Vehicle-mounted voice AI intelligent interaction system and device

Abstract

The invention discloses a vehicle-mounted voice AI intelligent interaction system and device, relates to the technical field of vehicle-mounted voice AI, and solves the problems that noise associated with a voice acquisition process cannot be removed and the specific accuracy of voice recognition cannot be guaranteed due to the fact that voice features of a wake-up instruction are not confirmed. According to the method, the processing of the wake-up sound signal is refined, the signal energy is confirmed firstly, the short-time energy is calculated by dividing the short time period and summed, and the sound energy characteristics are accurately measured; determining a correlation frequency, converting from a time domain to a frequency domain by using discrete Fourier transform, and extracting key frequency information; the double analysis of the energy and the frequency of the sound signal provides rich and key characteristic data for subsequent processing.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-mounted voice AI technology, and specifically to a vehicle-mounted voice AI intelligent interaction system and device. Background Art

[0002] In-car voice AI is an innovative application of artificial intelligence technology in the automotive field. It makes driving smarter, safer and more convenient through voice interaction. The driver only needs to say the destination, and the system can automatically plan the best route. It can also provide real-time traffic information, avoid congestion, and dynamically adjust the route according to traffic conditions. It can play music, radio, audiobooks, etc., and switch songs, adjust the volume, and search for what you want to listen to through voice commands, so that you can continue to entertain while driving;

[0003] The application with publication number CN116564302A discloses an intelligent voice interaction system based on AI autonomous learning, including an AI intelligent voice interactor, the output end of the AI ​​intelligent voice interactor is electrically connected to a starting power supply, the output end of the starting power supply is electrically connected to a central processing unit, the output end at the bottom of the central processing unit is bidirectionally electrically connected to a voice acquisition unit, the output end at the bottom of the voice acquisition unit is bidirectionally electrically connected to a data processing module, the output end at the bottom of the data processing module is bidirectionally electrically connected to a voice recognition unit, the output end at the bottom of the voice recognition unit is bidirectionally electrically connected to a computer understanding unit, and the output end on the right side of the computer understanding unit is bidirectionally electrically connected to a user identity determination module; the invention adds the functionality of high-intensity autonomous learning, so that it has the effect of AI autonomous learning of the intelligent voice interaction system, and prevents the occurrence of poor AI autonomous learning ability of the intelligent voice interaction system.

[0004] During the intelligent interaction process, after the corresponding voice AI is awakened, the influence of other noises will cause interference in the recognition of the in-vehicle voice AI, resulting in the corresponding in-vehicle voice AI recognizing content with low accuracy or failing to recognize specific voice commands. The sound characteristics of the wake-up command are not confirmed, resulting in the inability to remove the noise associated with the sound collection process, and the specific accuracy of sound recognition cannot be guaranteed. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an in-vehicle voice AI intelligent interaction system and device, which solves the problem that the sound characteristics of the wake-up command are not confirmed, resulting in the inability to remove the noise associated with the sound collection process and the inability to ensure the specific accuracy of sound recognition.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an in-vehicle voice AI intelligent interaction system, comprising:

[0007] The wake-up signal monitoring end monitors the wake-up sound signal set by the in-vehicle voice intelligent AI in real time, and executes the subsequent sound feature processing end and sound signal acquisition end based on the monitored wake-up sound signal;

[0008] The sound feature processing end performs feature processing on the wake-up sound signal detected by the vehicle-mounted voice intelligent AI, first confirms the signal energy of the corresponding wake-up sound signal, then confirms the relevant frequency of the corresponding wake-up sound signal, and transmits the confirmed signal energy and the associated relevant frequency to the single-frequency feature confirmation end;

[0009] The single-frequency feature confirmation end confirms the single-frequency feature associated with the corresponding wake-up sound signal based on the signal energy and related frequency confirmed by the corresponding wake-up sound signal, and performs interval confirmation on the confirmed single-frequency feature, and transmits the confirmed single-frequency feature interval to the sound filtering processing end, specifically in the following manner:

[0010] The signal energy confirmed by the corresponding wake-up sound signal is calibrated as XN i , and mark the confirmed relevant frequency as PL i , where i represents different wake-up sound signals, using D i =XN i ÷PL i Confirm single frequency feature D i ;

[0011] Based on the different single frequency features D associated with different wake-up sound signals i , select D from several groups of different single frequency characteristics i min and D i max, confirm the single frequency feature interval, and transmit the confirmed single frequency feature interval to the sound filtering processing end;

[0012] The sound signal acquisition end, after the in-vehicle voice intelligent AI is awakened, collects the surrounding sounds and transmits the real-time collected sounds to the sound filtering and processing end. If no sound is collected within a certain period of time, the collection process is stopped. The certain period of time is a preset period of time.

