CN114710419B - Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium - Google Patents
Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium Download PDFInfo
- Publication number
- CN114710419B CN114710419B CN202210157335.9A CN202210157335A CN114710419B CN 114710419 B CN114710419 B CN 114710419B CN 202210157335 A CN202210157335 A CN 202210157335A CN 114710419 B CN114710419 B CN 114710419B
- Authority
- CN
- China
- Prior art keywords
- sound
- characteristic
- signal
- identified
- power supply
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/12—Network monitoring probes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
Abstract
The invention relates to a device working state single-point monitoring method, device and storage medium based on switching power supply sound, wherein the method utilizes a microphone built in an intelligent sound box to realize remote single-point detection of the working condition of indoor electronic equipment without additional hardware equipment. The intelligent sound box prototype is particularly realized, sound emitted by the sound box switch power supply adapter is monitored through a microphone of the intelligent sound box, and the switch and the working state of indoor electronic equipment are monitored. The method realizes the single-point and remote detection of the working conditions of a plurality of electronic devices in the house without additional hardware equipment, and optimizes the intelligent home ecological system.
Description
Technical Field
The present invention relates to a method and an apparatus for monitoring a single point of an operating state of a device based on a switching power supply sound, and a storage medium.
Background
Smart home ecosystems have long attracted many technicians to study various technologies, hopefully creating sophisticated monitoring and communication mechanisms between different applications. And as the intelligent sound box brings the voice interaction function into the home, the use of the intelligent sound box in the intelligent home is rapidly increasing. Because of the convenience and rapidity of the intelligent sound box, users often use the intelligent sound box as a convenient hub for monitoring and controlling various intelligent devices. Intelligent sound boxes have gradually become a bridge for establishing communication between users and various electronic devices. However, to establish contact with the existing commercial smart devices, the electronic devices must be equipped with smart communication modules (such as bluetooth, wi-Fi, zigbee, etc.), which requires users to replace the conventional devices with smart devices or to install additional smart modules, thereby limiting the widespread development of smart home ecosystems. Therefore, breaking the limitation between the intelligent sound box and the traditional non-intelligent equipment, it is important to establish a bridge for communication between the intelligent sound box and the traditional non-intelligent equipment.
In order to solve the monitoring problem of electronic equipment, scientific researchers have proposed various solutions, which can be mainly divided into the following three categories: 1) Distributed sensor deployment. The implementation method of the scheme is that a sensor is arranged on each device to realize the monitoring of the device level, and the methods comprise methods based on vision (reference Zhang, chenyang, and YIngli Tian. "RGB-D camera-based daily living activity recognment." Journal of computer vision and image processing 2.4.4 (2012): 12.), sound (reference Pathak, nilavra, md Abdullah Al Hafiz Khan, and Nirmalya Roy. "Acoustic based appliance state identifications for fine-graded energy analysis systems." 2015 IEEE International Conference on Pervasive Computing and Communications (Percom. IEEE, 2015.) "or radiation signal (reference Laput, gierad, et al." Em-sense: touch recognition of uninstrumented, electrical and electromechanical objects. "Proceedings of the 28th Annual ACM Symposium on User Interface Software&Technology.2015.), however, the methods all require the installation of special sensors or intelligent modules, which not only cause inconvenience to users, but also raise costs; 2) Non-invasive load monitoring (NILM) (reference Hart, george William. "Nonintrusive appliance load monitoring." Proceedings of the IEEE 80.12.12 (1992): 1870-1891.). The scheme analyzes the total electricity consumption condition in the home environment, and presumes which devices are working in the current period through the total electricity consumption. For example, some methods (references Dan, wang, huang Xiao Li, and Ye Shu Ce. "Review of non-intrusive load appliance monitoring," 2018 IEEE 3rd Advanced Information Technology,Electronic and Automation Control Conference (IAEAC), "IEEE, 2018, or Zeifinan, m., craig Akers, and Kurt roth," Non intrusive appliance load monitoring (nialm) for energy control in residential building applications, "Energy Etliciency in Domestic Appliances and Lighting 20 II (2011): 24-26) extract power consumption change characteristics to infer the operation of the home device. However, these methods have limitations in that they are not effective for complex time-varying devices (e.g., computers, projectors) and some