Power dispatching operation instruction safety management method and system based on AI identification

Abstract

The invention discloses a power dispatching operation instruction safety management method and system based on AI recognition, and relates to the technical field of voice instruction recognition, and the method comprises the steps: carrying out the data collection, and carrying out the feature extraction of the data; recognizing a voice operation instruction based on a voice recognition model of multi-model fusion; and classifying the instruction sequence and carrying out instruction confirmation. According to the power dispatching operation instruction safety management method based on AI identification, the voice instruction is confirmed through a method of maximizing the posterior probability, accurate confirmation of the instruction classification result is achieved, it is ensured that each instruction is strictly verified before execution, and therefore the risk of misoperation is reduced, and the safety of the power dispatching operation instruction is improved. And the original voice signal is digitalized and characterized, so that the effectiveness and accuracy of subsequent processing are ensured. The method has better effects in the aspects of accuracy and precision.

Description

Technical Field

[0001] The present invention relates to the technical field of voice command recognition, and in particular to a method and system for safely managing electric power dispatching operation commands based on AI recognition. Background Art

[0002] In the process of power dispatching operations, it is crucial to accurately convey and execute instructions. Traditional power dispatching operations rely on manual communication and execution, which has many potential risks and challenges. In recent years, with the rapid development of artificial intelligence (AI) technology, AI has made significant progress in the fields of speech recognition, natural language processing, and machine learning, bringing new opportunities for the communication and management of power dispatching operation instructions. Speech recognition technology converts voice signals into text, enabling machines to understand and execute voice instructions, thereby improving the automation and accuracy of operations. Most existing speech recognition systems use a single model, such as models based on hidden Markov models (HMM) or deep neural networks (DNN). These models show certain limitations in practical applications, especially in complex power dispatching scenarios. When faced with diverse voice instructions and background noise, the recognition accuracy and real-time performance need to be improved.

[0003] Although the existing technology has made some progress in the voice recognition of power dispatching operation instructions, there are still many shortcomings. First, when processing complex power dispatching instructions, traditional voice recognition systems are easily affected by background noise and voice variation, resulting in a decrease in recognition accuracy. Secondly, the existing single model cannot provide sufficient robustness and generalization capabilities when faced with diverse instructions. In addition, the existing instruction confirmation mechanism usually relies on a single confirmation method and cannot effectively avoid the risks of misoperation and missed operations. Therefore, there is an urgent need for an AI recognition technology based on multi-model fusion to improve the recognition accuracy and robustness of voice instructions, and to ensure the safety and reliability of instruction execution by integrating multiple confirmation methods. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing scheduling instruction recognition method is easily affected by the compilation of noise and voice, has insufficient accuracy, poor generalization ability, and how to avoid erroneous operations.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for safe management of power dispatching operation instructions based on AI recognition, comprising data collection and feature extraction of the data; recognizing voice operation instructions based on a speech recognition model based on multi-model fusion; classifying instruction sequences and confirming the instructions.

[0007] As a preferred solution of the AI-recognition-based power dispatching operation instruction safety management method described in the present invention, the data collection includes voice data collection, and the voice data collection captures the dispatching instruction voice in real time through a microphone array, denoises the voice, and performs feature extraction after processing.

[0008] As a preferred solution of the power dispatching operation instruction safety management method based on AI recognition described in the present invention, wherein: the data recognition processing includes recognizing voice information, extracting the feature vector obtained by feature extraction of the voice signal, dividing the voice signal into multiple small frames, each frame length is 20-40 milliseconds, each frame x t It is expressed as:

[0009] x t =[x(t),x(t+1),…,x(t+N-1)]

[0010] Where N is the number of samples per frame;

[0011] Multiply each frame signal by the Hamming window function, expressed as:

[0012] x′ t =x t ·w(n)

[0013] Where w(n) is the Hamming window function;

[0014] Perform fast Fourier transform on each frame of processed signal to convert the time domain signal to the frequency domain, and divide the frequency domain signal into several Mel frequency bands through a set of Mel filters, which can be expressed as:

[0015]

[0016] Among them, X t (k) is the signal after fast Fourier transform, H m (k) is the frequency response of the mth Mel filter.

