth cycle, the first cycle and the second period are two adjacent periods, and the third period is located before the fourth period.
- the abnormal sound recognition module is specifically used to: process the audio data to obtain the sound feature parameters corresponding to the audio data; obtain the sound feature parameters corresponding to the audio data and pre-training.
- the audio template library is compared, and there is abnormal sound in the output audio data.
- the device for detecting tire abnormality further includes: a tire abnormality alarm module for issuing a tire abnormality alarm to prompt the user that the target tire is abnormal; an audio template library update module for receiving the The user updates the audio template library based on the feedback result of the tire abnormality alarm and the sound feature parameter corresponding to the feedback result and the audio data.
- the number of the sound-sensitive sensors is greater than 1, and the sound source localization module is specifically configured to: based on the audio data, obtain the audio energy collected by the sound-sensitive sensor; based on the audio energy, The distance of each acoustic sensor relative to the sound source at the unknown position is calculated, wherein the position of the acoustic sensor corresponding to the smallest distance relative to the sound source at the unknown position represents the abnormal sound source.
- the apparatus for detecting tire abnormality further includes: a sound-sensitive sensor activation module, configured to activate the sound-sensitive sensor associated with the target tire.
- an embodiment of the present application provides a control unit for tire abnormality detection, the control unit is configured with programmable instructions, and when the programmable instructions are executed, can implement the tire abnormality detection provided in the first aspect.
- an embodiment of the present application provides a system for detecting tire abnormality, the system including a tire pressure sensor, an acoustic sensor and a control unit.
- Tire pressure sensors are built into each wheel to collect tire pressure data for each wheel.
- Acoustic sensors are placed on the vehicle that can collect audio data at the tires, and are used to collect audio data.
- the control unit is coupled with the tire pressure sensor and the sound-sensitive sensor, and is used for executing the method for detecting tire abnormality provided by the first aspect.
- an embodiment of the present application provides a storage medium, where an instruction is stored in the storage medium, and when the instruction is executed, the method for detecting tire abnormality provided in the first aspect can be implemented.
- the method for detecting tire abnormality provided by the embodiments of the present application, combined with the tire pressure detection, abnormal sound recognition and sound source localization technology, can timely alarm when the tire is abnormal (for example, the tire is inserted into a foreign object, the tire is leaking, etc.), and the driving is improved. security, and the accuracy of recognition can be continuously increased through incremental learning.
- FIG. 1 is a schematic diagram of a tire abnormality detection system provided by an embodiment of the present application.
- FIG. 2 is a schematic diagram of a microphone installation position according to an embodiment of the present application.
- FIG. 3 is a flowchart of a tire abnormality detection method provided by an embodiment of the present application.
- FIG. 4 is a flowchart of a method for identifying an abnormal sound provided by an embodiment of the present application
- FIG. 5 is a schematic diagram of a sound source localization principle provided by an embodiment of the present application.
- FIG. 6 is a schematic diagram of a tire abnormality detection device provided by an embodiment of the present application.
- FIG. 7 is a schematic diagram of another tire abnormality detection device according to an embodiment of the present application.
- the tire abnormality detection system 100 includes: a data acquisition module 101 , a control module 102 , a display module 103 and a feedback module 104 .
- the data acquisition module 101 includes a pressure sensor module 1011 and an acoustic sensor module 1012 .
- the pressure sensor module 1011 includes a built-in pressure sensor on each wheel.
- one pressure sensor is set on one wheel.
- the pressure sensor is set on the hub of each wheel, and is used to collect the tires in the north of the corresponding wheel. pressure data.
- each pressure sensor has a built-in wireless communication component, the pressure sensor can convert the collected tire pressure signal into an analog signal, and then send the collected tire pressure data to the control module 102 through the wireless communication component, and the control module 102 performs the data. Analysis and Exception Judgment.
- the sound-sensitive sensor module 1012 includes one or more sound-sensitive sensors, and the one or more sound-sensitive sensors can be respectively deployed on the inner side of the vehicle door or near the wheels, and are used to collect audio data of the vehicle during driving.
- Acoustic sensors generally refer to devices that can collect audio data, such as microphones and pickups.
- the sound-sensitive sensor module uses four microphones as sound-sensitive sensors, which are respectively deployed outside the vehicle near the wheels for audio data during the driving process of the vehicle. It should be pointed out that the number and deployment positions of the acoustic sensors are not specifically limited.
- the voice recognition microphone in the car can be reused.
- one or more sound-sensitive sensors send the collected audio data to the control module 102 in a wireless manner or a wired manner.
- the control module 102 includes a storage unit 1021 , an abnormal sound recognition unit 1022 , and a sound source localization unit 1023 .
- the storage unit 1021 stores the tire pressure data, audio data and audio database received from the data acquisition module 101 .
- the abnormal sound identification unit 1022 performs abnormal sound analysis on the audio data received from the data acquisition module 101 .
- the sound source localization unit 1023 performs sound source localization analysis on the audio data received from the data acquisition module 101 .
- the control module 102 determines that the tire is abnormal based on the analysis results of the abnormal sound recognition unit 1022 and the sound source localization unit 1023, it sends a signal to the display module 103, and the display module 103 is used to display the abnormal tire to prompt the user that the tire is abnormal. In addition to prompting the user that the tire is abnormal through the display module 103, the user can also be prompted that the tire is abnormal by means of sound, vibration, vision, or the like.
- the audio template library updating module 104 is used to receive feedback information from the user on the display module 103 prompting the tire to be abnormal, for example, the user confirms that the abnormality does exist after checking that the display module 103 prompts the tire to be abnormal, or the user prompts the tire to be abnormal after the display module 103. Negative feedback that there is no abnormality after checking that there is no abnormality.
- the audio template library updating module 104 is further configured to update the audio template library according to the user's feedback, so as to improve the performance of the tire abnormality detection system 100 .
- the tire abnormality detection method 200 may include the following steps:
- the pressure sensor when the vehicle starts, activate each pressure sensor built in the hub of each wheel and collect tire pressure data of the corresponding wheel.
- the pressure sensor periodically collects the tire pressure data of the corresponding wheel, for example, the tire pressure data is collected every preset time (such as 20 seconds, 30 seconds, etc.), the tire pressure data can be collected once, It is also possible to collect multiple data tire pressures.
- the average tire pressure data may be taken as the tire pressure data collected in one cycle.
- S200 compare and analyze the tire pressure data collected by each pressure sensor in the current cycle and the tire pressure data collected in the previous cycle to detect the tire pressure state of the wheel corresponding to each pressure sensor. For example, compare and analyze the tire pressure data of wheel A collected by pressure sensor A in the current cycle and the tire pressure data of wheel A collected by pressure sensor A in the previous cycle, and the tire pressure data of wheel B collected by pressure sensor B in the current cycle The tire pressure data is compared and analyzed with the tire pressure data of wheel B collected by pressure sensor B in the previous cycle.
- each pressure sensor collects the tire pressure data of the corresponding wheel for the first time, in this case, S200 may not be executed, that is, no cycle time
- the collected tire pressure data were compared and analyzed.
- the tire pressure data collected by each pressure sensor in the current cycle can also be used as the tire pressure data collected by the respective pressure sensors in the previous cycle of the first cycle, that is, the tire pressure data collected by each pressure sensor in the current cycle is the same as the tire pressure data collected by each pressure sensor in the current cycle.
- the collected tire pressure data were compared and analyzed.
- the tire pressure data collected by each pressure sensor in the current cycle can also be compared and analyzed with the tire pressure data collected recently by each pressure sensor before the vehicle is started.
- the tire pressure status may include rapid decrease in tire pressure, non-rapid decrease in tire pressure, increase in tire pressure, and normal tire pressure, where rapid decrease in tire pressure indicates that the tire of the wheel has rapid air leakage, and non-rapid decrease in tire pressure indicates that there is no rapid tire leakage in the wheel.
- the S200 also includes:
- step S202 if the tire pressure state is that the tire pressure does not decrease rapidly, it is determined whether the value of the timer reaches a preset value. If the value of the timer does not reach the preset value, the value of the timer is increased by 1, and the process returns to step S100. If the value of the timer reaches the preset value, the value of the timer is reset to zero, and the following step S300 is performed. When the timer value reaches the preset value, the tire is abnormal. The timer is used to record the times of non-rapid reduction of tire pressure. The value of the timer indicates the number of times of non-rapid reduction of tire pressure.
- the continuous increase of the value of the timer indicates that the tire pressure is in a state of continuous reduction.
- the tire pressure status of each tire is that the number of times the tire pressure does not decrease rapidly is counted separately, which may be counted by one timer, or may be counted by multiple timers separately.
- a timer is a device that can count, and the specific structure and composition of the timer are not limited.
- the above-mentioned preset value may be set according to actual requirements, for example, 10, 20, which is used to indicate that the preset tire pressure state is the number of times that the tire pressure does not decrease rapidly and continuously.
- step S203 if the tire pressure state is that the tire pressure is increased, the value of the timer is cleared to zero, and further, the process returns to step S100.
- step S204 if the tire pressure state is normal, clear the value of the timer, and further, return to step S100.
- Step S200 specifically includes:
- the tire pressure state is that the tire pressure decreases rapidly, indicating that the tire of the wheel A has rapid air leakage.
- the tire pressure state is that the tire pressure does not decrease rapidly, which means that although the tire of the wheel A does not leak rapidly, the tire pressure is decreasing.
- the above P is a preset value, which is a positive number and is used to indicate the threshold value of tire air leakage.