[0013] The sound filtering and processing end processes the sounds collected from the surrounding area based on the confirmed single-frequency characteristic interval, confirms the signal energy and related frequency associated with the corresponding sound signal, and locks the sound characteristics of the corresponding sound signal. Based on the determined sound characteristics, the collected sound is filtered, the characteristic sound is confirmed, and the confirmed characteristic sound is transmitted to the text processing control end. The specific method is as follows:

[0014] The collected sound is processed in the same manner as the sound feature processing end, and the single frequency feature of the corresponding sound signal is confirmed based on the signal energy and related frequencies associated with different sound signals in the corresponding sound;

[0015] The sound signal of the single-frequency feature ∈ single-frequency feature interval is recorded as the characteristic signal, and the single-frequency feature The sound signal in the single-frequency characteristic interval is eliminated;

[0016] Based on the characteristic signals confirmed in sequence, the sound associated with the characteristic signal is recorded as the characteristic sound, and the characteristic sound is reordered according to the time characteristics before and after the characteristic sound, and the reordered characteristic sound is transmitted to the text processing control terminal.

[0017] Preferably, the sound feature processing end confirms the signal energy of the corresponding wake-up sound signal in a specific manner as follows:

[0018] The wake-up sound signal is divided into several short periods, each of which is a preset period, using: Confirm the short-term energy E associated with the corresponding short period n n , where X(m) represents the original wake-up sound signal, which is a discrete time series, and m is a discrete time variable, which represents the amplitude of the sound signal at different discrete time points; the sound signal is continuous in the time domain, but in digital signal processing, it needs to be sampled and quantized to convert it into a discrete digital signal. X(m) is such a discrete digital sound signal, which contains all the time domain information of the sound. W(nm) is a window function, which divides the sound signal into short time periods for analysis. n is the center position of the short time period currently concerned, and m is a discrete time variable. W(nm) means that the signals at different positions m are weighted in the short time period centered on n, where ∞ represents infinity, and m=-∞ represents that m starts from negative infinity and sums to positive infinity;

[0019] The short-term energy E confirmed by several short time periods n The sum is performed to confirm the signal energy belonging to the wake-up sound signal.

[0020] Preferably, the sound feature processing end confirms the relevant frequency of the corresponding wake-up sound signal in a specific manner:

[0021] Determine a sound signal sequence L(q) corresponding to the wake-up sound signal, where q=0, 1, ..., N-1, according to which, during the signal acquisition process, the continuous sound signal is sampled and quantized to obtain a set of discrete digital sequences;

[0022] use Confirm the frequency domain sequence X(k), where k = 0, 1, ..., N-1, where in and is the rotation factor, different combinations of q and k will produce different weighting coefficients;

[0023] Based on the different X(k)s corresponding to different k values, select the largest X(k) and substitute the associated k value into Identify the relevant frequency of this wake-up sound signal, where f s is the sampling frequency;

[0024] The signal energy and related frequency confirmed by the corresponding wake-up sound signal are transmitted to the single-frequency feature confirmation terminal.

[0025] Preferably, it also includes:

[0026] The text processing control terminal processes the received characteristic sounds, confirms the characteristic texts associated with several groups of characteristic sounds, and proofreads the confirmed characteristic texts with the execution texts pre-stored in the cloud library. Based on the proofreading results, the execution instruction is confirmed. The specific method is as follows:

[0027] Confirm the characteristic texts associated with the plurality of characteristic sounds, and collate the characteristic texts with the pre-stored execution texts: mark the same type of text characters at the same positions in the characteristic texts and the execution texts as the same type of characters, and confirm the proportion of the same type of characters in the total number of characters in the execution texts;

[0028] The execution text with the highest percentage during proofreading is confirmed, the confirmed execution text is used as an execution instruction, and the designated execution part is controlled based on the confirmed execution instruction.

