A method and system for intelligent control of automobile cockpit equipment based on voice recognition

Abstract

The present invention discloses a method and system for intelligent control of automobile cockpit equipment based on speech recognition, which relates to the field of intelligent control of automobile cockpit equipment, including: acquiring multi-source audio signals, analyzing the spectrum overlap between speech signals and noise signals and the phase interference between speech signals, and evaluating the current speech recognition interference level; categorizing and setting initial priorities for multi-source speech instructions, and executing multi-source speech instructions in order of instruction priority when the system is in a normal speech recognition state; identifying, classifying and asynchronously processing multi-source speech instructions when the system is in a speech recognition interference state, and dynamically adjusting the instruction execution priority weights in an asynchronous processing channel; dynamically planning system resource allocation based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel, and determining the specific execution priority order of the instructions in the asynchronous processing channel.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control of automobile cockpit equipment, and in particular to an intelligent control method and system of automobile cockpit equipment based on voice recognition. Background Art

[0002] In the car cockpit, voice recognition is used to control various devices, but the recognition accuracy is often affected by noise in noisy environments. In addition, different voice commands have different priorities, especially the conflicts between driving-related commands and entertainment commands. Traditional voice control systems lack dynamic recognition of command priorities and are difficult to adjust in real time according to the driver's identity and scenario. Therefore, how to set command priorities based on driving-related keywords, identity recognition, and scene matching while identifying noise levels is the key to improving the intelligent experience of the cockpit and ensuring driving safety.

[0003] Therefore, how to assign different priorities to different types of commands by identifying the noise level, identity confidence and scene adaptability of the voice signal, and realize a method based on multi-dimensional command priority recognition and applied to intelligent control of automotive cockpit equipment to meet the requirements of safe, efficient and personalized automotive cockpit control is a problem that needs to be solved. Summary of the invention

[0004] In order to solve the above technical problems, a method for intelligent control of automobile cockpit equipment based on voice recognition is provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for intelligently controlling vehicle cockpit equipment based on speech recognition, comprising:

[0007] Acquire multi-source audio signals, analyze the spectral overlap between speech signals and noise signals and the phase interference between speech signals, and evaluate the current speech recognition interference level;

[0008] Classify and initially prioritize multi-source voice commands. When the system is in a normal voice recognition state, execute multi-source voice commands in the order of command priority.

[0009] When the system is in a state of voice recognition interference, it recognizes, classifies and asynchronously processes multi-source voice commands, and dynamically adjusts the execution priority weight of commands in the asynchronous processing channel;

[0010] Based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel, the system resource allocation is dynamically planned to determine the specific execution priority order of the instructions in the asynchronous processing channel.

[0011] Preferably, the acquiring of multi-source audio signals, analyzing the spectrum overlap between the speech signal and the noise signal and the phase interference between the speech signals, and evaluating the current speech recognition interference level specifically includes:

[0012] Obtain real-time multi-source audio signals and set the frequency range of voice signals and noise signals;

[0013] The audio signal in the frequency range of the speech signal is marked as a speech signal, and the audio signal in the frequency range of the noise signal is marked as a noise signal;

[0014] Perform spectrum analysis on speech signal and noise signal through fast Fourier transform to obtain spectrum density in time and frequency;

[0015] Sampling multiple time periods, processing the spectrum information in each time period, and calculating the minimum spectrum density value of the time series speech signal and noise signal in each frequency band;

[0016] The dynamic spectrum overlap rate is evaluated by double integration of time and frequency, the overlapping spectrum is calculated in each time segment, and the calculation result is averaged over the entire time period to obtain the comprehensive spectrum overlap metric;

[0017] Through phase difference analysis, the phase relationship between multiple sound sources is identified and the impact of phase interference of multiple sound sources is quantified;

[0018] The phase information of multiple sound source signals is obtained by short-time Fourier transform, the phase value of each sound source on the time statistical scale is calculated, the cosine of the phase difference corresponding to each sound source is taken and integrated, and the average value of the integral calculation result is marked as the multi-source phase interference coefficient;

[0019] Based on the comprehensive spectrum overlap measurement and multi-source phase interference coefficient, the weighted phase superposition interference index is calculated comprehensively. The specific expression is: ;

[0020] In the formula, is the phase superposition interference index, is the comprehensive spectrum overlap measure, is the multi-source phase interference coefficient, for and The phase angle relationship between them is in the range of ;

[0021] Set the phase superposition interference index threshold. When the calculated phase superposition interference index is less than the phase superposition interference index threshold, it indicates that the system is in a normal speech recognition state. When the calculated phase superposition interference index is greater than or equal to the phase superposition interference index threshold, the system state is changed to a speech recognition interference state.