- the preset value can be a fixed threshold value (such as 5kpa), or a dynamic threshold value, such as based on ambient temperature, driving environment, weather, tire Threshold for dynamic adjustment of materials, road conditions and other information.
- S300 activate the sound-sensitive sensor, and collect audio data.
- the acoustic sensor For the description of the deployment position of the acoustic sensor, reference may be made to the description of the acoustic sensor module and the acoustic sensor in the corresponding embodiment of FIG. 1 .
- sound-sensitive sensors collect audio data from tires while the vehicle is running.
- the sound-sensitive sensor may periodically collect audio data for a period of time, for example, audio data is collected every minute, and each collection lasts for 10s.
- the number of activated sound-sensitive sensors can be determined according to requirements, preset rules, etc., and can only activate the tire pressure state of the wheel is that the tire pressure does not decrease rapidly and the value of the timer reaches one or a nearby value corresponding to or near the preset value.
- Multiple acoustic sensors all of which can be activated.
- other sound-sensitive sensors such as activating all or nearby sensors
- Audio data is used for abnormal sound recognition.
- step S200 it is detected that the tire state of the wheel corresponding to microphone 1 (the upper left wheel in the figure) is that the tire pressure does not decrease rapidly and the value of the timer reaches the preset value, then in step S300, you can To activate microphone 1, microphone 2, and microphone 3, it is not necessary to activate microphone 4, which is farther from the wheel corresponding to microphone 1, because microphone 1 is relatively far from the wheel corresponding to microphone 4 (the lower right wheel in the figure), and the wheel corresponding to microphone 4 is relatively far away.
- the sound at the microphone 1 has little impact on the audio data collected by the microphone 1, and not activating the microphone 4 can reduce energy consumption and reduce the occupation of computing resources.
- step S400 Perform abnormal sound identification on the collected audio data to identify whether there is an abnormal sound. If there is an abnormal sound, step S500 is executed. If there is no abnormal sound, clear the value of the timer, and return to S100.
- the method for identifying the abnormal sound in the prior art may be adopted, and you may also refer to the relevant description of the following embodiments, which will not be repeated here.
- step S500 use the sound source localization technology to locate the sound source of the abnormal sound, and identify the abnormal tire. Specifically, locate the sound sensitive sensor from which the abnormal sound identified in step S400 comes from, and then determine and locate the sound sensitive sensor. There is an abnormality in the tire of the wheel corresponding to the sensor. Combined with Figure 2, for example, the abnormal sound source and the microphone 1 are extracted, then the tire with abnormality is identified as the tire of the wheel corresponding to the microphone 1 (the upper left wheel in the figure).
- S600 Match the abnormal tire identified based on the sound source localization with the abnormal tire identified based on the tire pressure state.
- a tire abnormality warning is issued to the user, for example, the user is prompted that the tire is abnormal by means of display, sound, vibration and the like.
- the tire with abnormality identified based on sound source localization is A
- the tire with abnormality identified based on the tire pressure status is also A, indicating that the matching is successful, and a tire abnormality warning is issued to the user: the tire A is abnormal, you can use To prompt the user to get off the car to check whether there is any abnormality in the argument.
- S700 Receive feedback information from the user regarding the tire abnormality warning. For example, the user detects the feedback information that the tire is indeed abnormal, or the user detects the feedback information that the tire does not have abnormality.
- a warning is issued only when the monitored tire pressure value exceeds the tire leakage threshold and it is judged that the tire is leaking, and the above-mentioned embodiment provides a tire abnormality detection method, which can be used when the tire does not leak.
- the tire pressure is continuously reduced and the tire abnormal sound monitoring is used to timely identify the abnormality of the tire and issue a warning to ensure the driving safety of the vehicle.
- the embodiments of the present application provide an abnormal sound recognition method, which can be used to implement step S400 in the embodiment corresponding to FIG. 3 .
- the abnormal sound recognition method includes the following steps:
- S301 extracting sound feature parameters of the collected audio data.
- a valid sound segment is separated from the audio data, which is also called endpoint detection, and the starting point and the ending point of the abnormal sound are derived from the audio data.
- endpoint detection sound feature parameters are extracted from valid sound segments.
- Accurate endpoint detection can improve abnormal sound recognition accuracy. Endpoint detection may adopt methods based on short-term energy and short-term zero-crossing rate, short-term amplitude and short-term dynamic threshold rate in the prior art, and the endpoint detection method will not be repeated in this application.
- different feature parameters for reflecting audio data can be selected according to specific requirements, for example, short-term amplitude, Mel Frequency Cepstrum Coefficient (MFCC), MFCC first-order difference coefficient, etc.
- the MFCC is used to characterize the features of the audio data
- the method for calculating the MFCC generally includes: decomposing the audio signal (that is, the audio data or the effective sound segment, the same below) into multiple frames; pre-emphasizing the audio signal, through a high-pass filter; then perform Fourier transform on the audio signal that has passed the high-pass filter, and transform it to the frequency domain; pass the spectrum obtained by each frame through the Mel filter (triangular overlapping window) to obtain the Mel scale; The logarithmic energy is extracted on the Seoul scale; the inverse discrete Fourier transform is performed on the result obtained above, and it is transformed into the cepstral domain.
- the MFCC is the amplitudes of this cepstrogram. Generally, 12 coefficients are used, which are superimposed with the frame energy to obtain 13-dimensional coefficients. Further, the method for extracting audio feature parameters corresponding to audio data in the abnormal sound recognition stage adopts the same feature parameters as the audio feature parameters corresponding to the audio samples extracted in the audio template library stage for abnormal sound recognition.
- Specific content includes:
- the sound feature parameters corresponding to the audio data to be identified are used represents that the maximum posterior probability that the audio data is the j-th type of audio in the audio template library is P( ⁇ j
- the audio data to be recognized has the same probability of each type of sound in the audio template library, that is, And P(X) is constant and the same. Therefore, to find the maximum posterior probability is to find the value of ⁇ j to maximize P(X
- the recognition rate of abnormal sound largely depends on the accuracy of the training audio template library.
- the audio template library can be trained using Hidden Markov Model (HMM), Gaussian Mixture Model (GMM) and other sound models.
- HMM Hidden Markov Model
- GMM Gaussian Mixture Model
- Gaussian distribution combinations of different parameters can be used to represent different audios, that is, the characteristic parameters of each audio correspond to one GMM.
- the audio template library training process includes: a sound feature parameter extraction stage of audio samples and an audio template library training stage.
- the sound characteristic parameter adopts MFCC.
- the specific steps of the sound feature parameter extraction stage of the audio sample include the following:
- Each type of audio sample can have a unified standard, and the audio amplitude is normalized between [-1, 1] to eliminate the difference between different audio samples, that is, divide each sample value by The peak amplitude of the signal in this segment.
- the calculation formula is:
- x(i) is the original audio sample
- n is the audio sample length
- Pre-emphasis can boost high-frequency components and flatten the spectrum of the audio sample, which is convenient for spectrum analysis or channel parameter analysis.
- the transfer function of the pre-emphasis filter z is:
- ⁇ is a constant, usually 0.97.
- Windowing and framing In order to ensure the short-term stability of the audio samples, the Hamming window function is selected for framing processing. A typical window size is 25ms and the frame shift is 10ms. The window function is:
- the length of the window sequence is N, and ⁇ is taken as 0.4.
- FFT Fast Fourier Transform
- the mel-frequency filter formed by the band-pass filter performs the convolution operation, and then takes the logarithm of the output of each frequency band to obtain the logarithmic energy S(m) of each output.
- the discrete cosine transform is performed on the N parameters, and the Mel cepstral coefficients are obtained as audio feature parameters.
- the formula is expressed as:
- n is the number of MFCCs taken;
- C i (n) is the n-th MFCC coefficient of the i-th frame,
- S(m) is the logarithmic power spectrum of the audio sample, and
- M is the number of triangular filters.
- processing method for extracting the sound feature parameters of the audio samples in the training process of the audio template library can be applied to extracting the sound feature parameters of the collected audio data in step S301.
- the GMM can be used in the training phase of the audio template library.
- the specific steps include the following: for the tire abnormal sound recognition scene, the audio template library stores the audio samples of the vehicle in the process of driving under multiple scene road conditions and the audio of the tire pierced into the foreign object. After extracting the sound feature parameters, use GMM for training, and obtain the GMM corresponding to each type of audio sample, which can be referred to by ⁇ j , and finally obtain the ternary formula describing the GMM of each type of audio sample:
- P j is the weight of the mixed component
- ⁇ j is the mean vector
- ⁇ j is the covariance matrix
- N is the mixing order
- j is the sample number
- the triple combination of GMM describing each type of audio sample is the audio template library.
- the embodiments of the present application provide a sound source localization analysis method, which can be used to implement step S500 in the embodiment corresponding to FIG. 5 , including the following steps:
- S502 Acquire the audio energy collected by a certain sound-sensitive sensor (eg, a microphone) at a certain moment according to the audio data with abnormal sound in S501.
- a certain sound-sensitive sensor eg, a microphone
- S503 Calculate the distance between the microphone and the sound source position based on the two-dimensional position coordinates of the corresponding microphone.
- the figure contains n+1 indifference microphones and unknown sound sources.
- S(t) is the energy value of the sound source
- gi is the gain coefficient of the ith microphone
- d i is the distance between the ith microphone and the unknown sound source
- ⁇ i (t) is the superimposed background noise energy value.