[0017] As a preferred solution of the power dispatching operation instruction safety management method based on AI recognition described in the present invention, the data recognition processing also includes discrete cosine transforming the energy of the logarithmic Mel frequency band to obtain the Mel frequency cepstrum coefficient MFCC, which represents the feature vector of the voice dispatching instruction signal, expressed as:

[0018]

[0019] Among them, O t (n) is the n-th MFCC coefficient of the t-th frame, and M represents the number of Mel frequency bands.

[0020] As a preferred solution of the AI-based recognition-based power dispatching operation instruction safety management method described in the present invention, the speech recognition model based on multi-model fusion recognizes the voice operation instruction, including processing sequence data through a recurrent neural network GUR, assigning weights to the feature vectors of each time step through an attention mechanism, and the weighted summed output is used as the final feature representation sequence y of the operation instruction.

[0021] As a preferred solution of the power dispatching operation instruction safety management method based on AI recognition described in the present invention, wherein: the classification of instruction sequences includes classifying operation instruction sequences by similarity, extracting data points from sequence y for input, expressed as:

[0022]

[0023] in, is the instruction label after classification, α j and b are the parameters of the support vector machine, K(x,x j ) is the kernel function, which represents the data point x and the support vector x j The similarity between .

[0024] As a preferred solution of the power dispatching operation instruction safety management method based on AI recognition described in the present invention, the instruction confirmation includes maximizing the posterior of the voice instruction after completing the instruction classification to detect whether the voice instruction is correct, which is expressed as:

[0025]

[0026] Among them, C is the final confirmed result, C′ represents the candidate data, and P(c i |C′,y t ) is the confirmation probability of a single operator, P(C ′ ) is the prior probability, λ is the attenuation coefficient of the exponential function, μ k and σ 2 are the mean and variance of the Gaussian distribution respectively;

[0027] The results are displayed to the on-duty personnel, and after the command is confirmed, the system begins to execute the voice command.

[0028] Another object of the present invention is to provide an AI-based recognition-based power dispatching operation instruction safety management system, which can confirm voice instructions by maximizing the posterior probability method, thereby achieving accurate confirmation of instruction classification results, ensuring that each instruction is strictly verified before execution, thereby reducing the risk of misoperation and solving the problem of current voice dispatching instruction recognition containing erroneous operations.

[0029] As a preferred solution of the AI-recognition-based power dispatching operation instruction safety management system described in the present invention, it includes: a data acquisition module, an instruction extraction module, and an instruction confirmation module; the data acquisition module is used to collect data and extract features from the data; the instruction extraction module is used to recognize voice operation instructions based on a multi-model fusion speech recognition model; the instruction confirmation module is used to classify instruction sequences and confirm instructions.

[0030] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a method for safely managing power dispatching operation instructions based on AI recognition.

[0031] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for safely managing power dispatching operation instructions based on AI recognition.

[0032] Beneficial effects of the present invention: The power dispatching operation instruction safety management method based on AI recognition provided by the present invention confirms the voice instruction by maximizing the posterior probability method, realizes the accurate confirmation of the instruction classification result, ensures that each instruction is strictly verified before execution, thereby reducing the risk of misoperation, digitizes and characterizes the original voice signal, and ensures the effectiveness and accuracy of subsequent processing. The present invention achieves better results in terms of accuracy and precision. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0034] Figure 1 An overall flow chart of a method for safe management of electric power dispatching operation instructions based on AI recognition provided in the first embodiment of the present invention.

[0035] Figure 2 A schematic diagram of a module of an electric power dispatching operation instruction safety management system based on AI recognition provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0037] Example 1

[0038] Reference Figure 1 , as an embodiment of the present invention, provides a method for safe management of power dispatching operation instructions based on AI recognition, comprising:

[0039] S1: Collect data and extract features from the data.