[0029] The present invention provides a vehicle-mounted voice AI intelligent interaction system and device. Compared with the prior art, it has the following beneficial effects:

[0030] The present invention processes the wake-up sound signal in detail through the sound feature processing end. First, the signal energy is confirmed, and the short-time energy is calculated and summed by dividing the short time period to accurately measure the sound energy characteristics; then the relevant frequency is determined, and the discrete Fourier transform is used to convert from the time domain to the frequency domain to extract the key frequency information; this dual analysis of the sound signal energy and frequency provides rich and key feature data for subsequent processing;

[0031] The single-frequency feature confirmation end determines the single-frequency feature based on signal energy and frequency, and cleverly integrates the single-frequency features of different wake-up sound signals into a valid interval through interval confirmation; the sound filtering processing end processes the collected sound based on this interval in the same way as the sound feature processing end, accurately filters out characteristic sounds, and eliminates noise, greatly improving the purity of the sound signal, significantly improving the accuracy of intelligent AI sound reading, and reducing the impact of external noise interference on the interaction effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] First embodiment

[0035] See also Figure 1 The present application provides an in-vehicle voice AI intelligent interaction system, including: a wake-up signal monitoring end, a sound feature processing end, a single-frequency feature confirmation end, a sound signal acquisition end, a sound filtering processing end, and a text processing control end. The wake-up signal monitoring end, the sound feature processing end, and the single-frequency feature confirmation end are electrically connected from the output node to the input node in sequence, and the sound signal acquisition end and the single-frequency feature confirmation end are both electrically connected to the sound filtering processing end input node, and the sound filtering processing end is electrically connected to the text processing control end input node;

[0036] Among them, the wake-up signal monitoring end monitors the wake-up sound signal set by the vehicle-mounted voice intelligent AI in real time, and executes the subsequent sound feature processing end and the sound signal acquisition end based on the monitored wake-up sound signal. The wake-up sound signal is the wake-up command of a specific AI, and the corresponding AI can recognize it by itself. For example: Hello, so-and-so, the intelligent AI will answer automatically;

[0037] Its sound feature processing end performs feature processing on the wake-up sound signal detected by the vehicle-mounted voice intelligent AI, first confirms the signal energy of the corresponding wake-up sound signal, then confirms the relevant frequency of the corresponding wake-up sound signal, and transmits the confirmed signal energy and the associated relevant frequency to the single-frequency feature confirmation end. Specifically, a series of voice commands has a single voice character, and the single group of sounds corresponding to such voice characters belongs to a single group of sound signals, and each group of sound signals has different signal energies and associated signal frequencies;

[0038] The specific method of confirming the signal energy of the corresponding wake-up sound signal is as follows:

[0039] The wake-up sound signal is divided into several short periods, each of which is a preset period, which is set in advance by the relevant operator, using: Confirm the short-term energy E associated with the corresponding short period n n , where X(m) represents the original wake-up sound signal, which is a discrete time series, and m is a discrete time variable, which represents the amplitude of the sound signal at different discrete time points. The sound signal is continuous in the time domain, but in digital signal processing, it needs to be sampled and quantized to convert it into a discrete digital signal. X(m) is such a discrete digital sound signal, which contains all the time domain information of the sound. W(nm) is the window function, which is used to divide the sound signal into short time periods for analysis. n is the center position of the short time period currently concerned, and m is a discrete time variable. W(nm) means that in the short time period centered on n, the signals at different positions m are weighted. Common window functions include Hanming window, rectangular window, Haining window, etc. Different window functions have different shapes and characteristics, which will have a certain impact on the calculation results of short-time energy. The window function is selected in advance by the relevant operator, where ∞ represents infinity, and m=-∞ represents that m starts from negative infinity and sums to positive infinity. In actual calculations, since the sound signal is of finite length and is processed in frames, the range of summation is usually within the length range of a frame of signal currently processed.