- ⁇ i (t) is set to 0
- g i is set to 1
- the energy value S(t) of the sound source can be set to a fixed constant (such as 1000).
- the distance between the ith microphone and the unknown sound source can be calculated according to the energy measurement result E i (t)
- S504 Use the least squares method to identify the sound source of the abnormal sound. Specifically, the position of the sound source of the abnormal sound is most likely to be located at which sound-sensitive sensor (eg, a microphone).
- sound-sensitive sensor eg, a microphone
- the specific implementation is as follows:
- FIG. 3 corresponds to step S600 described in the embodiment.
- the abnormal tires identified based on the sound source localization are matched with the abnormal tires identified based on the tire pressure state.
- the unknown tires found based on the sound source localization method are matched Whether the tire at the position pointed by the position sound source is consistent with the tire with abnormality identified based on the tire pressure status.
- the unknown sound source that is, the abnormal sound source, corresponds to the abnormal sound identified by the abnormal sound.
- the traditional TPMS cannot give an alarm in time when the tire pressure drops.
- the nails are embedded in the tire for a long time, threatening the safety of passengers.
- the tire pressure is too low to drive.
- the audio data of vehicle driving is collected through a microphone.
- the tire pressure data collected by the tire pressure sensor and the audio data collected by the sound sensitive sensor are collected by the GMM-based incremental learning abnormal sound recognition technology and sound source localization technology. By analyzing, it can alarm in time when there is an abnormality in the tire (for example, the tire is stuck in a foreign object), which can improve driving safety, and the accuracy of recognition can be continuously increased through incremental learning.
- the device 1000 for detecting tire abnormality includes: a tire pressure data acquisition module 1010, a tire pressure analysis module 1020, an audio data acquisition module 1030, and an abnormal sound recognition module Module 1040 , sound source localization module 1050 , matching module 1060 , tire leakage alarm module 1070 , tire abnormality alarm module 1080 , audio template library update module 1090 and sound sensitive sensor activation module 1110 .
- the tire pressure data collection module 1010 is used to periodically collect the tire pressure data of the target tire through the pressure sensor; the tire pressure analysis module 1020 is used to analyze the periodically collected tire pressure data and identify the tire pressure of the target tire.
- the audio data collection module 1030 is used to collect audio data through the sound-sensitive sensor; the abnormal sound recognition module 1040 is used to identify the abnormal sound in the collected audio data, and identify the abnormal sound in the audio data; the sound source
- the positioning module 1050 is used to perform sound source localization analysis on the collected audio data to identify the abnormal sound source; the matching module 1060 is used to match the pointing position of the abnormal sound source with the target tire, and the pointing position of the abnormal sound source is the same as that of the target tire.
- a target tire match indicates that the target tire is abnormal.
- the continuous decrease in the tire pressure of the target tire includes: the first tire pressure data is greater than the second tire pressure data, and the difference between the first tire pressure data and the second tire pressure data is not greater than a preset value; wherein, the first tire pressure data
- the data is the tire pressure data of the target tire collected in the first cycle, and the tire pressure data of the target tire collected in the second cycle.
- the first cycle and the second cycle are any two adjacent consecutive cycles. period, and the first period precedes the second period.
- the multiple cycles include at least 3 acquisition cycles.
- the tire pressure analysis module 1020 is further configured to: analyze the tire pressure data collected periodically, and identify that the target tire has air leakage.
- the apparatus 1000 further includes: a tire air leakage alarm module, configured to issue an alarm to prompt the user that the target tire has air leakage.
- the occurrence of air leakage in the target tire includes: the third tire pressure data is greater than the fourth tire pressure data, and the difference between the third tire pressure data and the fourth tire pressure data is greater than a preset value, and the third tire pressure data is in the third cycle.
- the collected tire pressure data of the target tire, the fourth tire pressure data of the target tire collected in the fourth cycle, the first cycle and the second cycle are two adjacent cycles, and the third cycle is located before the fourth cycle.
- the abnormal sound recognition module 1040 is specifically configured to: process the audio data to obtain sound feature parameters corresponding to the audio data; compare the sound feature parameters corresponding to the audio data with the audio template library obtained by pre-training, There is an abnormal sound in the output audio data.
- Apparatus 1000 also includes:
- the tire abnormality alarm module is used to issue a tire abnormality alarm to prompt the user that the target tire is abnormal;
- the audio template library update module 1090 is used to receive the feedback result of the user for the tire abnormality alarm, and based on the Feedback results and sound feature parameters corresponding to the audio data, and update the audio template library.
- the sound source localization module 1050 is specifically used to: based on the audio data, obtain the audio energy collected by the sound sensitive sensors; The distance of the acoustic sensor corresponding to the smallest distance from the unknown sound source indicates an abnormal sound source.
- the apparatus 1000 also includes an acoustic sensor activation module 1110 for activating the acoustic sensor associated with the target tire.
- the embodiment of the present application provides another tire abnormality detection device, referring to FIG. 7 , the tire abnormality detection device 200 can implement the tire abnormality detection method described in the corresponding embodiment of FIG. 3 .
- the tire abnormality detection device 200 includes: a memory 201 , a processor 202 , a communication interface 203 , and a bus 204 .
- the memory 201 , the processor 202 , and the communication interface 203 are connected to each other through the bus 204 for communication.
- the memory 201 may be a read-only memory, a static storage device, a dynamic storage device, or a random access memory.
- the memory 201 may store a program, and when the program stored in the memory 201 is executed by the processor 202, the processor 202 is configured to execute the tire abnormality detection method described in the embodiment of the present application corresponding to FIG. 3 .
- the processor 202 may adopt a general-purpose central processing unit, a microprocessor, an application-specific integrated circuit, a graphics processing unit (graphics processing unit, GPU) or one or more integrated circuits for executing relevant programs to implement the embodiments of the present application.
- the processor can realize the functions of each module in FIG. 6 .
- the processor 202 can also be an integrated circuit chip with signal processing capability. In the implementation process, each step of the image segmentation method of the present application can be completed by an integrated logic circuit of hardware in the processor 202 or instructions in the form of software.
- the above-mentioned processor 202 can also be a general-purpose processor, a digital signal processor (Digital Signal Processing, DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices. , discrete gate or transistor logic devices, discrete hardware components.
- a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
- the steps of the method disclosed in conjunction with the embodiments of the present application may be directly embodied as executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor.
- the software modules may be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other storage media mature in the art.
- the storage medium is located in the memory 201, and the processor 202 reads the information in the memory 201, and in combination with its hardware (such as a display screen), completes the functions required to be performed by the modules included in the tire abnormality detection apparatus 100 of the embodiment of the present application, or executes this The tire abnormality detection method 100 of the application method embodiment.
- the communication interface 203 uses a transceiver such as but not limited to a transceiver to implement communication between the tire abnormality detection device 200 and other devices or a communication network. For example, the data of the actual parking spaces around the vehicle to be parked may be received through the communication interface 203 .
- the bus 204 may include a pathway for communicating information between the various components of the tire anomaly detection device 200 (eg, the memory 201, the processor 202, the communication interface 203).
- the tire abnormality detection device 200 shown in FIG. 7 only shows a memory, a processor, and a communication interface, in the specific implementation process, those skilled in the art should understand that the tire abnormality detection device 200 also includes a Other devices necessary for operation. Meanwhile, according to specific needs, those skilled in the art should understand that the tire abnormality detection apparatus 200 may further include hardware devices that implement other additional functions. In addition, those skilled in the art should understand that the tire abnormality detection apparatus 200 may also only include the necessary components to implement the embodiments of the present application, rather than all the components shown in FIG. 7 .