[0022] Preferably, the categorization and initial priority setting of the multi-source voice commands, when the system is in a normal voice recognition state, executing the multi-source voice commands in the order of command priority specifically includes:

[0023] The voice commands executable by cockpit equipment are classified into command types, with voice commands related to driving assistance and driving safety classified as Class I commands, and voice commands related to infotainment and environmental control classified as Class II commands;

[0024] Setting the first priority weight for all instructions in the instruction set of the first category of instructions and the second category of instructions, wherein the first priority weight of the first category of instructions must be higher than the first priority weight of the second category of instructions;

[0025] When it is detected that the current system is in a normal voice recognition state, the multi-source voice commands received by the system are executed sequentially according to the time sequence of command reception and the first priority weight sequence assigned to the commands.

[0026] Preferably, when the system is in a speech recognition interference state, the multi-source speech commands are recognized, classified and asynchronously processed, and the execution priority weight of the commands in the asynchronous processing channel is dynamically adjusted, specifically including:

[0027] When the system is in a speech recognition interference state, the received multi-source voice commands are split and processed;

[0028] The system identifies and distinguishes the types of commands to which voice commands belong, and uses a multi-threaded processing framework to place the first and second types of commands in different asynchronous processing channels, and analyzes each asynchronous processing channel separately through asynchronous processing strategies;

[0029] The driving scene is determined based on the current real-time driving parameters of the car, and the real-time driving score of the current driving scene is calculated. The ideal driving score is set based on the ideal driving parameters of the driving scene.

[0030] Analyze the execution frequency and effect of the current instruction in historical driving scenarios, and dynamically evaluate the driving scenario matching degree of the instruction based on the deviation between the driving score and the ideal driving score. The driving scenario matching degree of the instruction is calculated as follows: ;

[0031] In the formula, , are the actual driving score and ideal driving score of the current scene after the kth instruction is executed, is the effect evaluation value after the execution of the kth instruction, is the execution frequency weight of the kth instruction, is the number of instructions in the history, is the real-time driving scene matching degree of the kth instruction;

[0032] The voice feature matching model is used to identify the identity information of the command initiator of the voice, and the command initiation confidence of the current voice command is evaluated based on the identity information recognition result and the real-time signal-to-noise ratio of the current voice information;

[0033] The execution priority weight of the command in the asynchronous processing channel is dynamically set according to the real-time driving scene matching degree of the command and the confidence degree of the command initiation, and marked as the second priority weight of the command.

[0034] Preferably, the dynamically planning system resource allocation based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel to determine the specific execution priority order of the instructions in the asynchronous processing channel specifically includes:

[0035] Obtaining the first priority weight and the second priority weight of the current multi-source voice command, normalizing the scales of the first priority weight and the second priority weight, performing weighted summation, and using the calculation result as the final execution priority weight of the command;

[0036] Get the total amount of available resources for system voice command processing;

[0037] Obtain all voice commands in the asynchronous processing channel, and dynamically allocate available system resources in proportion to the final execution priority weight of each voice command relative to the total final execution priority weight of all voice commands;

[0038] Based on the system resources planned and allocated for each instruction, a priority execution queue is set in the asynchronous processing channel, and the instructions are added to the execution queue in sequence according to the final execution priority weight of the instructions and executed sequentially through system scheduling.