low-power devices (e.g., lamps), and they cannot distinguish between home appliances of the same type. 3) Infrastructure-mediated awareness (IMS). Still other approaches utilize single point sensing techniques to detect the effects of equipment on the medium of the overall building infrastructure, including effects on air, water flow, power lines, or vibrations of the building infrastructure. Such as the vibroseis (Sun, wei, et al, "vibroseis: recognizing Home Activities by Deep Learning Subtle Vibrations on an Interior Surface of a House from a Single Point Using Laser Doppler vibrometry" Proceedings of the ACM on Interactive, mobile, wearable and Ubiquitous Technologies 4.3.4.3 (2020): 1-28.) uses a doppler vibrometer to measure the effect of the appliance operating on the entire house structure vibration, and Gulati et al designs electromagnetic interference sensors or USRP monitors the electromagnetic interference conditions of the appliance operating on the power lines, and NoDE and OutletSpy use oscilloscopes to monitor the socket voltage for appliance operation. While these methods enable single point detection of home appliance operating conditions within a house, they typically require the use of specialized equipment that is impractical for a typical home environment.
Disclosure of Invention
The invention aims to provide a device working state single-point monitoring method and device based on switching power supply sound and a storage medium. The microphone built in the intelligent sound box is utilized to realize the working condition of the remote single-point detection indoor electronic equipment on the premise of not using additional hardware equipment. The intelligent sound box prototype is realized, sound emitted by a sound box switch power supply (adapter) is monitored through a microphone of the intelligent sound box, and the switch and the working state of indoor electronic equipment are monitored. The principle is that a PFC circuit in an electronic device can generate a current containing high-frequency characteristics, and the current can affect all circuit branches under the same power network, so that the branch current contains specific high-frequency components. Meanwhile, the branch current acts on the electronic component in the switching power supply (adapter) due to magnetostriction and piezoelectric effect and makes the component vibrate to make sound. The microphone monitors the sound, extracts relevant high-frequency characteristics, and further analyzes the working state of the electronic equipment. The method realizes the single-point and remote detection of the working conditions of a plurality of electronic devices in the house without additional hardware equipment, and optimizes the intelligent home ecological system.
The aim of the invention can be achieved by the following technical scheme:
a device working state single-point monitoring method based on switching power supply sound comprises the following steps:
collecting sound of a switching power supply adapter when equipment to be detected does not work as a first environmental background sound, wherein the switching power supply adapter is a switching power supply adapter of a microphone;
respectively establishing a characteristic template for each device to be monitored based on the collected first environmental background sound, wherein the process for establishing the characteristic template for the device to be monitored comprises the following steps: when the equipment to be monitored works independently, the sound of the switching power supply adapter is used as a first sampling sound; subtracting the first environmental background sound from the first sampled sound to obtain a characterization sound corresponding to the equipment, extracting a characteristic signal of the characterization sound, and obtaining a characteristic template based on the extracted characteristic signal;
collecting the sound of the switching power supply adapter and subtracting the first environmental background sound to obtain an original sound to be identified, and extracting a characteristic signal of the sound to be identified;
signal comparison: comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain a characteristic template with the smallest difference with the characteristic signals of the to-be-identified sound, and judging whether the smallest difference is smaller than a set threshold value;
and if the minimum difference is smaller than the set threshold value, outputting that the equipment to be monitored corresponding to the characteristic template is in a working state, subtracting the characteristic template from the characteristic signal of the sound to be identified, and then taking the characteristic signal as the characteristic signal of the new sound to be identified, and repeating the signal comparison step.
The method further comprises the steps of:
and when the trigger signal is received, the sound of the switching power supply adapter is collected again to update the background sound of the first environment when the equipment to be detected works.
The trigger signal comprises a timing trigger signal and a starting monitoring signal, wherein the starting monitoring is used for representing the starting of the equipment working state monitoring.
Comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the minimum difference between the characteristic signals of the to-be-identified sound, wherein the characteristic template specifically comprises the following components:
and calculating the Euclidean distance between the characteristic signal of the to-be-identified sound and the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the minimum Euclidean distance between the characteristic signal of the to-be-identified sound and the characteristic templates.
The extraction process of the characteristic signals comprises the following steps:
performing parameter-optimized variation modal decomposition on the audio signal by pre-configuring the number of decomposition signals and the value range of penalty factors, calculating the total bandwidth of the obtained inherent mode function under different decomposition conditions,
and selecting the combination of the number of the decomposition signals and the penalty factors corresponding to the minimum total bandwidth, screening the signal components through periodic detection, and taking the sum of all the screened signal components as a characteristic signal.
The value range of the number of the decomposition signals is 2-40, and the value range of the penalty factor is 1000-100000.
The method for obtaining the characteristic template based on the extracted characteristic signal comprises the following steps:
splitting the characteristic signal into a plurality of characteristic signal fragments, storing,
and after the characteristic signals of all the equipment to be monitored are acquired and the characteristic signal fragments are obtained, linear characteristic extraction is carried out to obtain the characteristic templates of all the equipment to be monitored.
The method comprises the steps of collecting the sound of a switching power supply adapter and subtracting a first environmental background sound to obtain an original sound to be identified, and extracting a characteristic signal of the sound to be identified, wherein the characteristic signal specifically comprises: collecting the sound of a switching power supply adapter, subtracting a first environmental background sound to obtain an original sound to be identified, extracting a characteristic signal of the sound to be identified, and splitting the characteristic signal into a plurality of characteristic signal fragments of the sound to be identified;
comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the minimum difference between the characteristic signals of the to-be-identified sound, wherein the characteristic template specifically comprises the following components: comparing the characteristic signal segments of each to-be-identified sound with the characteristic templates of the to-be-monitored equipment to obtain the characteristic template with the minimum total difference between the to-be-identified sound and the characteristic signal segments;
the step of subtracting the characteristic template from the characteristic signal of the to-be-identified sound to serve as the characteristic signal of the new to-be-identified sound and repeating the signal comparison step specifically comprises the following steps: and subtracting the characteristic templates from the characteristic signal fragments of all the sounds to be identified to serve as the characteristic signal fragments of new sounds to be identified, and repeating the signal comparison step.
The device for monitoring the single point of the working state of the whole house electronic equipment based on the switching power supply sound comprises a memory, a processor and a program stored in the memory, wherein the processor realizes the method when executing the program.
A storage medium having stored thereon a program which when executed performs a method as described above.
Compared with the prior art, the invention has the following beneficial effects:
1. the remote single-point electronic equipment working state detection function is realized, no extra equipment is needed, and the remote single-point electronic equipment working state detection function can be realized only by using equipment with a microphone.
2. The method for monitoring the power grid environment by utilizing the sound is innovatively provided, and the channel conversion of current, magnetic field and sound is constructed. The use of sound detection saves energy consumption, does not need additional hardware, reduces cost, and provides new possibility for monitoring the power grid environment and the equipment working state.
3. The high-frequency sound signal sent by the switching power supply is utilized, and the signal can be received by the microphone but cannot be heard by human ears, so that the user experience is improved.
4. Only data of single-label equipment is collected, multi-label equipment classification is achieved, training complexity is reduced, training time is shortened, and convenience is brought to users.
5. The electronic equipment monitoring device not only can monitor the switch of the electronic equipment, but also can further monitor the working state of the electronic equipment, including the working gear, the charging state and the App detection of the electronic equipment, and expands the application scene.
6. The detection of the state of the electronic device by using the sound signal has the advantages of long detection distance compared with the existing work (magnetic field or voltage), no need of additional equipment, only single-point detection and the like.
7. When classifying, the accuracy can be improved by comparing a plurality of fragments.
Drawings
FIG. 1 is a schematic diagram of the structure of the present invention;
FIG. 2 is a flow chart of the method of the present invention;
FIG. 3 is a schematic diagram of the beneficial effects of the method of the present invention in extracting characteristic bands of registered devices;
FIG. 4 is the accuracy of each work device in the case of single tag classification;
fig. 5 shows the accuracy of each working device in the case of multiple labels.