Abstract
一种检测轮胎异常的方法，可以应用在智能驾驶汽车、新能源汽车、自动驾驶汽车上，该方法包括：通过压力传感器周期性采集目标轮胎的胎压数据；对周期性采集到的胎压数据进行分析，识别出目标轮胎的胎压出现连续降低；通过声敏传感器采集音频数据；对音频数据进行异常音识别，识别出音频数据存在异常音；对音频数据进行声源定位分析，识别出异常声源；将异常声源的指向位置与目标轮胎进行匹配，异常声源的指向位置与目标轮胎匹配表示目标轮胎出现异常。
Description
本申请涉及汽车领域，尤其涉及一种检测轮胎异常的方法。
近年来，我国的汽车保有量已达2.6亿辆，并且这一数字仍在快速增长，汽车安全也越发受到人们重视。在车辆行驶过程中，汽车轮胎故障最令驾驶员担心，也最难预防。据统计在中国每年约30％的交通事故是胎压过低或过高引起轮胎爆胎，有高达50％的高速交通事故是汽车轮胎胎压异常引起的，如何监测预防和检测轮胎异常已经成为汽车安全的重要课题。
应用于汽车上的轮胎压力监测系统(Tire Pressure Monitoring System,TPMS)，能够实时检测轮胎胎压，在检测到胎压异常时能够及时发出警报甚至提前预警，从而保障车辆和乘客的安全。TPMS的基本原理是首先设定车辆轮胎胎压的合理范围，确保设定值与车辆轮胎参数一致，当胎压偏离设定值不在设定的车辆轮胎胎压的合理范围内，TPMS将发出警报。TPMS能够在轮胎胎压太低或者有气体泄露时，及时报警。
目前市面上的TPMS大致可分为两类：间接式TPMS、直接式TPMS。间接式TPMS是指通过测量车辆其他参数，如四轮转速，而非直接测量汽车轮胎压力的系统，但是间接式TPMS无法测量出车辆轮胎的瞬时气压，且灵敏度较低。直接式TPMS是指通过安装在车辆轮胎内部的压力传感器对胎压进行直接测量，并将数据直接发送至接收器上的系统，直接式TMPS监测精度高，无论车辆在何种工况下都能进行监测，且灵敏度高，不会出现误报，但是这种直接式TPMS安装成本较高，电池耗电较快。
现有的TPMS，通常是在轮胎胎压数值低于一定预设值时才会进行报警，无法及时进行报警，提示驾驶员，存在潜在的驾驶安全。
发明内容
针对现有技术中TPMS在轮胎存在异常时无法及时进行报警的问题，本申请实施例提供了一种轮胎异常检测方法和装置。
第一方面，本申请实施例提供了一种检测轮胎异常的方法，可以用于检测轮胎是否存在异常，包括轮胎是否漏气等问题。具体步骤包括：通过车辆车轮上内置的压力传感器周期性地采集目标轮胎的胎压数据；对周期性采集到的胎压数据进行分析，识别出所述目标轮胎的胎压出现连续降低的情况；通过车辆配置的声敏传感器采集音频数据；对所述音频数据进行异常音识别，识别出所述音频数据存在异常音；对所述音频数据进行声源定位分析，识别出异常声源；将所述异常声源的指向位置与所述目标轮胎进行匹配，所述异常声源的指向位置与所述目标轮胎匹配表示所述目标轮胎出现异常。
结合胎压检测、异常音识别和声源定位技术，能够及时检测出车辆的轮胎存在异常并及时发出报警，且相对于传统TPMS仅依据胎压阈值进行判定的技术方案，轮胎 异常检测的准确率更高。
一种可能的实现方式，所述目标轮胎的胎压出现连续降低包括：第一胎压数据大于第二胎压数据，且所述第一胎压数据减去所述第二胎压数据的差值不大于预设值；
其中，所述第一胎压数据为在第一周期采集的所述目标轮胎的胎压数据，所述第二在第二周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是连续的多个周期中任意相邻的两个周期，且所述第一周期位于所述第二周期之前。
一种可能的实现方式，该检测轮胎异常的方法还包括：对周期性采集到的胎压数据进行分析，识别出所述目标轮胎出现漏气；发出报警，以向用户提示所述目标轮胎出现漏气。
一种可能的实现方式，所述目标轮胎出现漏气包括：第三胎压数据大于第四胎压数据，且所述第三胎压数据减去所述第四胎压数据的差值大于预设值，所述第三胎压数据为在第三周期采集的所述目标轮胎的胎压数据，所述第四在第四周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是相邻的两个周期，且所述第三周期位于所述第四周期之前。
一种可能的实现方式，所述对所述音频数据进行异常音识别，识别出所述音频数据存在异常音，包括：对所述音频数据进行处理，获得所述音频数据对应的声音特征参数；将所述音频数据对应的声音特征参数与预训练得到的音频模板库进行比对，输出所述音频数据存在异常音。
一种可能的实现方式，该检测轮胎异常的方法还包括：发出轮胎异常报警，以向用户提示所述目标轮胎出现异常；接收所述用户针对所述轮胎异常报警的反馈结果；基于所述反馈结果和所述音频数据对应的声音特征参数，更新所述音频模板库。
一种可能的实现方式，所述声敏传感器的数量大于1，所述对所述音频数据进行声源定位分析，识别出异常声源，包括：基于所述音频数据，获取所述声敏传感器采集到音频能量；基于所述音频能量，计算每个声敏传感器相对于未知位置声源的距离，其中相对于未知位置声源的距离最小所对应的声敏传感器所处位置表示所述异常声源。
一种可能的实现方式，在所述通过声敏传感器采集音频数据之前，还包括：激活与所述目标轮胎关联的所述声敏传感器。
第二方面，本申请实施例提供了一种检测轮胎异常的装置，用于实现第一方面提供的检测轮胎异常的方法，该检测轮胎异常的装置包括：胎压数据采集模块，用于通过压力传感器周期性采集目标轮胎的胎压数据；胎压分析模块，用于对周期性采集到的胎压数据进行分析，识别出所述目标轮胎的胎压出现连续降低；音频数据采集模块，用于通过声敏传感器采集音频数据；异常音识别模块，用于对所述音频数据进行异常音识别，识别出所述音频数据存在异常音；声源定位模块，用于对所述音频数据进行声源定位分析，识别出异常声源；匹配模块，用于将所述异常声源的指向位置与所述目标轮胎进行匹配，所述异常声源的指向位置与所述目标轮胎匹配表示所述目标轮胎出现异常。
一种可能的实现方式，所述目标轮胎的胎压出现连续降低包括：第一胎压数据大于第二胎压数据，且所述第一胎压数据减去所述第二胎压数据的差值不大于预设值；
其中，所述第一胎压数据为在第一周期采集的所述目标轮胎的胎压数据，所述第 二在第二周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是连续的多个周期中任意相邻的两个周期，且所述第一周期位于所述第二周期之前。
一种可能的实现方式，所述胎压分析模块还用于：对周期性采集到的胎压数据进行分析，识别出所述目标轮胎出现漏气；所述装置还包括：轮胎漏气报警模块，用于发出报警，以向用户提示所述目标轮胎出现漏气。
一种可能的实现方式，所述目标轮胎出现漏气包括：第三胎压数据大于第四胎压数据，且所述第三胎压数据减去所述第四胎压数据的差值大于预设值，所述第三胎压数据为在第三周期采集的所述目标轮胎的胎压数据，所述第四在第四周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是相邻的两个周期，且所述第三周期位于所述第四周期之前。
一种可能的实现方式，所述异常音识别模块具体用于：对所述音频数据进行处理，获得所述音频数据对应的声音特征参数；将所述音频数据对应的声音特征参数与预训练得到的音频模板库进行比对，输出所述音频数据存在异常音。
一种可能的实现方式，该检测轮胎异常的装置还包括：轮胎异常报警模块，用于发出轮胎异常报警，以向用户提示所述目标轮胎出现异常；音频模板库更新模块，用于接收所述用户针对所述轮胎异常报警的反馈结果，并基于所述反馈结果和所述音频数据对应的声音特征参数，更新所述音频模板库。
一种可能的实现方式，所述声敏传感器的数量大于1，所述声源定位模块具体用于：基于所述音频数据，获取所述声敏传感器采集到音频能量；基于所述音频能量，计算每个声敏传感器相对于未知位置声源的距离，其中相对于未知位置声源的距离最小所对应的声敏传感器所处位置表示所述异常声源。
一种可能的实现方式，该检测轮胎异常的装置还包括：声敏传感器激活模块，用于激活与所述目标轮胎关联的所述声敏传感器。
第三方面，本申请实施例提供了一种控制单元，用于轮胎异常检测，该控制单元配置有可编程指令，所述可编程指令被执行时可以实施用于第一方面提供的检测轮胎异常的方法。
第四方面，本申请实施例提供了一种检测轮胎异常的系统，该系统包括胎压传感器、声敏传感器和控制单元。胎压传感器内置在各车轮中，用于采集个车轮的胎压数据。声敏传感器安置在可以采集轮胎处音频数据的车辆上，用于采集音频数据。控制单元与胎压传感器、声敏传感器耦合，用于执行第一方面提供的检测轮胎异常的方法。
第五方面，本申请实施例提供了一种存储介质，该存储介质存储有指令，指令被执行时可以实施第一方面提供的检测轮胎异常的方法。
本申请实施例提供的检测轮胎异常的方法，结合胎压检测、异常音识别和声源定位技术，能够在轮胎存在异常时(例如，轮胎扎入异物、轮胎漏气等)及时报警，提升驾驶安全性，而且通过增量学习方式可以不断增加识别的准确性。
图1为本申请实施例提供的一种轮胎异常检测系统示意图；
图2为本申请实施例提供的麦克风安装位置示意图；
图3为本申请实施例提供的一种轮胎异常检测方法流程图；
图4为本申请实施例提供的一种异常音识别方法流程图；
图5为本申请实施例提供的一种声源定位原理示意图；
图6为本申请实施例提供的一种轮胎异常检测装置示意图；
图7为本申请实施例提供的另一种轮胎异常检测装置示意图。
为了使本申请的目的、技术方案和优点更加清楚，下面将结合附图，对本申请实施例中的技术方案进一步地详细描述。显然，所描述的实施例仅是本申请的部分实施例，而不是全部的实施例。基于本申请中的实施例，本技术领域技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
本申请实施例提供了一种轮胎异常检测系统，如图1所示，轮胎异常检测系统100包括：数据采集模块101、控制模块102、显示模块103和反馈模块104。