[0040] Furthermore, the data collection includes voice data collection, which captures the dispatch instruction voice in real time through a microphone array, denoises the voice, and extracts features after processing.

[0041] It should be noted that the data recognition processing includes recognizing the voice information, extracting the feature vectors from the voice signal, dividing the voice signal into multiple small frames, each frame length is 20-40 milliseconds, and each frame x t It is expressed as:

[0042] x t =[x(t),x(t+1),…,x(t+N-1)]

[0043] Where N is the number of samples per frame;

[0044] Multiply each frame signal by the Hamming window function, expressed as:

[0045] x′ t =x t ·w(n)

[0046] Where w(n) is the Hamming window function;

[0047] Perform fast Fourier transform on each frame of processed signal to convert the time domain signal to the frequency domain, and divide the frequency domain signal into several Mel frequency bands through a set of Mel filters, which can be expressed as:

[0048]

[0049] Among them, X t (k) is the signal after fast Fourier transform, H m (k) is the frequency response of the mth Mel filter.

[0050] It should also be noted that the data recognition processing also includes performing discrete cosine transform on the energy of the logarithmic Mel frequency band to obtain Mel frequency cepstrum coefficients MFCC, which represent the feature vector of the voice scheduling instruction signal, expressed as:

[0051]

[0052] Among them, O t (n) is the n-th MFCC coefficient of the t-th frame, and M represents the number of Mel frequency bands.

[0053] S2: Recognize voice operation instructions based on the speech recognition model of multi-model fusion.

[0054] Furthermore, the speech recognition model based on multi-model fusion recognizes voice operation instructions, including processing sequence data through a recurrent neural network GUR, assigning weights to the feature vectors of each time step through an attention mechanism, and outputting the weighted sum as the final feature representation sequence y of the operation instruction.

[0055] S3: Classify the instruction sequence and confirm the instruction.

[0056] Furthermore, the classification of the instruction sequence includes classifying the operation instruction sequence by similarity, extracting data points from the sequence y for input, which is expressed as:

[0057]

[0058] in, is the instruction label after classification, α j and b are the parameters of the support vector machine, K(x,x j ) is the kernel function, which represents the data point x and the support vector x j The similarity between .

[0059] It should be noted that the instruction confirmation includes maximizing the posterior of the voice instruction after completing the instruction classification to detect whether the voice instruction is correct, which is expressed as:

[0060]

[0061] Among them, C is the final confirmed result, C′ represents the candidate data, and P(c i |C′,y t ) is the confirmation probability of a single operator, P(C ′ ) is the prior probability, λ is the attenuation coefficient of the exponential function, μ k and σ 2 are the mean and variance of the Gaussian distribution respectively;

[0062] The results are displayed to the on-duty personnel, and after the command is confirmed, the system starts to execute the voice command. By classifying voice commands and calculating the maximum posterior probability, the accuracy and reliability of command recognition can be effectively improved. Not only can it ensure that the system executes the correct commands and reduce operational errors, but it can also improve the security and stability of the system through multiple verifications.

[0063] Example 2

[0064] One embodiment of the present invention provides a method for safe management of power dispatching operation instructions based on AI recognition. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0065] First, data collection is carried out, mainly including voice data collection. Voice data is captured in real time through a microphone array to capture the power dispatch command voice. The captured voice signal is processed by noise reduction to remove environmental noise and ensure data quality. Each group of samples contains 1,000 sets of voice dispatch data, and a total of four samples are generated.

[0066] Table 1 Experimental data table

[0067]

[0068]

[0069] First, in terms of voice command recognition rate, the average recognition rate of the prior art is about 84.75%, while the average recognition rate of the present invention reaches 94.5%, which is an increase of about 9.75 percentage points. This shows that the present invention has higher recognition accuracy and can more effectively recognize power dispatching instructions. Its main effect has been basically achieved based on more effective voice feature extraction.

[0070] Secondly, in terms of false alarm rate, the prior art has a high false alarm rate, averaging about 10.5%, while the false alarm rate of the present invention is significantly reduced, averaging only 3%. This shows that the subsequent scheduling strategy recognition technology avoids the implementation of erroneous instructions. The present invention can effectively reduce false alarms and improve the reliability and security of the system in practical applications.