Detailed Description
The invention will now be described in detail with reference to the drawings and specific examples. The present embodiment is implemented on the premise of the technical scheme of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following examples.
A single-point monitoring method for the working state of a full-house electronic device based on the sound of a switching power supply is characterized in that a microphone is utilized, (an intelligent sound box is used in the embodiment) to receive the sound sent out by a switching power supply adapter, frequency characteristics are extracted, and the working states of other electronic devices in the current circuit environment are obtained through analysis. The method comprises two parts of single-label equipment registration and multi-label classification, and mainly comprises the following steps:
1. single tag device registration
1.1 obtaining a section of background sound of a switching power supply in the current environment by using a microphone
1.2 turning on only one electronic device in the same environment, collecting the switching power supply sound at that time
1.3 filtering out Low frequency interference and noise Using Low pass Filter
1.4 weakening the self sound of the equipment and improving the signal to noise ratio
1.5 detection and extraction of frequency bands containing signals from other electronic devices
1.6, carrying out segmentation cutting on the residual signals, extracting features from the processed data set, registering the features as a feature template of the equipment, and storing the feature template in a model library.
2. Multi-tag device classification
2.1 recording a background sound with microphone without device operation
2.2 uninterrupted recording of sound with microphone when the device is in operation (there may be multiple devices)
2.3 sliding window slicing is carried out on the received sound, and the signal to noise ratio is improved on the slice
2.4 detecting and separating Signal bands for the slice
2.5 extracting a composite signal characteristic model from the residual signal, calculating the gap between the composite signal characteristic model and the registered equipment in the step 1 by using the single characteristic model, recording the nearest equipment label, and judging that the equipment is working. The device feature model is then subtracted from the composite signal model and a second round of similarity calculation is performed. Repeating the above operation until the distance is greater than a certain threshold value.
Specifically, this application need not extra equipment, utilizes intelligent audio amplifier from the microphone of taking can realize, and intelligent audio amplifier and electronic equipment need not to place in same room simultaneously. Fig. 1 is a schematic view of a usage scenario of the present application. The method and the device can break the bottleneck that the traditional electronic equipment cannot be monitored by the existing intelligent sound box, and expand the application ecology of the intelligent home.
FIG. 2 is a flow chart of the apparatus and method, the invention is realized by the following steps:
step S1: single tag device registration, i.e. registering a single device to be monitored, comprises:
collecting sound of a switching power supply adapter when equipment without detection works as a first environmental background sound, wherein the switching power supply adapter is a switching power supply adapter of a microphone;
respectively establishing a characteristic template for each device to be monitored based on the collected first environmental background sound, wherein the process for establishing the characteristic template for the device to be monitored comprises the following steps: collecting the sound of a switching power supply adapter as a first sampling sound when the equipment to be monitored works independently; subtracting the first environmental background sound from the first sampled sound to obtain a characterization sound corresponding to the equipment, extracting a characteristic signal of the characterization sound, and obtaining a characteristic template based on the extracted characteristic signal;
the detailed process is as follows
Step S101: first, in the absence of device operation, the microphone collects ambient background sound Y bg The ambient background sound signal contains low frequency ambient sound and the sound produced by the adapter itself during normal operation.
Step S102: the single registration device (i.e. the device to be monitored) is operated, and the microphone is again used to collect the sound signal, denoted as Y r The signal contains low-frequency environmental sounds, the sound of the adapter itself when the adapter is in normal operation and the characteristic sound S of the vibration generated by the current acting on the electronic components of the adapter when the registration device is in operation.
Step S103: and the signal to noise ratio is improved. In this embodiment, a feasible implementation manner is provided, but not limited to this manner: estimating ambient background sound Y bg Distribution, denoted as E (Y bg ) The signal-to-noise ratio is improved by using spectral subtraction, and the sound signal containing the registration device, excluding the background sound signal, is extracted and denoted as S'. Satisfy the expression S' =y r -E(Y bg ) =s+_e, where e represents the estimation error.