数据采集模块101包括压力传感器模组1011和声敏传感器模组1012。具体地，压力传感器模组1011包括内置在各车轮上地压力传感器，一般情况，一个车轮上设置一个压力传感器，具体地，压力传感器设置各车轮的轮毂上，用于采集相应车轮轮胎北部的胎压数据。可选的，各压力传感器内置有无线通信组件，压力传感器可以将采集到的胎压信号转换成模拟信号，再通过无线通信组件向控制模块102发送采集到胎压数据，由控制模块102进行数据分析和异常判断。具体地，声敏传感器模组1012包括一个或多个声敏传感器，一个或多个声敏传感器可以分别部署在车门内侧或车轮附近位置，用于采集车辆在行驶过程中的音频数据。声敏传感器泛指能够进行音频数据采集的器件，例如麦克风、拾音器。如图2所示，声敏传感器模组以4个麦克风作为声敏传感器，分别部署在车外靠近车轮的位置，用于在车辆在行驶过程中的音频数据。需要指出的，声敏传感器数量和部署位置不做具体限定。可选的，可以复用车内的语音识别麦克风。可选的，一个或多个声敏传感器将采集到的音频数据通过无线方式或有线方式发送给控制模块102。
控制模块102包括存储单元1021、异常音识别单元1022、声源定位单元1023。存储单元1021存储从数据采集模块101接收到的胎压数据、音频数据以及音频数据库。异常音识别单元1022将从数据采集模块101接收到的音频数据进行异常音分析。声源定位单元1023对从数据采集模块101接收到的音频数据进行声源定位分析。在控制模块102基于异常音识别单元1022和声源定位单元1023分析结果判断轮胎存在异常时，发送信号给显示模块103，显示模块103用于显示轮胎异常，以提示用户轮胎存在异常。除了通过显示模块103提示用户轮胎存在异常，还可以通过声音、震动、视觉等方式提示用户轮胎存在异常。
音频模板库更新模块104用于接收用户对于显示模块103提示轮胎存在异常的反馈信息，例如用户在显示模块103提示轮胎存在异常后经检查确实存在异常的肯定反馈，或用户在显示模块103提示轮胎存在异常后经检查不存在异常的否定反馈。音频模板库更新模块104还用于根据用户的反馈更新音频模板库，以提高轮胎异常检测系统100的性能。
本申请实施例提供了一种轮胎异常检测方法，如图3所示，轮胎异常检测方法200可以包括以下步骤：
S100，车辆启动时，激活内置于各车轮轮毂上的各压力传感器并采集相应车轮的胎压数据。在完整的轮胎异常检测方法流程中，压力传感器周期性采集相应车轮的胎压数据，例如，每隔预设时间(如20秒、30秒等)采集胎压数据，可以采集一次胎压数据，也可以采集多次数据胎压。当每隔预设时间采集多次胎压数据时，可以取平均胎压数据作为一个周期所采集的胎压数据。
S200，将各压力传感器在当前周期采集到的胎压数据与各自在上一周期采集到的胎压数据进行对比分析，以检测各压力传感器分别对应车轮的胎压状态。例如，压力传感器A在当前周期采集到的车轮A的胎压数据与压力传感器A在上一个周期采集到的车轮A的胎压数据进行对比分析，压力传感器B在当前周期采集到的车轮B的胎压数据与压力传感器B在上一个周期采集到的车轮B的胎压数据进行对比分析。需要说明的，如果当前周期是第一个周期，即车辆启动后，各压力传感器第一次对相应车轮的胎压数据进行采集，在这种情况下，可以不执行S200，即不进行周期间采集到的胎压数据进行对比分析。也可以将各压力传感器在当前周期采集的胎压数据作为各自压力传感器在第一个周期的上一个周期采集的胎压数据，即将各压力传感器在当前周期采集的胎压数据与各自在当前周期采集的胎压数据进行对比分析。也可以将各压力传感器在当前周期采集的胎压数据与各压力传感器在车辆启动之前最近采集的胎压数据进行对比分析。
胎压状态可以包括胎压快速降低、胎压非快速降低、胎压增高和胎压正常，其中胎压快速降低表示车轮的轮胎存在快速漏气，胎压非快速降低表示车轮不存在轮胎快速漏气但是胎压出现降低，胎压增高表示车轮的胎压发生上升，胎压正常表示车轮的胎压未发生变化。
可选的，S200还包括：
S201，如果胎压状态是胎压快速降低，表示轮胎出现漏气，则通过显示、声音、振动等方式报警，提示用户轮胎出现漏气。可选的，提示用户存在轮胎漏气后结束胎压异常检测方法流程。
S202，如果胎压状态是胎压非快速降低，则判断计时器的数值是否达到预设数值。如果计时器的数值没有达到预设数值，则计时器的数值增加1，并返回执行步骤S100。如果计时器的数值达到预设数值，则将计时器的数值清零，并执行下面步骤S300。计时器的数值达到预设数值表示轮胎存在异常。计时器用于记录出现胎压非快速降低的次数，计时器的数值表示存现胎压非快速降低的次数，也就是说，计时器的数值连续增加则表示胎压处于连续降低状态。需要说明的，各轮胎的胎压状态是胎压非快速降低的次数是分开计数的，可以是一个计时器计数，也可以是多个计时器分别计数。计时器是一种可以记数的装置，计时器的具体结构和组成不作限定。可选的，上述预设数值可以根据实际需求进行设置，例如10，20，用于表示预设的胎压状态是胎压非快速降低连续出现的次数。
S203，如果胎压状态是胎压增高，则将计时器的数值清零，进一步的，返回执行步骤S100。
S204，如果胎压状态是胎压正常，则将计时器的数值清零，进一步的，返回执行步骤S100。
一个示例，假设压力传感器A内置在车轮A的轮毂上，在当前周期n，压力传感器A采集到的车轮A的轮胎数据为P
n，单位为千帕(Kpa)。在上一周期n-1，压力传感器A采集到的车轮A的轮胎数据为P
n-1。步骤S200具体包括：
若P
n-1-P
n＞P，则胎压状态是胎压快速降低，表示车轮A的轮胎存在快速漏气。
若0＜P
n-1-P
n≤P，则胎压状态是胎压非快速降低，表示车轮A的轮胎虽然不存在快速漏，但是胎压在降低。
若P
n-1-P
n＜0，则胎压状态是胎压增高，表示车轮A的胎压上升。
若P
n-1-P
n＝0，则胎压状态是胎压正常，表示车轮A的胎压没有出现变化。需要注意的，在实际应用中，胎压一般都会出现波动，不会出现P
n-P
n-1＝0，仅是本申请的示例。其实P
n与P
n-1差值不大，例如在-0.5到0.5之间，都可以判断胎压状态是胎压正常。同样的，-0.5＜P
n-1-P
n＜0.5，也就不能判断为胎压状态是胎压非快速降低或者胎压增高。
上述P为预设值，是正数，用于表示轮胎漏气的阈值，预设值可以为某一固定阈值(如5kpa)，也可以是动态阈值，例如基于环境温度、行驶环境、天气、轮胎材质、路况等信息动态调整的阈值。
S300，激活声敏传感器，并采集音频数据。关于声敏传感器部署位置的描述可以参照图1对应实施例中关于声敏传感器模组和声敏传感器的描述。一般而言，声敏传感器是在车辆行驶的过程中针对轮胎进行音频数据采集。可选的，声敏传感器可以是周期性采集一段时间内音频数据，例如每分钟采集一次音频数据，每次采集持续10s。可选的，激活声敏传感器的数量可以根据需求、预设规则等确定，可以仅激活车轮的胎压状态是胎压非快速降低且计时器的数值达到预设数值对应的或附近的一个或多个声敏传感器，可以激活全部声敏传感器。除了激活车轮的胎压状态是胎压非快速降低且计时器的数值达到预设数值对应的声敏传感器外，还激活其他声敏传感器(例如激活全部或者附近传感器)，是为了更全面的采集音频数据用于异常音识别。结合图2所示，例如，在S200步骤中检测到麦克风1对应的车轮(图中左上车轮)的轮胎状态是胎压非快速降低且计时器的数值达到预设数值，那么在S300步骤中可以激活麦克风1,麦克风2,麦克风3，不需要激活距离麦克风1对应的车轮较远的麦克风4，因为麦克风1距离麦克风4对应的车轮(图中右下车轮)相对较远，麦克风4对应的车轮处的声音对麦克风1采集的音频数据影响较小，不激活麦克风4可以减少能耗，降低计算资源的占用。
S400，对采集到的音频数据进行异常音识别，以识别是否存在异常音。若存在异常音，则执行步骤S500。若不存在异常音，将计时器的数值清零，并返回执行S100。关于如何对音频数据进行异常音识别，可以采用现有技术中的异常音识别方法，也可以参阅下面实施例相关描述，此处不再赘述。
S500，采用声源定位技术对异常音进行声源定位，识别出存在异常的轮胎，具体的，对S400步骤识别出的异常音来源于哪个声敏传感器进行定位，进而确定与定位到的声敏传感器所对应车轮的轮胎存在异常，结合图2，例如是逼出异常音来源与麦克风1，那么识别出存在异常的轮胎是麦克风1对应的车轮(图中左上车轮)的轮胎。
S600，将基于声源定位识别出的存在异常的轮胎与基于胎压状态识别出的存在异 常的轮胎进行匹配。
S601，如果匹配成功，则向用户发出轮胎异常警告，例如通过显示、声音、振动等方式提示用户轮胎存在异常。
一个示例，基于声源定位识别出的存在异常的轮胎是A，基于胎压状态识别出的存在异常的轮胎也是A，则表示匹配成功，向用户发出轮胎异常警告：轮胎A存在异常，可以用于提示用户下车检测论调是否存在异常。
S602，如果匹配不成功，即基于声源定位识别出的存在异常的轮胎与基于胎压状态识别出的存在异常的轮胎不是同一个轮胎，关闭声敏传感器。进一步，将计时器的数值清零。
S700，接收用户针对轮胎异常警告的反馈信息。例如，用户检测轮胎确实存在异常的反馈的信息，或者用户检测轮胎不存在异常的反馈的信息。
S800，基于用户的反馈信息和反馈信息所针对的轮胎异常警告对应的音频数据，更新音频数据库，以实现通过增量学习方式提升异常音识别算法的准确率。
现有技术是在监测到的胎压值超出轮胎漏气的阈值，判断轮胎存在漏气时才发出警告，而上述实施例提供的一种轮胎异常检测方法，能够在轮胎未出现漏气的情况下，通过胎压连续降低和轮胎异常音监测及时识别出轮胎存在异常并发出警告，保证车辆驾驶安全。
结合上面本申请实施例，本申请实施例提供了一种异常音识别方法，可以用于实现图3对应实施例中步骤S400。如图4所示，异常音识别方法包括以下步骤：
S301,提取采集到的音频数据的声音特征参数。可选的，在提取音频数据的声音特征参数提取之前，从音频数据中分离出有效声音段，也称之为端点检测，从音频数据中出道异常声的起始点和截止点。端点检测之后，再是从有效声音段中提取声音特征参数。准确的端点检测可以提高异常音识别准确率。端点检测可以采用现有技术中的基于短时能量和短时过零率、短时幅度和短时动态门限率等方法，本申请关于端点检测方法不再赘述。