[0040] The short-term energy E confirmed by several short time periods n Perform summation to confirm the signal energy belonging to the wake-up sound signal;

[0041] The specific method of determining the relevant frequency corresponding to the wake-up sound signal is as follows:

[0042] Determine a sound signal sequence L(q) corresponding to the wake-up sound signal, where q=0, 1, ..., N-1, according to which, during the signal acquisition process, the continuous sound signal is sampled and quantized to obtain a set of discrete digital sequences;

[0043] use Confirm the frequency domain sequence X(k), where k = 0, 1, ..., N-1, where in and is the rotation factor. Different combinations of q and k will produce different weighting coefficients. Its essence is a complex exponential signal. With the change of q and k, it exhibits the characteristics of rotation on the complex plane. Through this rotation, the transformation from time domain to frequency domain is realized, and the information of different frequency components is extracted.

[0044] Based on the different X(k)s corresponding to different k values, select the largest X(k) and substitute the associated k value into Identify the relevant frequency of this wake-up sound signal, where f s is the sampling frequency;

[0045] The signal energy and related frequency confirmed by the corresponding wake-up sound signal are transmitted to the single-frequency feature confirmation terminal.

[0046] Specifically, each group of sound characters corresponds to a group of sound signals, and each group of sound signals is associated with corresponding signal energy and related frequencies. Based on the confirmed signal energy and related frequencies, the single-frequency characteristics associated with the corresponding sound signals are confirmed.

[0047] Among them, the single-frequency feature confirmation end confirms the single-frequency feature associated with the corresponding wake-up sound signal based on the signal energy and related frequency confirmed by the corresponding wake-up sound signal, and performs interval confirmation on the confirmed single-frequency feature, and transmits the confirmed single-frequency feature interval to the sound filtering processing end, wherein the specific method of interval confirmation is:

[0048] The signal energy confirmed by the corresponding wake-up sound signal is calibrated as XN i , and mark the confirmed relevant frequency as PL i , where i represents different wake-up sound signals, using D i =XN i ÷PL i Confirm single frequency feature D i ;

[0049] Based on the different single frequency features D associated with different wake-up sound signals i , select D from several groups of different single frequency characteristics i min and D i max, confirm the single frequency feature interval, and transmit the confirmed single frequency feature interval to the sound filtering processing end;

[0050] Specifically, such single-frequency feature intervals are the voice features of the corresponding persons. Based on such voice features, the corresponding noise can be effectively eliminated to achieve better voice filtering effects. The corresponding voice filtering processing end filters the voices associated with the relevant stages.

[0051] Among them, the sound signal acquisition end, after the in-vehicle voice intelligent AI is awakened, collects the surrounding sounds and transmits the real-time collected sounds to the sound filtering and processing end. If no sound is collected within a certain period of time, the collection process is stopped. The certain period of time is a preset period of time, which is formulated by relevant operators based on experience and is generally 3 seconds.

[0052] Among them, the sound filtering and processing end processes the sounds collected from the surrounding area based on the confirmed single-frequency characteristic interval, confirms the signal energy and related frequency associated with the corresponding sound signal, and locks the sound characteristics of the corresponding sound signal. Based on the determined sound characteristics, the collected sound is filtered, the characteristic sound is confirmed, and the confirmed characteristic sound is transmitted to the text processing control end, wherein the specific method of confirming the characteristic sound is:

[0053] The collected sound is processed in the same manner as the sound feature processing end, and the single frequency feature of the corresponding sound signal is confirmed based on the signal energy and related frequencies associated with different sound signals in the corresponding sound;

[0054] The sound signal of the single-frequency feature ∈ single-frequency feature interval is recorded as the characteristic signal, and the single-frequency feature The sound signal in the single-frequency characteristic interval is eliminated;

[0055] Based on the characteristic signals confirmed in sequence, the sounds associated with the characteristic signals are recorded as characteristic sounds, and the characteristic sounds are reordered according to the time characteristics before and after the characteristic sounds, and the reordered characteristic sounds are transmitted to the text processing control terminal;

[0056] Specifically, if the sound characteristics associated with the determined characteristic sound are relatively consistent with the wake-up characteristics in the corresponding wake-up process, then such characteristic sound and the related wake-up characteristics all come from the same sound source. Therefore, after the associated characteristic sound is confirmed, a better noise filtering effect can be obtained, thereby improving the specific accuracy of the intelligent AI sound reading process.

[0057] The text processing control terminal processes the received characteristic sounds, confirms the characteristic texts associated with the characteristic sounds, and compares the confirmed characteristic texts with the execution texts pre-stored in the cloud library. Based on the comparison results, the execution instruction is confirmed. The specific method of confirming the execution instruction is as follows:

[0058] Confirm the characteristic texts associated with the plurality of characteristic sounds, and collate the characteristic texts with the pre-stored execution texts: mark the same type of text characters at the same positions in the characteristic texts and the execution texts as the same type of characters, and confirm the proportion of the same type of characters in the total number of characters in the execution texts;

[0059] Confirm the execution text with the highest proportion during proofreading, use the confirmed execution text as the execution instruction, and control the specified execution part based on the confirmed execution instruction to complete the corresponding execution process. During consistency proofreading, if the same text parameters exist, when the feature text and the execution text proofreading are completely consistent, the corresponding execution instruction can be directly determined.

[0060] Second embodiment

[0061] An in-vehicle voice AI intelligent interaction device, comprising:

[0062] Audio playback module: supports external audio playback via Bluetooth, supports external audio playback via AUX, and supports simultaneous playback of Bluetooth and AUX;

[0063] Audio input module: voice input and wake-up are performed directly through the device's MIC; due to the delay of several hundred milliseconds when using Bluetooth external speakers, the microphone should be able to normally eliminate the voice input caused by this delay; weak sound input can still be recognized, such as in drawers, armrest boxes, etc.;

[0064] Network module: Supports plug-and-play Internet access services based on Cat1 for Android devices (may involve driver debugging for specific device solutions);

[0065] Indicator light: Provides a multi-color indicator light to display the following four states: network connected, network normal, no network, and device failure;

[0066] SIM card reading area: The board supports both SMD and plug-in cards. If a plug-in card exists, the plug-in card takes priority. If no plug-in card is available, the preset SMD card is used. The plug-in card uses a Nano pop-up plug-in card slot.

[0067] Control module: The device does not support turning off the microphone, it can only be powered off; the device supports manual pause, which can only be paused by voice; the device does not support volume adjustment; the device does not support hardware-side reset.

[0068] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0069] The above embodiments are only used to illustrate the technical method 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 method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A vehicle-mounted voice AI intelligent interaction system, characterized in that: include: The wake-up signal monitoring end monitors the wake-up sound signal set by the in-vehicle voice intelligent AI in real time, and executes the subsequent sound feature processing end and sound signal acquisition end based on the monitored wake-up sound signal; The sound feature processing end performs feature processing on the wake-up sound signal detected by the vehicle-mounted voice intelligent AI, first confirms the signal energy of the corresponding wake-up sound signal, then confirms the relevant frequency of the corresponding wake-up sound signal, and transmits the confirmed signal energy and the associated relevant frequency to the single-frequency feature confirmation end; The single-frequency feature confirmation end confirms the single-frequency feature associated with the corresponding wake-up sound signal based on the signal energy and related frequency confirmed by the corresponding wake-up sound signal, and performs interval confirmation on the confirmed single-frequency feature, and transmits the confirmed single-frequency feature interval to the sound filtering processing end; The sound signal acquisition end, after the in-vehicle voice intelligent AI is awakened, collects the surrounding sounds and transmits the real-time collected sounds to the sound filtering and processing end. If no sound is collected within a certain period of time, the collection process is stopped. The certain period of time is a preset period of time. The sound filtering and processing end processes the sounds collected from the surrounding area based on the confirmed single-frequency feature interval, confirms the signal energy and related frequency associated with the corresponding sound signal, to lock the sound characteristics of the corresponding sound signal, filters the collected sounds based on the determined sound characteristics, confirms the characteristic sounds, and transmits the confirmed characteristic sounds to the text processing control end.