Step S104: extracting the characteristic frequency band of the registration equipment from the residual signal S', wherein the characteristic frequency band comprises the following steps: and carrying out parameter-optimized variation modal decomposition on the audio signal by pre-configuring the number of decomposition signals and the value range of penalty factors, calculating the total bandwidth of the natural mode function obtained under different decomposition conditions, selecting the combination of the number of decomposition signals and the penalty factors corresponding to the minimum total bandwidth, screening the signal components through periodic detection, and taking the sum of all the screened signal components as a characteristic signal. In this embodiment, the range of the number of the decomposed signals is 2-40, and the range of the penalty factor is 1000-100000.
Namely, the present embodiment provides a feasible implementation manner, but is not limited to this manner: and performing parameter-optimized Variational Modal Decomposition (VMD) on the signal, defining the range of signal quantity k (set as 2-40 in scheme) and penalty factor alpha (set as 1000-100000 in scheme) of parameter decomposition, calculating the total bandwidth of the Intrinsic Mode Function (IMF) obtained under different decomposition conditions, selecting the combination of k and alpha with minimum total bandwidth, and screening the IMF through period detection. The sum of all the screened IMFs obtained at this time is used as the characteristic signal of the registration device.
Step S105: splitting the processed characteristic signal into a plurality of fragments and storing the fragments in a database.
Step S106: and after collecting a plurality of devices or working states, performing linear feature extraction on the database, and taking the result after feature extraction as a feature template of each device for subsequent classification.
Step S2: multi-tag device classification, comprising:
collecting the sound of the switching power supply adapter and subtracting the first environmental background sound to obtain an original sound to be identified, and extracting a characteristic signal of the sound to be identified;
signal comparison: comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain a characteristic template with the smallest difference with the characteristic signals of the to-be-identified sound, and judging whether the smallest difference is smaller than a set threshold value;
if the minimum difference is smaller than the set threshold, outputting that the equipment to be monitored corresponding to the characteristic template is in a working state, subtracting the characteristic template from the characteristic signal of the sound to be identified, and then taking the characteristic signal as the characteristic signal of the new sound to be identified, and repeating the signal comparison step.
The detailed process is as follows
Step S201: similar to step S101, the microphone first collects ambient background sound Y without the device in operation bg
Step S202: and (3) utilizing the microphone to collect sound signals in real time as a sliding window to split the sound signals into a plurality of signal fragments, and carrying out processing similar to the steps S103 and S104 on each fragment to obtain the characteristic frequency band of the fragment.
Step S203: extracting the characteristic signal by using the characteristic model obtained in the step S105 to obtain a characteristic signal Y
Step S204: and calculating the similarity between the characteristic signal and a template (marked as X) in a characteristic template library. The present embodiment provides a feasible implementation manner, but is not limited to this manner: using Euclidean distance calculation, the method is thatCalculating the characteristic signal closest to the device and less than the threshold value indicates that the corresponding device is operating.
In this embodiment, a manner of splitting a feature signal of a sound to be identified into a plurality of segments is adopted, so that each segment needs to be compared with a feature template, then a feature template closest to each segment is obtained, voting is finally performed, the feature template with the largest voting is used as a final identification result, and equipment corresponding to the feature template is used as equipment which is identified to be running.
Step S205: because of the linear additivity of feature extraction, each time a device is determined by S204, the feature template of the device is subtracted, and the operation of S204 is repeated by the remaining feature signals until the calculated similarity is greater than the threshold, at which time all the working devices are found.
Step S206: and feeding back the monitored working state to the user.
In this embodiment, after receiving the trigger signal, the sound of the switching power supply adapter is collected again to update the first environmental background sound when the equipment to be detected is not in operation. The trigger signal comprises a timing trigger signal and a start monitoring signal, wherein the start monitoring is used for representing the starting of the equipment working state monitoring. Therefore, the updating of the environmental background sound can be realized, and the accuracy is improved.