[0071] Example 3

[0072] Reference Figure 2 , as an embodiment of the present invention, provides an electric power dispatching operation instruction safety management system based on AI recognition, including a data acquisition module, an instruction extraction module, and an instruction confirmation module.

[0073] Among them, the data acquisition module is used to collect data and extract features from the data; the instruction extraction module is used to recognize voice operation instructions based on a multi-model fusion speech recognition model; and the instruction confirmation module is used to classify instruction sequences and perform instruction confirmation.

[0074] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0075] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0076] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0077] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for safe management of power dispatching operation instructions based on AI recognition, characterized in that: include: Collect data and extract features from the data; A speech recognition model based on multi-model fusion recognizes voice operation instructions; Classify instruction sequences and perform instruction confirmation.

2. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 1, characterized in that: The data collection includes voice data collection, which captures the dispatch instruction voice in real time through a microphone array, denoises the voice, and extracts features after processing.

3. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 2, characterized in that: The data recognition processing includes recognizing the voice information, extracting the feature vector obtained by the feature extraction of the voice signal, dividing the voice signal into multiple small frames, each frame length is 20-40 milliseconds, and each frame x t It is expressed as: x t =[x(t),x(t+1),…,x(t+N-1)] Where N is the number of samples per frame; Multiply each frame signal by the Hamming window function, expressed as: x′ t =x t ·w(n) Where w(n) is the Hamming window function; Perform fast Fourier transform on each frame of processed signal to convert the time domain signal to the frequency domain, and divide the frequency domain signal into several Mel frequency bands through a set of Mel filters, which can be expressed as: Among them, X t (k) is the signal after fast Fourier transform, H m (k) is the frequency response of the mth Mel filter.

4. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 3, characterized in that: The data recognition processing also includes performing discrete cosine transform on the energy of the logarithmic Mel frequency band to obtain Mel frequency cepstrum coefficients MFCC, which represent the feature vector of the voice scheduling instruction signal, expressed as: Among them, O t (n) is the nth MFCC coefficient of the tth frame, and M represents the number of Mel frequency bands.

5. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 4, characterized in that: The multi-model fusion-based speech recognition model recognizes speech operation instructions, including processing sequence data through a recurrent neural network GUR, assigning weights to the feature vectors of each time step through an attention mechanism, and outputting the weighted sum as the final feature representation sequence y of the operation instruction.

6. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 5, characterized in that: The classification of the instruction sequence includes classifying the operation instruction sequence by similarity, extracting data points from the sequence y for input, which is expressed as: in, is the instruction label after classification, α j and b are the parameters of the support vector machine, K(x,x j ) is the kernel function, which represents the data point x and the support vector x j The similarity between .

7. The method for safe management of power dispatching operation instructions based on AI recognition according to claim 6, characterized in that: The instruction confirmation includes maximizing the posterior of the voice instruction after completing the instruction classification to detect whether the voice instruction is correct, which is expressed as: Among them, C is the final confirmed result, C′ represents the candidate data, and P(c i |C′,y t ) is the confirmation probability of a single operator, P(C′) is the prior probability, λ is the attenuation coefficient of the exponential function, μ k and σ 2 are the mean and variance of the Gaussian distribution respectively; The results are displayed to the on-duty personnel, and after the command is confirmed, the system begins to execute the voice command.

8. A system using the power dispatching operation instruction safety management method based on AI recognition as claimed in any one of claims 1 to 7, characterized in that: It includes data acquisition module, instruction extraction module and instruction confirmation module; The data acquisition module is used to collect data and extract features from the data; The instruction extraction module is used to recognize voice operation instructions based on a multi-model fusion speech recognition model; The instruction confirmation module is used to classify instruction sequences and perform instruction confirmation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the power dispatching operation instruction safety management method based on AI identification described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power dispatching operation instruction safety management method based on AI recognition described in any one of claims 1 to 7 are implemented.

Images