声音特征参数可以根据具体需求选择不同用于反应音频数据的特征参数，例如，短时幅度、梅尔频率倒谱系数(Mel Frequency Cepstrum Coefficient,MFCC)、MFCC一阶差分系数等。可选的，采用MFCC表征音频数据的特征，计算MFCC的方法大致包括：将音频信号(即音频数据或者有效声音段，下面同样)分解为多个讯框；将音频信号预强化，通过一个高通滤波器；再在将经过高通滤波器的音频信号进行傅立叶变换，变换至频域；将每个讯框获得的频谱通过梅尔滤波器(三角重叠窗口)，得到梅尔刻度；在每个梅尔刻度上提取对数能量；对上面获得的结果进行离散傅里叶反变换，变换到倒频谱域。MFCC就是这个倒频谱图的幅度(amplitudes)。一般使用12个系数，与讯框能量叠加得13维的系数。进一步的，异常音识别阶段中提取音频数据对应音频特征参数的方法，与训练用于异常音识别的音频模板库阶段中提取音频样本对应的音频特征参数采用相同的特征参数。
S302，将提取获得的音频数据对应的声音特征参数与音频模板库做比对，并输出识别结果。具体内容包括：
待识别的音频数据对应的声音特征参数用
表示，该音频数据 是音频模板库中第j类音频的最大后验概率为P(λ
j|X)。根据贝叶斯准则，最大后验概率可表示为：
假定待识别的音频数据是音频模板库中每类声音的概率相等，即
且P(X)常量且相同。因此求最大后验概率即找到λ
j的值使P(X|λ
j)最大。由于一条音频的似然概率由每一帧的似然概率相乘得到，因此
其对数形式为
其识别结果为
也就是待识别的音频数据是第j类音频的概率最大，如果在音频模板库中，第j类音频是异常音，则输出待识别的音频数据是异常音。
异常音的识别率很大程度依赖于训练得到的音频模板库的准确程度。音频模板库可以采用隐马尔科夫模型(Hidden Markov Model,HMM)、高斯混合模型(Gaussian Mixture Model,GMM)等声音模型进行训练。可选的，采用GMM进行音频模板库训练，不同参数的高斯分布组合可以用来表征不同的音频，即每种音频的特征参数对应一个GMM。
音频模板库训练过程包括：音频样本的声音特征参数提取阶段和音频模板库训练阶段。在本申请的实例中，声音特征参数采用MFCC。音频样本的声音特征参数提取阶段具体步骤包括以下：
1)归一化：可以使每类音频样本有统一的标准，音频幅值归一在[-1,1]之间，消除不同音频样本之间的差异，也就是把每一个采样值除以本段信号的幅度峰值。计算公式为：
其中x(i)是原始音频样本，
是归一化后的音频样本，n是音频样本长度。
2)预加重：预加重可以提升高频成分，使音频样本的频谱变得平坦，以便于进行频谱分析或声道参数分析。预加重滤波器z的传递函数为：
H(z)＝1-μ·z
-1，
其中，μ为常数，通常取0.97。
3)加窗分帧：为了保证音频样本的短时平稳性，选择汉明窗函数进行分帧处理。典型的窗口大小是25ms，帧移是10ms。窗函数为：
w(n)＝(1-α)-α·cos(2πn/(N-1)),0≤n≤N-1，
其中，窗口序列长度为N，α取0.4。
4)MFCC提取：取帧长N＝256点，对每一帧作快速傅里叶变换(Fast Fourier Transformation,FFT)变换求出频谱参数，再将每帧数据的频谱参数通过一组N个三角 形带通滤波器构成的梅尔频率滤波器做卷积运算，之后对每个频带的输出取对数，求出每个输出的对数能量S(m)。最后对此N个参数进行离散余弦变换，求出梅尔倒谱系数作为音频特征参数。公式表达为：
其中，n为所取的MFCC个数；C
i(n)为第i帧的第n个MFCC系数，S(m)为音频样本的对数功率谱，M为三角滤波器个数。
值得注意的，音频模板库训练过程中关于音频样本的声音特征参数提取的处理方法可以适用于步骤S301中的提取采集到的音频数据的声音特征参数。
音频模板库训练阶段可以采用GMM，具体步骤包括以下：针对轮胎异常音识别场景，音频模板库中存入车辆在多个场景路况下行驶过程中的音频样本及车胎扎入异物行驶过程中的音频样本，进行声音特征参数提取后采用GMM进行训练，获得每类音频样本分别对应的GMM，可用λ
j来指代，最终得到描述每类音频样本的GMM的三元式：
λ
j＝{P
j,μ
j,Σ
j},j＝1,2,…,N，
其中，P
j为混合分量的权值，μ
j为均值矢量，Σ
j为协方差矩阵，N为混合阶数，j为样本序号，描述每类音频样本的GMM的三元式组合为音频模板库。
结合上面本申请实施例，本申请实施例提供了一种声源定位分析方法，可以用于实现图5对应实施例中步骤S500，包括以下步骤：
S501，将识别出存在异常音的音频数据作为输入。
S502：根据S501中的存在异常音的音频数据，获取到某时刻某声敏传感器(例如麦克风)采集到的音频能量。
S503：基于对应麦克风的二维位置坐标，计算麦克风与声源位置间的距离。
结合图5，可能的具体实现如下：
图中包含n+1个无差异的麦克风和未知位置声源，第i个麦克风在t时刻接收到能量的表达式为：
S(t)是声源的能量值，g
i是第i个麦克风的增益系数，d
i是第i个麦克风与未知位置声源之间的距离，ε
i(t)为叠加的背景噪声能量值。为简化计算，ε
i(t)为0，g
i设为1，声源的能量值S(t)可设为某固定常量(如1000)。可根据能量测量结果E
i(t)计算第i个麦克风到未知位置声源间的距离
S504：采用最小二乘法，识别出异常音的声源。具体的，异常音的声源指向的位置在哪个声敏传感器(例如麦克风)处的概率最大。
可选的，以声敏传感器为麦克风为例，具体实现：
设第i个麦克风的二维位置坐标为r
i＝(x
i,y
i),i＝1,2,…n，各个麦克风i与未知位置声源之间的距离为
可得公式：
其中1≤i≤n。由最小二乘法可得：
上述方程可表示为：
即
其中
为估计未知位置声源的指向位置，再根据
的值，求其与四个麦克风的最小距离，取其中的最小值，可得到t时刻未知位置声源指向的某麦克风i。
S505：根据上一步骤的估计结果，识别出存在异常的轮胎。
进一步的，基于S504中得到的t时刻未知位置声源指向的某麦克风i，继续计算其它时刻未知位置声源指向的麦克风，最后计算出整个音频数据中未知位置声源指向哪个麦克风的概率最大，再基于麦克风与车轮的位置关系，找到与概率最大的麦克风对应的车轮，进而可以定位到存在异常的轮胎。声源定位就是找到音频数据未知位置声源所指向的位置，以此识别出存在异常的轮胎。图3对应实施例所描述的步骤S600，将基于声源定位识别出的存在异常的轮胎与基于胎压状态识别出的存在异常的轮胎进行匹配，实际上就是将基于声源定位方法找到的未知位置声源所指向的位置所在的轮胎是否与基于胎压状态识别出的存在异常的轮胎一致。需要注意的是，未知位置声源也就是异常声源，是与异常音识别出的异常音相对应的。
传统TPMS无法在胎压发生降低时及时进行报警，钉子长期嵌在轮胎中，威胁着乘客的生命安全，且往往发生报警时，胎压已经过低无法行驶。针对以上问题，通过麦克风采集车辆行驶音频数据，本申请实施例通过基于GMM的增量学习式异常音识别技术和声源定位技术对胎压传感器采集的胎压数据和声敏传感器采集的音频数据进行分析，能够在轮胎存在异常时(例如，轮胎扎入异物)及时报警，提升驾驶安全性，而且通过增量学习方式可以不断增加识别的准确率。
本申请实施例提供了一种检测轮胎异常的装置，如图6所示，检测轮胎异常的装置1000包括：胎压数据采集模块1010、胎压分析模块1020、音频数据采集模块1030、异常音识别模块1040、声源定位模块1050、匹配模块1060、轮胎漏气报警模块1070、轮胎异常报警模块1080、音频模板库更新模块1090和声敏传感器激活模块1110。
胎压数据采集模块1010，用于通过压力传感器周期性采集目标轮胎的胎压数据；胎压分析模块1020，用于对周期性采集到的胎压数据进行分析，识别出所述目标轮胎的胎压出现连续降低；音频数据采集模块1030，用于通过声敏传感器采集音频数据；异常音识别模块1040，用于对采集到的音频数据进行异常音识别，识别出音频数据存在异常音；声源定位模块1050，用于对采集到的音频数据进行声源定位分析，识别出 异常声源；匹配模块1060，用于将异常声源的指向位置与目标轮胎进行匹配，异常声源的指向位置与目标轮胎匹配表示目标轮胎出现异常。目标轮胎的胎压出现连续降低包括：第一胎压数据大于第二胎压数据，且第一胎压数据减去第二胎压数据的差值不大于预设值；其中，第一胎压数据为在第一周期采集的目标轮胎的胎压数据，第二在第二周期采集的目标轮胎的胎压数据，第一周期和第二周期是连续的多个周期中任意相邻的两个周期，且第一周期位于第二周期之前。多个周期包括至少3个采集周期。
进一步的，胎压分析模块1020还用于：对周期性采集到的胎压数据进行分析，识别出目标轮胎出现漏气。
可选的，装置1000还包括：轮胎漏气报警模块，用于发出报警，以向用户提示所述目标轮胎出现漏气。目标轮胎出现漏气包括：第三胎压数据大于第四胎压数据，且第三胎压数据减去第四胎压数据的差值大于预设值，第三胎压数据为在第三周期采集的目标轮胎的胎压数据，第四在第四周期采集的目标轮胎的胎压数据，第一周期和第二周期是相邻的两个周期，且第三周期位于第四周期之前。
异常音识别模块1040具体用于：对所述音频数据进行处理，获得所述音频数据对应的声音特征参数；将所述音频数据对应的声音特征参数与预训练得到的音频模板库进行比对，输出所述音频数据存在异常音。
装置1000还包括：
轮胎异常报警模块，用于发出轮胎异常报警，以向用户提示所述目标轮胎出现异常；音频模板库更新模块1090，用于接收所述用户针对所述轮胎异常报警的反馈结果，并基于所述反馈结果和所述音频数据对应的声音特征参数，更新所述音频模板库。
可选的，声敏传感器的数量大于1，声源定位模块1050具体用于：基于音频数据，获取声敏传感器采集到音频能量；基于音频能量，计算每个声敏传感器相对于未知位置声源的距离，其中相对于未知位置声源的距离最小所对应的声敏传感器所处位置表示异常声源。
装置1000还包括声敏传感器激活模块1110，用于激活与所述目标轮胎关联的所述声敏传感器。
本申请实施例提供了另一种轮胎异常检测装置，参见图7，该轮胎异常检测装置200可以实现图3对应实施例描述的轮胎异常检测方法。轮胎异常检测装置200包括：存储器201、处理器202、通信接口203以及总线204。其中，存储器201、处理器202、通信接口203通过总线204实现彼此之间的通信连接。
存储器201可以是只读存储器，静态存储设备，动态存储设备或者随机存取存储器。存储器201可以存储程序，当存储器201中存储的程序被处理器202执行时，处理器202用于执行图3对应的本申请实施例中描述的轮胎异常检测方法。