The following description will illustrate specific examples that were tested in a real operating environment, containing 34 states for 28 different operating devices. The sound signal is sampled with a microphone with a sampling rate of 192 kHz. In the registration stage, firstly, under the condition of no working equipment, collecting environment background sounds with the duration of 30 seconds, and estimating the distribution of the environment background sounds; and then, respectively collecting sound signals with the time length of 2 minutes for each state of each working device, carrying out low-pass filtering and spectral subtraction on the signals to improve the signal to noise ratio, and extracting sound signals containing the working states of the registration devices except environmental background sounds. Next, the characteristic frequency band where the registration device is located is extracted from the remaining signals, in this embodiment, a parameter-optimized Variational Mode Decomposition (VMD) method is selected, the preset parameter range includes the number k (2-40) of decomposed signals and the alpha (1000-100000) of penalty factors, and when k and alpha respectively take 10 and 10000 in the present data set, the total bandwidth of the decomposed Intrinsic Mode Function (IMF) is minimum, so that the combination is taken as an optimal combination, and then the IMF with a period is screened through period detection. The sum of all the screened IMFs obtained at this time is used as the characteristic signal of the registration device. Fig. 3 shows the beneficial effect of extracting the characteristic frequency band of the registration device. After the above steps are completed, a feature template of the feature signal is extracted. In the step, the original 2-minute signal is firstly sliced and segmented, the segmentation length is 2s in the embodiment, all the segmented signals are used as a training set, the Principal Component Analysis (PCA) is used for extracting linear additive features to perform dimension reduction, a PCA model of a corresponding data set is trained, and feature vectors after dimension reduction are used as feature templates of working equipment and stored in a feature template library.
In the classification phase, ambient background sounds are first collected (30 seconds) without the equipment in operation, and as the system is running, the microphone collects the sound signals in real time and cuts into segments (one segment every 2 seconds) using a sliding window, where only one segment is used for classification, i.e. once every 2 seconds. And generating feature vectors of the fragments after the dimension reduction of the corresponding fragments by using a trained PCA model, and calculating the similarity between the vectors and templates in a feature template library. And then feeding back the monitored working state to the user.
Fig. 4 and 5 show the accuracy of each working device in the case of single-label classification and in the case of multi-label classification, respectively, wherein the average accuracy in the case of single-label classification is higher than 99.5%, and the average accuracy in the case of multi-label classification is higher than 99%, which represents the effectiveness of the present application.
The above functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Claims (10)
1. The equipment working state single-point monitoring method based on the switching power supply sound is characterized by comprising the following steps of:
collecting sound of a switching power supply adapter when equipment to be detected does not work as a first environmental background sound, wherein the switching power supply adapter is a switching power supply adapter of a microphone;
respectively establishing a characteristic template for each device to be monitored based on the collected first environmental background sound, wherein the process for establishing the characteristic template for the device to be monitored comprises the following steps: when the equipment to be monitored works independently, the sound of the switching power supply adapter is used as a first sampling sound; subtracting the first environmental background sound from the first sampled sound to obtain a characterization sound corresponding to the equipment, extracting a characteristic signal of the characterization sound, and obtaining a characteristic template based on the extracted characteristic signal;
collecting the sound of the switching power supply adapter and subtracting the first environmental background sound to obtain an original sound to be identified, and extracting a characteristic signal of the sound to be identified;
signal comparison: comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain a characteristic template with the smallest difference with the characteristic signals of the to-be-identified sound, and judging whether the smallest difference is smaller than a set threshold value;
and if the minimum difference is smaller than the set threshold value, outputting that the equipment to be monitored corresponding to the characteristic template is in a working state, subtracting the characteristic template from the characteristic signal of the sound to be identified, and then taking the characteristic signal as the characteristic signal of the new sound to be identified, and repeating the signal comparison step.
2. The method for single point monitoring of the operating state of a full house electronic device based on switching power supply sound according to claim 1, further comprising:
and when the trigger signal is received, the sound of the switching power supply adapter is collected again to update the background sound of the first environment when the equipment to be detected works.