2. The vehicle-mounted voice AI intelligent interaction system according to claim 1, characterized in that: The specific method of the sound feature processing end to confirm the signal energy of the corresponding wake-up sound signal is: The wake-up sound signal is divided into several short periods, each of which is a preset period, using: Confirm the short-term energy E associated with the corresponding short period n n , where X(m) represents the original wake-up sound signal, which is a discrete time series, and m is a discrete time variable, which represents the amplitude of the sound signal at different discrete time points; the sound signal is continuous in the time domain, but in digital signal processing, it needs to be sampled and quantized to convert it into a discrete digital signal. X(m) is such a discrete digital sound signal, which contains all the time domain information of the sound. W(nm) is a window function, which divides the sound signal into short time periods for analysis. n is the center position of the short time period currently concerned, and m is a discrete time variable. W(nm) means that the signals at different positions m are weighted in the short time period centered on n, where ∞ represents infinity, and m=-∞ represents that m starts from negative infinity and sums to positive infinity; The short-term energy E confirmed by several short time periods n The sum is performed to confirm the signal energy belonging to the wake-up sound signal.

3. The vehicle-mounted voice AI intelligent interaction system according to claim 1, characterized in that: The specific method of the sound feature processing end to confirm the relevant frequency of the corresponding wake-up sound signal is: Determine a sound signal sequence L(q) corresponding to the wake-up sound signal, where q=0, 1, ..., N-1, according to which, during the signal acquisition process, the continuous sound signal is sampled and quantized to obtain a set of discrete digital sequences; use Confirm the frequency domain sequence X(k), where k = 0, 1, ..., N-1, where in and is the rotation factor, different combinations of q and k will produce different weighting coefficients; Based on the different X(k)s corresponding to different k values, select the largest X(k) and substitute the associated k value into Identify the relevant frequency of this wake-up sound signal, where f s is the sampling frequency; The signal energy and related frequency confirmed by the corresponding wake-up sound signal are transmitted to the single-frequency feature confirmation terminal.

4. The vehicle-mounted voice AI intelligent interaction system according to claim 1, characterized in that: The specific manner in which the single-frequency feature confirmation terminal performs interval confirmation on the confirmed single-frequency feature is: The signal energy confirmed by the corresponding wake-up sound signal is calibrated as XN i , and mark the confirmed relevant frequency as PL i , where i represents different wake-up sound signals, using D i =XN i ÷PL i Confirm single frequency feature D i ; Based on the different single frequency features D associated with different wake-up sound signals i , select D from several groups of different single frequency characteristics i min and D i max, confirm the single-frequency feature interval, and transmit the confirmed single-frequency feature interval to the sound filtering processing end.

5. The vehicle-mounted voice AI intelligent interaction system according to claim 1, characterized in that: The specific method of confirming the characteristic sound at the sound filtering and processing end is: The collected sound is processed in the same manner as the sound feature processing end, and the single frequency feature of the corresponding sound signal is confirmed based on the signal energy and related frequencies associated with different sound signals in the corresponding sound; The sound signal with single-frequency feature ∈ single-frequency feature interval is recorded as feature signal; Based on the characteristic signals confirmed in sequence, the sound associated with the characteristic signal is recorded as the characteristic sound, and the characteristic sound is reordered according to the time characteristics before and after the characteristic sound, and the reordered characteristic sound is transmitted to the text processing control terminal.

6. The vehicle-mounted voice AI intelligent interaction system according to claim 5, characterized in that: The single frequency feature The sound signal in the single-frequency feature interval is eliminated.

7. The vehicle-mounted voice AI intelligent interaction system according to claim 5, characterized in that: Also includes: The text processing control end processes the received characteristic sounds, confirms the characteristic texts associated with several groups of characteristic sounds, and proofreads the confirmed characteristic texts with the execution texts pre-stored in the cloud library, and confirms the execution instructions based on the proofreading results.

8. The vehicle-mounted voice AI intelligent interaction system according to claim 7, characterized in that: The specific method for the text processing control terminal to confirm the execution of the instruction is: Confirm the characteristic texts associated with the plurality of characteristic sounds, and collate the characteristic texts with the pre-stored execution texts: mark the same type of text characters at the same positions in the characteristic texts and the execution texts as the same type of characters, and confirm the proportion of the same type of characters in the total number of characters in the execution texts; The execution text with the highest percentage during proofreading is confirmed, the confirmed execution text is used as an execution instruction, and the designated execution part is controlled based on the confirmed execution instruction.

Images