[0039] Furthermore, a vehicle cockpit equipment intelligent control system based on voice recognition is proposed, which is used to implement the above-mentioned vehicle cockpit equipment intelligent control method based on voice recognition, including:

[0040] A speech recognition interference level assessment module, which comprehensively calculates a weighted phase superposition interference index based on the analysis of the spectrum overlap between the speech signal and the noise signal and the phase interference between the speech signals, and updates the system state to a normal speech recognition state or a speech recognition interference state in real time based on the calculation result of the phase superposition interference index;

[0041] A command priority calculation module, which divides the voice commands executable by the cockpit device into command types, determines the initial priority weights of the commands and calculates the asynchronous processing execution priority weights, comprehensively calculates the initial priority settings of the commands and the dynamic command execution priority weights in the asynchronous processing channel, and determines the specific execution priority order of the commands in the asynchronous processing channel;

[0042] A system resource dynamic allocation module, which dynamically allocates available system resources in proportion to all voice commands in the asynchronous processing channel based on the proportion of the final execution priority weight of each voice command relative to the total final execution priority weights of all voice commands;

[0043] An instruction scheduling execution module automatically selects an instruction scheduling execution strategy according to the current system state. When it is detected that the current system is in a normal speech recognition state, the system receives multi-source speech instructions and executes the instructions sequentially according to the time sequence of instruction reception and the first priority weight sequence assigned to the instructions. When the system is in a speech recognition interference state, the execution is scheduled in the execution queue sequence of the asynchronous processing channel.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] Analyze the spectral overlap between the voice signal and the noise signal and the phase interference between the voice signals to evaluate the current voice recognition interference level. Categorize and set initial priorities for multi-source voice commands. When the system is in a normal voice recognition state, execute multi-source voice commands in the order of command priority. When the system is in a voice recognition interference state, identify and classify multi-source voice commands and asynchronously process them, and dynamically adjust the command execution priority weights in the asynchronous processing channel. Dynamically plan system resource allocation based on the initial priority setting of the commands and the dynamic command execution priority weights in the asynchronous processing channel to determine the specific execution priority order of the commands in the asynchronous processing channel.

[0046] By identifying the noise level, identity confidence and scene adaptability of the voice signal, different priorities are assigned to different types of commands, realizing multi-dimensional command priority recognition, which can be applied to the intelligent control of automotive cockpit equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of an intelligent control method for vehicle cockpit equipment based on speech recognition according to the present invention;

[0048] Figure 2A flowchart of obtaining multi-source audio signals, analyzing the spectrum overlap between speech signals and noise signals and the phase interference between speech signals, and evaluating the current speech recognition interference level in the present invention;

[0049] Figure 3 The present invention is to classify and initially prioritize multi-source voice commands, and when the system is in a normal voice recognition state, execute the multi-source voice command flow chart in the order of command priority;

[0050] Figure 4 The flow chart of the present invention is to identify, classify and asynchronously process multi-source voice commands and dynamically adjust the execution priority weight of commands in the asynchronous processing channel when the system is in a voice recognition interference state;

[0051] Figure 5 A flowchart of the present invention for dynamically planning system resource allocation based on the initial priority setting of instructions and the dynamic instruction execution priority weights in the asynchronous processing channel, and determining the specific execution priority sequence of instructions in the asynchronous processing channel;

[0052] Figure 6 The present invention is a schematic structural diagram of an intelligent control system for vehicle cockpit equipment based on speech recognition. DETAILED DESCRIPTION

[0053] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0054] Reference Figure 1 As shown, a method for intelligent control of automobile cockpit equipment based on voice recognition includes:

[0055] Acquire multi-source audio signals, analyze the spectral overlap between speech signals and noise signals and the phase interference between speech signals, and evaluate the current speech recognition interference level;

[0056] Classify and initially prioritize multi-source voice commands. When the system is in a normal voice recognition state, execute multi-source voice commands in the order of command priority.

[0057] When the system is in a state of voice recognition interference, it recognizes, classifies and asynchronously processes multi-source voice commands, and dynamically adjusts the execution priority weight of commands in the asynchronous processing channel;

[0058] Based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel, the system resource allocation is dynamically planned to determine the specific execution priority order of the instructions in the asynchronous processing channel.

[0059] Reference Figure 2 As shown, multi-source audio signals are acquired, the spectrum overlap between speech signals and noise signals and the phase interference between speech signals are analyzed, and the current speech recognition interference level is evaluated.