3. The method for single-point monitoring of the operating state of a full-house electronic device based on switching power supply sound according to claim 2, wherein the trigger signal comprises a timing trigger signal and a start monitoring signal, and the start monitoring is used for representing that the device operating state monitoring is started.
4. The method for single-point monitoring of the working state of a full-house electronic device based on switching power supply sound according to claim 1, wherein the comparing the characteristic signal of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the smallest difference with the characteristic signal of the to-be-identified sound comprises the following steps:
and calculating the Euclidean distance between the characteristic signal of the to-be-identified sound and the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the minimum Euclidean distance between the characteristic signal of the to-be-identified sound and the characteristic templates.
5. The method for single-point monitoring of the working state of a whole-house electronic device based on switching power supply sound according to claim 1, wherein the extraction process of the characteristic signal of the characteristic sound and the characteristic signal of the sound to be identified comprises the following steps:
performing parameter-optimized variation modal decomposition on the audio signal by pre-configuring the number of decomposition signals and the value range of penalty factors, calculating the total bandwidth of the obtained inherent mode function under different decomposition conditions,
and selecting the combination of the number of the decomposition signals and the penalty factors corresponding to the minimum total bandwidth, screening the signal components through periodic detection, and taking the sum of all the screened signal components as a characteristic signal.
6. The method for single-point monitoring of the working state of the whole-house electronic equipment based on the switching power supply sound according to claim 5, wherein the number of the decomposed signals is 2-40, and the penalty factor is 1000-100000.
7. The method for single-point monitoring of the operating state of a whole-house electronic device based on switching power supply sound according to claim 1, wherein the method for obtaining the feature template based on the extracted feature signal comprises the following steps:
splitting the characteristic signal of the characteristic sound into a plurality of characteristic signal segments, storing,
and after the characteristic signals of the characteristic sounds of all the equipment to be monitored are acquired and the characteristic signal fragments are obtained, linear characteristic extraction is carried out to obtain the characteristic templates of all the equipment to be monitored.
8. The method for monitoring the single point of the working state of the whole house electronic equipment based on the sound of the switching power supply according to claim 1, wherein the steps of collecting the sound of the switching power supply adapter and subtracting the background sound of the first environment to obtain the original sound to be identified, extracting the characteristic signal of the sound to be identified are as follows: collecting the sound of a switching power supply adapter, subtracting a first environmental background sound to obtain an original sound to be identified, extracting a characteristic signal of the sound to be identified, and splitting the characteristic signal into a plurality of characteristic signal fragments of the sound to be identified;
comparing the characteristic signals of the to-be-identified sound with the characteristic templates of the to-be-monitored devices to obtain the characteristic template with the minimum difference between the characteristic signals of the to-be-identified sound, wherein the characteristic template specifically comprises the following components: comparing the characteristic signal segments of each to-be-identified sound with the characteristic templates of the to-be-monitored equipment to obtain the characteristic template with the minimum total difference between the to-be-identified sound and the characteristic signal segments;
the step of subtracting the characteristic template from the characteristic signal of the to-be-identified sound to serve as the characteristic signal of the new to-be-identified sound and repeating the signal comparison step specifically comprises the following steps: and subtracting the characteristic templates from the characteristic signal fragments of all the sounds to be identified to serve as the characteristic signal fragments of new sounds to be identified, and repeating the signal comparison step.
9. A full-house electronic equipment working state single-point monitoring device based on switching power supply sound, which comprises a memory, a processor and a program stored in the memory, wherein the processor realizes the method as set forth in any one of claims 1-8 when executing the program.