处理器202可以采用通用的中央处理器，微处理器，应用专用集成电路，图形处理器(graphics processing unit,GPU)或者一个或多个集成电路，用于执行相关程序，以实现本申请实施例的信令分析装置中的单元所需执行的功能，或者执行本申请方法实施例的图像分割方法。处理器可实现图6中各模块的功能。
处理器202还可以是一种集成电路芯片，具有信号的处理能力。在实现过程中，本申请的图像分割方法的各个步骤可以通过处理器202中的硬件的集成逻辑电路或者 软件形式的指令完成。上述的处理器202还可以是通用处理器、数字信号处理器(Digital Signal Processing,DSP)、专用集成电路(ASIC)、现成可编程门阵列(Field Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器201，处理器202读取存储器201中的信息，结合其硬件(例如显示屏)完成本申请实施例的轮胎异常检测装置100中包括的模块所需执行的功能，或者执行本申请方法实施例的轮胎异常检测方法100。
通信接口203使用例如但不限于收发器一类的收发装置，来实现轮胎异常检测装置200与其他设备或通信网络之间的通信。例如，可以通过通信接口203接收待泊车车辆周边实际停车位的数据。
总线204可包括在轮胎异常检测装置200各个部件(例如，存储器201、处理器202、通信接口203)之间传送信息的通路。
应注意，尽管图7所示的轮胎异常检测装置200仅仅示出了存储器、处理器、通信接口，但是在具体实现过程中，本领域的技术人员应当理解，轮胎异常检测装置200还包括实现正常运行所必须的其他器件。同时，根据具体需要，本领域的技术人员应当理解，轮胎异常检测装置200还可包括实现其他附加功能的硬件器件。此外，本领域的技术人员应当理解，轮胎异常检测装置200也可仅仅包括实现本申请实施例所必须的器件，而不必包括图7所示的全部器件。
以上所述，仅为本发明的具体实施方式，但本发明的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本发明的保护范围之内。因此，本发明的保护范围应以所述权利要求的保护范围为准。
Claims (17)
- 一种检测轮胎异常的方法，其特征在于，包括：通过压力传感器周期性采集目标轮胎的胎压数据；对周期性采集到的胎压数据进行分析，识别出所述目标轮胎的胎压出现连续降低；通过声敏传感器采集音频数据；对所述音频数据进行异常音识别，识别出所述音频数据存在异常音；对所述音频数据进行声源定位分析，识别出异常声源；将所述异常声源的指向位置与所述目标轮胎进行匹配，所述异常声源的指向位置与所述目标轮胎匹配表示所述目标轮胎出现异常。
- 根据权利要求1所述的方法，其特征在于，所述目标轮胎的胎压出现连续降低包括：第一胎压数据大于第二胎压数据，且所述第一胎压数据减去所述第二胎压数据的差值不大于预设值；其中，所述第一胎压数据为在第一周期采集的所述目标轮胎的胎压数据，所述第二胎压数据为在第二周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是连续的多个周期中任意相邻的两个周期，且所述第一周期位于所述第二周期之前。
- 根据权利要求1或2所述的方法，其特征在于，还包括：对周期性采集到的胎压数据进行分析，识别出所述目标轮胎出现漏气；发出报警，以向用户提示所述目标轮胎出现漏气。
- 根据权利要求3所述的方法，其特征在于，所述目标轮胎出现漏气包括：第三胎压数据大于第四胎压数据，且所述第三胎压数据减去所述第四胎压数据的差值大于预设值，所述第三胎压数据为在第三周期采集的所述目标轮胎的胎压数据，所述第四胎压数据为在第四周期采集的所述目标轮胎的胎压数据，所述第三周期和所述第四周期是相邻的两个周期，且所述第三周期位于所述第四周期之前。
- 根据权利要求1-4中任一项所述的方法，其特征在于，所述对所述音频数据进行异常音识别，识别出所述音频数据存在异常音，包括：对所述音频数据进行处理，获得所述音频数据对应的声音特征参数；将所述音频数据对应的声音特征参数与预训练得到的音频模板库进行比对，输出所述音频数据存在异常音。
- 根据权利要求5所述的方法，其特征在于，还包括：发出轮胎异常报警，以向用户提示所述目标轮胎出现异常；接收所述用户针对所述轮胎异常报警的反馈结果；基于所述反馈结果和所述音频数据对应的声音特征参数，更新所述音频模板库。
- 根据权利要求1-6中任一项所述的方法，其特征在于，所述声敏传感器的数量大于1，所述对所述音频数据进行声源定位分析，识别出异常声源，包括：基于所述音频数据，获取所述声敏传感器采集到音频能量；基于所述音频能量，计算每个声敏传感器相对于未知位置声源的距离，其中相对于未知位置声源的距离最小所对应的声敏传感器所处位置表示所述异常声源。
- 根据权利要求1-7中任一项所述的方法，其特征在于，在所述通过声敏传感器 采集音频数据之前，还包括：激活与所述目标轮胎关联的所述声敏传感器。
- 一种检测轮胎异常的装置，其特征在于，包括：胎压数据采集模块，用于通过压力传感器周期性采集目标轮胎的胎压数据；胎压分析模块，用于对周期性采集到的胎压数据进行分析，识别出所述目标轮胎的胎压出现连续降低；音频数据采集模块，用于通过声敏传感器采集音频数据；异常音识别模块，用于对所述音频数据进行异常音识别，识别出所述音频数据存在异常音；声源定位模块，用于对所述音频数据进行声源定位分析，识别出异常声源；匹配模块，用于将所述异常声源的指向位置与所述目标轮胎进行匹配，所述异常声源的指向位置与所述目标轮胎匹配表示所述目标轮胎出现异常。
- 根据权利要求9所述的装置，其特征在于，所述目标轮胎的胎压出现连续降低包括：第一胎压数据大于第二胎压数据，且所述第一胎压数据减去所述第二胎压数据的差值不大于预设值；其中，所述第一胎压数据为在第一周期采集的所述目标轮胎的胎压数据，所述第二胎压数据为在第二周期采集的所述目标轮胎的胎压数据，所述第一周期和所述第二周期是连续的多个周期中任意相邻的两个周期，且所述第一周期位于所述第二周期之前。
- 根据权利要求9或10所述的装置，其特征在于，所述胎压分析模块还用于：对周期性采集到的胎压数据进行分析，识别出所述目标轮胎出现漏气；所述装置还包括：轮胎漏气报警模块，用于发出报警，以向用户提示所述目标轮胎出现漏气。
- 根据权利要求11所述的装置，其特征在于，所述目标轮胎出现漏气包括：第三胎压数据大于第四胎压数据，且所述第三胎压数据减去所述第四胎压数据的差值大于预设值，所述第三胎压数据为在第三周期采集的所述目标轮胎的胎压数据，所述第四胎压数据为在第四周期采集的所述目标轮胎的胎压数据，所述第三周期和所述第四周期是相邻的两个周期，且所述第三周期位于所述第四周期之前。
- 根据权利要求9-12中任一项所述的装置，其特征在于，所述异常音识别模块具体用于：对所述音频数据进行处理，获得所述音频数据对应的声音特征参数；将所述音频数据对应的声音特征参数与预训练得到的音频模板库进行比对，输出所述音频数据存在异常音。
- 根据权利要求13所述的装置，其特征在于，还包括：轮胎异常报警模块，用于发出轮胎异常报警，以向用户提示所述目标轮胎出现异常；音频模板库更新模块，用于接收所述用户针对所述轮胎异常报警的反馈结果，并基于所述反馈结果和所述音频数据对应的声音特征参数，更新所述音频模板库。
- 根据权利要求9-14中任一项所述的装置，其特征在于，所述声敏传感器的数 量大于1，所述声源定位模块具体用于：基于所述音频数据，获取所述声敏传感器采集到音频能量；基于所述音频能量，计算每个声敏传感器相对于未知位置声源的距离，其中相对于未知位置声源的距离最小所对应的声敏传感器所处位置表示所述异常声源。
- 根据权利要求9-15中任一项所述的装置，其特征在于，还包括：声敏传感器激活模块，用于激活与所述目标轮胎关联的所述声敏传感器。
- 一种控制单元，其特征在于，包括可编程指令，所述可编程指令被调用时用于执行权利要求1-8中任一项所述的方法。
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202180000477.8A CN115052761B (zh) | 2021-01-13 | 2021-01-13 | 一种检测轮胎异常的方法和装置 |
| PCT/CN2021/071370 WO2022150999A1 (zh) | 2021-01-13 | 2021-01-13 | 一种检测轮胎异常的方法和装置 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2021/071370 WO2022150999A1 (zh) | 2021-01-13 | 2021-01-13 | 一种检测轮胎异常的方法和装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2022150999A1 true WO2022150999A1 (zh) | 2022-07-21 |
Family
ID=82446479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2021/071370 WO2022150999A1 (zh) | 2021-01-13 | 2021-01-13 | 一种检测轮胎异常的方法和装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN115052761B (zh) |
| WO (1) | WO2022150999A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116051897A (zh) * | 2023-01-28 | 2023-05-02 | 合肥中科类脑智能技术有限公司 | 门开关状态的检测方法、装置及边缘处理设备、存储介质 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09145501A (ja) * | 1995-11-20 | 1997-06-06 | Yokohama Rubber Co Ltd:The | タイヤ嵌合圧測定方法及び装置 |
| CN101973192A (zh) * | 2010-10-03 | 2011-02-16 | 燕山大学 | 汽车转弯行驶中的轮胎气压监测方法 |
| CN102717671A (zh) * | 2012-07-10 | 2012-10-10 | 江苏科技大学 | 一种汽车爆胎检测应急方法 |
| CN203580528U (zh) * | 2013-11-29 | 2014-05-07 | 山东永泰化工有限公司 | 轮胎异常监测装置 |
| CN104908645A (zh) * | 2015-06-08 | 2015-09-16 | 上海斐讯数据通信技术有限公司 | 一种行车辅助驾驶系统及其工作方法 |
| CN107499069A (zh) * | 2017-09-08 | 2017-12-22 | 江苏金坛绿能新能源科技有限公司 | 车辆轮胎爆胎预警装置 |
| CN107933212A (zh) * | 2017-10-23 | 2018-04-20 | 深圳市招科智控科技有限公司 | 港口集装箱水平搬运无人车车胎胎爆检测系统及方法 |
| CN110789278A (zh) * | 2018-08-02 | 2020-02-14 | 郑州宇通客车股份有限公司 | 一种胎压监测方法及装置 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012162241A2 (en) * | 2011-05-20 | 2012-11-29 | Northeastern University | Real-time wireless dynamic tire pressure sensor and energy harvesting system |
| KR102303225B1 (ko) * | 2014-12-03 | 2021-09-16 | 현대모비스 주식회사 | 휠 얼라인먼트 이상 검출방법 및 장치 |
| TWI593571B (zh) * | 2016-05-06 | 2017-08-01 | Mobiletron Electronics Co Ltd | Tire monitoring methods |
| CN107458383B (zh) * | 2016-06-03 | 2020-07-10 | 法拉第未来公司 | 使用音频信号的车辆故障自动检测 |
| CN109017166B (zh) * | 2018-08-16 | 2020-09-08 | 杭州容大智造科技有限公司 | 一种利用声音检测轮胎气压的设备 |
| CN111890850A (zh) * | 2020-09-10 | 2020-11-06 | 李光明 | 防止汽车爆胎的胎噪监测和报警系统及方法 |
-
2021
- 2021-01-13 CN CN202180000477.8A patent/CN115052761B/zh active Active
- 2021-01-13 WO PCT/CN2021/071370 patent/WO2022150999A1/zh active Application Filing
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09145501A (ja) * | 1995-11-20 | 1997-06-06 | Yokohama Rubber Co Ltd:The | タイヤ嵌合圧測定方法及び装置 |
| CN101973192A (zh) * | 2010-10-03 | 2011-02-16 | 燕山大学 | 汽车转弯行驶中的轮胎气压监测方法 |
| CN102717671A (zh) * | 2012-07-10 | 2012-10-10 | 江苏科技大学 | 一种汽车爆胎检测应急方法 |
| CN203580528U (zh) * | 2013-11-29 | 2014-05-07 | 山东永泰化工有限公司 | 轮胎异常监测装置 |
| CN104908645A (zh) * | 2015-06-08 | 2015-09-16 | 上海斐讯数据通信技术有限公司 | 一种行车辅助驾驶系统及其工作方法 |
| CN107499069A (zh) * | 2017-09-08 | 2017-12-22 | 江苏金坛绿能新能源科技有限公司 | 车辆轮胎爆胎预警装置 |
| CN107933212A (zh) * | 2017-10-23 | 2018-04-20 | 深圳市招科智控科技有限公司 | 港口集装箱水平搬运无人车车胎胎爆检测系统及方法 |
| CN110789278A (zh) * | 2018-08-02 | 2020-02-14 | 郑州宇通客车股份有限公司 | 一种胎压监测方法及装置 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116051897A (zh) * | 2023-01-28 | 2023-05-02 | 合肥中科类脑智能技术有限公司 | 门开关状态的检测方法、装置及边缘处理设备、存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115052761A (zh) | 2022-09-13 |
| CN115052761B (zh) | 2023-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5837341B2 (ja) | 路面状態判定方法とその装置 | |
| CN107985225B (zh) | 提供声音追踪信息的方法、声音追踪设备及具有其的车辆 | |
| WO2020220439A9 (zh) | 基于深度神经网络的高速公路交通流量状态识别方法 | |
| CN109632086A (zh) | 车内噪音检测方法、装置、可读存储介质及计算机设备 | |
| US9733346B1 (en) | Method for providing sound detection information, apparatus detecting sound around vehicle, and vehicle including the same | |
| JP2017505430A (ja) | 道路及びタイヤの状態を音響検出する方法 | |
| CN109949823B (zh) | 一种基于dwpt-mfcc与gmm的车内异响识别方法 | |
| JP5097470B2 (ja) | 路面状況判定方法及びその装置 | |
| US20150117652A1 (en) | Sound source detection device, noise model generation device, noise reduction device, sound source direction estimation device, approaching vehicle detection device and noise reduction method | |
| US9934793B2 (en) | Method for determining alcohol consumption, and recording medium and terminal for carrying out same | |
| Xu et al. | ER: Early recognition of inattentive driving leveraging audio devices on smartphones | |
| US9858813B2 (en) | Method for providing sound detection information, apparatus detecting sound around vehicle, and vehicle including the same | |
| Socoró et al. | Development of an Anomalous Noise Event Detection Algorithm for dynamic road traffic noise mapping | |
| WO2022150999A1 (zh) | 一种检测轮胎异常的方法和装置 | |
| JP2020152372A (ja) | 車両の外部の音響シーンを監視するためのシステム | |
| CN113837150B (zh) | 基于计算机视觉的非接触式轮胎气压获取方法及相关装置 | |
| CN113008985A (zh) | 一种利用轮胎/路面噪声评价道路路面构造的方法 | |
| CN113345443A (zh) | 基于梅尔频率倒谱系数的海洋哺乳动物发声检测识别方法 | |
| Kubo et al. | Design of ultra low power vehicle detector utilizing discrete wavelet transform | |
| CN112309423A (zh) | 驾驶环境下基于智能手机音频感知的呼吸道症状检测方法 | |
| CN112230208B (zh) | 一种基于智能手机音频感知的汽车行驶速度检测方法 | |
| CN214796752U (zh) | 一种基于声纹识别和声源定位的工程车辆识别定位装置 | |
| CN104991245A (zh) | 一种无人飞行器预警装置及其预警方法 | |
| JP7112909B2 (ja) | タイヤ種判別方法及びタイヤ種判別装置 | |
| CN205003281U (zh) | 一种无人飞行器预警装置 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21918217 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 21918217 Country of ref document: EP Kind code of ref document: A1 |