[0060] Acquire real-time multi-source audio signals, set the voice signal frequency range and the noise signal frequency range, wherein the voice signal frequency range and the noise signal frequency range are allowed to overlap each other.

[0061] The audio signal in the frequency range of the speech signal is marked as a speech signal, and the audio signal in the frequency range of the noise signal is marked as a noise signal;

[0062] Perform spectrum analysis on speech signal and noise signal through fast Fourier transform to obtain spectrum density in time and frequency;

[0063] Sampling multiple time segments, processing the spectrum information in each time segment, and calculating the minimum spectrum density value of the time-series speech signal and the noise signal in each frequency band;

[0064] The dynamic spectrum overlap rate is evaluated by double integration of time and frequency. The dynamic spectrum overlap rate calculation expression is: ;

[0065] In the formula, and Respectively represent the spectral density of the target speech signal and the noise signal at time t and frequency f, T is the set time period, is the frequency component differential within the spectrum, Indicates the dynamic spectrum overlapping ratio.

[0066] The overlapping spectrum is calculated in each time segment, and the calculation result is averaged over the entire time period to obtain a comprehensive spectrum overlap measure. The comprehensive spectrum overlap measure value indicates that the overlap ratio of the speech signal and the noise signal in the spectrum is high, which may cause greater interference.

[0067] Through phase difference analysis, the phase relationship between multiple sound sources is identified and the impact of phase interference of multiple sound sources is quantified;

[0068] The phase information of multiple sound source signals is obtained by short-time Fourier transform, the phase value of each sound source on the time statistical scale is calculated, the cosine of the phase difference corresponding to each sound source is taken and integrated, and the average value of the integral calculation result is marked as the multi-source phase interference coefficient. The interference coefficient calculation expression is: ;

[0069] Where N is the total number of sound sources, , are the phases of sound sources i and j at time t, respectively. A larger interference coefficient value indicates greater phase interference between multiple sound sources, which may have a negative impact on the speech signal.

[0070] Based on the comprehensive spectrum overlap measurement and multi-source phase interference coefficient, the weighted phase superposition interference index is calculated comprehensively. The specific expression is: ;

[0071] In the formula, is the phase superposition interference index, is the comprehensive spectrum overlap measure, is the multi-source phase interference coefficient, for and The phase angle relationship between them is in the range of ;

[0072] A phase superposition interference index threshold is set. When the calculated phase superposition interference index is less than the phase superposition interference index threshold, it indicates that the system is in a normal state of speech recognition. When the calculated phase superposition interference index is greater than or equal to the phase superposition interference index threshold, the system state is changed to a speech recognition interference state. The phase superposition interference index threshold is set by professionals in this field taking into account the expected use site of the car and the actual noise resistance level, and will not be repeated here.

[0073] Reference Figure 3 As shown, the multi-source voice commands are categorized and initially prioritized. When the system is in a normal voice recognition state, the multi-source voice commands are executed in the order of command priority.

[0074] The system divides voice commands into two sets: driving-related commands and entertainment or comfort commands, and performs asynchronous processing separately. Driving-related commands (such as navigation, driving mode switching, etc.) have higher priority, while entertainment commands (such as audio control, air conditioning settings, etc.) are classified as secondary command sets. Each command set is processed separately and executed first when a high-priority command appears to ensure real-time response to key commands.

[0075] The voice commands executable by cockpit equipment are classified into command types, with voice commands related to driving assistance and driving safety classified as Class I commands, and voice commands related to infotainment and environmental control classified as Class II commands.

[0076] Setting the first priority weight for all instructions in the instruction set of the first category of instructions and the second category of instructions, wherein the first priority weight of the first category of instructions must be higher than the first priority weight of the second category of instructions;

[0077] When it is detected that the current system is in a normal voice recognition state, the multi-source voice commands received by the system are executed sequentially according to the time sequence of command reception and the first priority weight sequence assigned to the commands.

[0078] Reference Figure 4 As shown, when the system is in a speech recognition interference state, multi-source voice commands are recognized, classified and asynchronously processed, and the command execution priority weights in the asynchronous processing channel are dynamically adjusted.

[0079] When the system is in a speech recognition interference state, the received multi-source voice commands are split and processed;

[0080] The system identifies and distinguishes the types of commands to which voice commands belong, and uses a multi-threaded processing framework to place the first and second types of commands in different asynchronous processing channels, and analyzes each asynchronous processing channel separately through asynchronous processing strategies;

[0081] By calculating the current driving scenario score in real time and comparing it with the suitability of various voice commands, the system can automatically adjust the execution priority of voice commands.

[0082] The driving scenario is determined based on the current real-time driving parameters of the car, and the real-time driving score of the current driving scenario is calculated. The ideal driving score is set based on the ideal driving parameters of the driving scenario. The ideal driving parameters are set by professionals based on the model of the car and the risk conditions in the actual driving scenario, which will not be repeated here.

[0083] The driving score is a comprehensive score obtained after digital conversion of multiple relevant driving data such as body temperature, visibility, vehicle speed, vehicle driving stability, driving time, etc.

[0084] Analyze the execution frequency and effect of the current instruction in historical driving scenarios, and dynamically evaluate the driving scenario matching degree of the instruction based on the deviation between the driving score and the ideal driving score. The driving scenario matching degree of the instruction is calculated as follows: ;

[0085] In the formula, , are the actual driving score and ideal driving score of the current scene after the kth instruction is executed, is the effect evaluation value after the execution of the kth instruction, is the execution frequency weight of the kth instruction, is the number of instructions in the history, is the real-time driving scene matching degree of the kth instruction;

[0086] Different voice commands have specific requirements for driving scenarios. For example, commands such as navigation adjustment and brake assistance have higher priority in complex scenarios. Based on the applicability of the command under different driving conditions, ensure that specific commands are executed first in specific scenarios.

[0087] The voice feature matching model is used to perform feature recognition on the voice to obtain the identity information of the command initiator. The command initiation confidence of the current voice command is evaluated based on the identity information recognition result combined with the real-time signal-to-noise ratio of the current voice information.

[0088] Among them, the identity recognition results need to include multi-dimensional output features, including age, gender, distance from the driver's seat, etc., and these features need to be mathematically transformed to comprehensively evaluate the command confidence of the identity information of the command initiator. For example, when the identity recognition age is low and the distance from the cab is far, the confidence of the command will be lower. On the contrary, when the age is high and the distance from the cab is close, the confidence of the command will be increased.

[0089] The execution priority weight of the command in the asynchronous processing channel is dynamically set according to the real-time driving scene matching degree of the command and the confidence degree of the command initiation, and marked as the second priority weight of the command.

[0090] Reference Figure 5 As shown, the system resource allocation is dynamically planned based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel to determine the specific execution priority order of the instructions in the asynchronous processing channel.

[0091] Obtaining the first priority weight and the second priority weight of the current multi-source voice command, normalizing the scales of the first priority weight and the second priority weight, performing weighted summation, and using the calculation result as the final execution priority weight of the command;

[0092] Get the total amount of available resources for system voice command processing;

[0093] Obtain all voice commands in the asynchronous processing channel, and dynamically allocate available system resources in proportion to the final execution priority weight of each voice command relative to the total final execution priority weight of all voice commands;

[0094] Based on the system resources planned and allocated for each instruction, a priority execution queue is set in the asynchronous processing channel, and the instructions are added to the execution queue in sequence according to the final execution priority weight of the instructions and executed sequentially through system scheduling.

[0095] Reference Figure 6 As shown, an intelligent control system for automobile cockpit equipment based on voice recognition includes:

[0096] A speech recognition interference level assessment module, which comprehensively calculates a weighted phase superposition interference index based on the analysis of the spectrum overlap between the speech signal and the noise signal and the phase interference between the speech signals, and updates the system state to a normal speech recognition state or a speech recognition interference state in real time based on the calculation result of the phase superposition interference index;

[0097] A command priority calculation module, which divides the voice commands executable by the cockpit device into command types, determines the initial priority weights of the commands and calculates the asynchronous processing execution priority weights, comprehensively calculates the initial priority settings of the commands and the dynamic command execution priority weights in the asynchronous processing channel, and determines the specific execution priority order of the commands in the asynchronous processing channel;

[0098] A system resource dynamic allocation module, which dynamically allocates available system resources in proportion to all voice commands in the asynchronous processing channel based on the proportion of the final execution priority weight of each voice command relative to the total final execution priority weights of all voice commands;

[0099] An instruction scheduling execution module automatically selects an instruction scheduling execution strategy according to the current system state. When it is detected that the current system is in a normal speech recognition state, the system receives multi-source speech instructions and executes the instructions sequentially according to the time sequence of instruction reception and the first priority weight sequence assigned to the instructions. When the system is in a speech recognition interference state, the execution is scheduled in the execution queue sequence of the asynchronous processing channel.

[0100] Furthermore, the present solution also proposes a storage medium for a method for intelligent control of vehicle cockpit equipment based on voice recognition, on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned method for intelligent control of vehicle cockpit equipment based on voice recognition is executed.

[0101] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0102] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for intelligent control of automobile cockpit equipment based on voice recognition, characterized in that: include: Acquire multi-source audio signals, analyze the spectral overlap between speech signals and noise signals and the phase interference between speech signals, and evaluate the current speech recognition interference level; Classify and initially prioritize multi-source voice commands. When the system is in a normal voice recognition state, execute multi-source voice commands in the order of command priority. When the system is in a state of voice recognition interference, it recognizes, classifies and asynchronously processes multi-source voice commands, and dynamically adjusts the execution priority weight of commands in the asynchronous processing channel; Dynamically plan system resource allocation based on initial priority setting of instructions and dynamic instruction execution priority weights in the asynchronous processing channel, and determine the specific execution priority order of instructions in the asynchronous processing channel; When the system is in a speech recognition interference state, the recognition classification and asynchronous processing of multi-source speech commands and the dynamic adjustment of the execution priority weight of the commands in the asynchronous processing channel specifically include: When the system is in a speech recognition interference state, the received multi-source voice commands are split and processed; The system identifies and distinguishes the types of commands to which voice commands belong, and uses a multi-threaded processing framework to place the first and second types of commands in different asynchronous processing channels, and analyzes each asynchronous processing channel separately through asynchronous processing strategies; The driving scene is determined based on the current real-time driving parameters of the car, and the real-time driving score of the current driving scene is calculated. The ideal driving score is set based on the ideal driving parameters of the driving scene. Analyze the execution frequency and effect of the current instruction in historical driving scenarios, and dynamically evaluate the driving scenario matching degree of the instruction based on the deviation between the driving score and the ideal driving score. The driving scenario matching degree of the instruction is calculated as follows: ； In the formula, , are the actual driving score and ideal driving score of the current scene after the kth instruction is executed, is the effect evaluation value after the execution of the kth instruction, is the execution frequency weight of the kth instruction, is the number of instructions in the history, is the real-time driving scene matching degree of the kth instruction; The voice feature matching model is used to identify the identity information of the command initiator of the voice, and the command initiation confidence of the current voice command is evaluated based on the identity information recognition result and the real-time signal-to-noise ratio of the current voice information; The execution priority weight of the command in the asynchronous processing channel is dynamically set according to the real-time driving scene matching degree of the command and the confidence degree of the command initiation, and marked as the second priority weight of the command.

2. According to claim 1, a method for intelligent control of automobile cockpit equipment based on voice recognition is characterized in that: The acquiring of multi-source audio signals, analyzing the spectrum overlap between the speech signal and the noise signal and the phase interference between the speech signals, and evaluating the current speech recognition interference level specifically includes: Obtain real-time multi-source audio signals and set the frequency range of voice signals and noise signals; The audio signal in the frequency range of the speech signal is marked as a speech signal, and the audio signal in the frequency range of the noise signal is marked as a noise signal; Perform spectrum analysis on speech signal and noise signal through fast Fourier transform to obtain spectrum density in time and frequency; Sampling multiple time periods, processing the spectrum information in each time period, and calculating the minimum spectrum density value of the time series speech signal and noise signal in each frequency band; The dynamic spectrum overlap rate is evaluated by double integration of time and frequency, the overlapping spectrum is calculated in each time segment, and the calculation result is averaged over the entire time period to obtain the comprehensive spectrum overlap metric; Through phase difference analysis, the phase relationship between multiple sound sources is identified and the impact of phase interference of multiple sound sources is quantified; The phase information of multiple sound source signals is obtained by short-time Fourier transform, the phase value of each sound source on the time statistical scale is calculated, the cosine of the phase difference corresponding to each sound source is taken and integrated, and the average value of the integral calculation result is marked as the multi-source phase interference coefficient; Based on the comprehensive spectrum overlap measurement and multi-source phase interference coefficient, the weighted phase superposition interference index is comprehensively calculated. The specific expression is: , In the formula, is the phase superposition interference index, is the comprehensive spectrum overlap measure, is the multi-source phase interference coefficient, for and The phase angle relationship between them is in the range of ; Set the phase superposition interference index threshold. When the calculated phase superposition interference index is less than the phase superposition interference index threshold, it indicates that the system is in a normal speech recognition state. When the calculated phase superposition interference index is greater than or equal to the phase superposition interference index threshold, the system state is changed to a speech recognition interference state.

3. The method for intelligent control of vehicle cockpit equipment based on voice recognition according to claim 2 is characterized in that: The categorization and initial priority setting of the multi-source voice commands, when the system is in a normal voice recognition state, executing the multi-source voice commands in the order of command priority specifically includes: The voice commands executable by cockpit equipment are classified into command types, with voice commands related to driving assistance and driving safety classified as Class I commands, and voice commands related to infotainment and environmental control classified as Class II commands; Setting the first priority weight for all instructions in the instruction set of the first category of instructions and the second category of instructions, wherein the first priority weight of the first category of instructions must be higher than the first priority weight of the second category of instructions; When it is detected that the current system is in a normal voice recognition state, the multi-source voice commands received by the system are executed sequentially according to the time sequence of command reception and the first priority weight sequence assigned to the commands.

4. The method for intelligent control of vehicle cockpit equipment based on voice recognition according to claim 3 is characterized in that: The dynamically planning system resource allocation based on the initial priority setting of the instructions and the dynamic instruction execution priority weights in the asynchronous processing channel to determine the specific execution priority order of the instructions in the asynchronous processing channel specifically includes: Obtaining the first priority weight and the second priority weight of the current multi-source voice command, normalizing the scales of the first priority weight and the second priority weight, performing weighted summation, and using the calculation result as the final execution priority weight of the command; Get the total amount of available resources for system voice command processing; Obtain all voice commands in the asynchronous processing channel, and dynamically allocate available system resources in proportion to the final execution priority weight of each voice command relative to the total final execution priority weight of all voice commands; Based on the system resources planned and allocated for each instruction, a priority execution queue is set in the asynchronous processing channel, and the instructions are added to the execution queue in sequence according to the final execution priority weight of the instructions and executed sequentially through system scheduling.

5. An intelligent control system for automobile cockpit equipment based on voice recognition, used to implement an intelligent control method for automobile cockpit equipment based on voice recognition as claimed in any one of claims 1 to 4, characterized in that: include: A speech recognition interference level assessment module, which comprehensively calculates a weighted phase superposition interference index based on the analysis of the spectrum overlap between the speech signal and the noise signal and the phase interference between the speech signals, and updates the system state to a normal speech recognition state or a speech recognition interference state in real time based on the calculation result of the phase superposition interference index; A command priority calculation module, which divides the voice commands executable by the cockpit device into command types, determines the initial priority weights of the commands and calculates the asynchronous processing execution priority weights, comprehensively calculates the initial priority settings of the commands and the dynamic command execution priority weights in the asynchronous processing channel, and determines the specific execution priority order of the commands in the asynchronous processing channel; A system resource dynamic allocation module, which dynamically allocates available system resources in proportion to all voice commands in the asynchronous processing channel based on the proportion of the final execution priority weight of each voice command relative to the total final execution priority weights of all voice commands; An instruction scheduling execution module automatically selects an instruction scheduling execution strategy according to the current system state. When it is detected that the current system is in a normal speech recognition state, the system receives multi-source speech instructions and executes the instructions sequentially according to the time sequence of instruction reception and the first priority weight sequence assigned to the instructions. When the system is in a speech recognition interference state, the execution is scheduled in the execution queue sequence of the asynchronous processing channel.

Images