10. A storage medium having a program stored thereon, wherein the program, when executed, implements the method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210157335.9A CN114710419B (en) | 2022-02-21 | 2022-02-21 | Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210157335.9A CN114710419B (en) | 2022-02-21 | 2022-02-21 | Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114710419A CN114710419A (en) | 2022-07-05 |
| CN114710419B true CN114710419B (en) | 2023-07-28 |
Family
ID=82166299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210157335.9A Active CN114710419B (en) | 2022-02-21 | 2022-02-21 | Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114710419B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228979A (en) * | 2016-08-16 | 2016-12-14 | 重庆大学 | A kind of abnormal sound in public places feature extraction and recognition methods |
| CN109819060A (en) * | 2018-12-15 | 2019-05-28 | 深圳壹账通智能科技有限公司 | Method for detecting abnormality, device, computer installation and storage medium |
| CN112667717A (en) * | 2020-12-23 | 2021-04-16 | 贵州电网有限责任公司电力科学研究院 | Transformer substation inspection information processing method and device, computer equipment and storage medium |
| CN114024560A (en) * | 2021-12-15 | 2022-02-08 | 宁波伊士通技术股份有限公司 | Echo suppression and howling prevention voice intercom system based on program-controlled electronic attenuator |
-
2022
- 2022-02-21 CN CN202210157335.9A patent/CN114710419B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106228979A (en) * | 2016-08-16 | 2016-12-14 | 重庆大学 | A kind of abnormal sound in public places feature extraction and recognition methods |
| CN109819060A (en) * | 2018-12-15 | 2019-05-28 | 深圳壹账通智能科技有限公司 | Method for detecting abnormality, device, computer installation and storage medium |
| CN112667717A (en) * | 2020-12-23 | 2021-04-16 | 贵州电网有限责任公司电力科学研究院 | Transformer substation inspection information processing method and device, computer equipment and storage medium |
| CN114024560A (en) * | 2021-12-15 | 2022-02-08 | 宁波伊士通技术股份有限公司 | Echo suppression and howling prevention voice intercom system based on program-controlled electronic attenuator |
Non-Patent Citations (1)
| Title |
|---|
| 一种阈值自适应调整的实时音频分割方法;李超;熊璋;薛玲;刘云;;北京航空航天大学学报(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114710419A (en) | 2022-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gopinath et al. | Energy management using non-intrusive load monitoring techniques–State-of-the-art and future research directions | |
| Alcalá et al. | Event-based energy disaggregation algorithm for activity monitoring from a single-point sensor | |
| US10679135B2 (en) | Periodicity analysis on heterogeneous logs | |
| Scanlon et al. | Residual life prediction of rotating machines using acoustic noise signals | |
| US20140207398A1 (en) | Transient Normalization for Appliance Classification, Disaggregation, and Power Estimation in Non-Intrusive Load Monitoring | |
| Ma et al. | Toward energy-awareness smart building: Discover the fingerprint of your electrical appliances | |
| CN109714324B (en) | User network abnormal behavior discovery method and system based on machine learning algorithm | |
| CN103295584B (en) | Speech data detection device, speech supervisory system and method thereof | |
| CN105143892A (en) | Power entry point energy measuring apparatus and energy measurement information labeling system using same | |
| Ntalampiras et al. | Acoustic detection of human activities in natural environments | |
| WO2012160062A1 (en) | Transition detection method for automatic-setup non-intrusive appliance load monitoring | |
| Guvensan et al. | Energy monitoring in residential spaces with audio sensor nodes: TinyEARS | |
| CN111382789A (en) | Power load identification method and system based on machine learning | |
| Majid et al. | Continuous sensing on intermittent power | |
| Chen et al. | Significant correlation pattern mining in smart homes | |
| CN103778916A (en) | Method and system for monitoring environmental sound | |
| Englert et al. | Reduce the number of sensors: Sensing acoustic emissions to estimate appliance energy usage | |
| CN114710419B (en) | Equipment working state single-point monitoring method and device based on switching power supply sound and storage medium | |
| CN116933895B (en) | Internet of things data mining method and system based on machine learning | |
| Jorde et al. | Event detection for energy consumption monitoring | |
| Zoha et al. | Acoustic and device feature fusion for load recognition | |
| Wüstrich et al. | Cyber-physical anomaly detection for ICS | |
| Adabi et al. | Seads: A modifiable platform for real time monitoring of residential appliance energy consumption | |
| JP6766346B2 (en) | Information processing device, activity status detection program and activity status detection method | |
| CN109946538B (en) | Mobile equipment monitoring method and system based on magnetic induction signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |