Linear array differential beam forming method and system based on non-uniform directivity microphone orientation optimization, storage medium and electronic equipment

Abstract

The invention discloses a linear array differential beam forming method and system based on non-uniform directional microphone orientation optimization, a storage medium and electronic equipment. The method comprises the following steps: constructing a linear microphone array of a directional microphone combination of an omnidirectional directional microphone and a directional microphone combination with a non-uniform microphone orientation angle; calculating a steering vector of the microphone array based on the directivity and the position of each microphone in the microphone array; a linear equation set is obtained based on the steering vector, the undistorted constraint condition and the zero point constraint condition; optimizing a microphone orientation angle of the pointing microphone by adopting a grid search strategy; and constructing an optimization problem based on a linear equation set to obtain a weight vector in an actual wave beam. Compared with a traditional design, higher white noise gain is achieved, deflection of beams in any direction is supported, and excellent performance is kept at all deflection angles; compared with a zero point constraint method, the method has the advantage that the aspects of WNG-error equalization, beam approximation, frequency consistency and the like are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of beamforming technology, and more specifically to a linear array differential beamforming method, system, storage medium, and electronic device based on non-uniform directional microphone orientation optimization. Background Technology

[0002] Microphone array signal processing is a field that focuses on speech signals, building upon traditional array signal processing. Initially, research on microphone arrays primarily concentrated on additive microphone arrays (AMAs). It wasn't until the 1930s, with the invention of directional microphones, that the concept of differential microphone arrays was proposed, and research on differential microphone arrays gradually began thereafter. The first literature on differential microphone arrays appeared in the 1940s.

[0003] Differential beamforming is one of many beamforming algorithms. When designing beamformers using differential beamforming algorithms, the spacing between adjacent microphones in the microphone array is typically small, allowing the difference between the outputs of each microphone to approximate the differential of the sound pressure field. Generally, a first-order differential beamformer is formed by combining the differential outputs of two adjacent omnidirectional microphones, a second-order differential beamformer by combining the differential outputs of two adjacent first-order differential beamformers, and an N+1-order differential beamformer by combining the differential outputs of two adjacent N (N≥2)-order differential beamformers. Generally, a higher order of differential beamformer indicates a higher directivity factor (DF) and lower white noise gain (WNG). For an N (N≤M-1)-order beamformer composed of M omnidirectional microphones, its beam pattern typically forms N nulls between 0 and π. The beamformer forms the main beam in the desired direction and creates nulls in the direction of interference and noise, thereby enhancing the desired signal and suppressing noise and interference.

[0004] Existing methods include linear superarrays based on Jacobi-Anger series expansions. These methods utilize Jacobi-Anger series expansions to design omnidirectional adjustable differential beamformers and approximate beamforming through series truncation. However, this approach relies on analytical expressions for ideal beammaps, limiting beammap design flexibility. Furthermore, the series truncation error increases with frequency, leading to main lobe shift and sidelobe increase in high-frequency bands. The lack of clear criteria for selecting the truncation order makes it difficult to balance beammap error and white noise gain. Alternative methods employ linear arrays using similar directional microphones. This approach optimizes microphone layout and beam weights, achieving differential beamforming in the end-fire direction with high directivity. However, this approach is limited by a single type of directional microphone, resulting in low freedom in beamform design. Beam performance significantly degrades in non-end-fire directions, white noise gain is low, and there is a lack of support for mixed configurations of multi-directional microphones, limiting spatial processing capabilities.

[0005] In summary, it is necessary to design a novel beamforming scheme to address the problems of traditional schemes, such as dependence on the analytical expression of the ideal beam pattern, deterioration of beam performance in the high-frequency band, low degree of freedom in beam shape design, low white noise gain performance, and lack of support for mixed configurations of multidirectional microphones. Summary of the Invention

[0006] The purpose of this invention is to provide a linear array differential beamforming method, system, storage medium, and electronic device based on non-uniform directional microphone orientation optimization.

[0007] To achieve the above objectives, in one aspect, this invention proposes a linear array differential beamforming method based on non-uniform directional microphone orientation optimization, comprising:

[0008] S1, Construct a linear microphone array, the microphone array comprising a plurality of omnidirectional microphones and directional microphones arranged at intervals, each of the directional microphones having a variable microphone orientation angle;

[0009] S2, Calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array;

[0010] S3. Based on the aforementioned guiding vector, distortion-free constraint conditions, and zero-point constraint conditions, a system of linear equations is obtained.

[0011] S4. A grid search strategy is used to optimize the microphone orientation angle pointing to the microphone in the microphone array. The orientation angle corresponding to the white noise gain with the largest average value in the broadband range is searched as the optimal microphone orientation angle.

[0012] S5. Based on the linear equations and white noise gain constraints, an optimization problem is constructed to minimize the error between the actual beam and the ideal beam. The weight vector in the actual beam is obtained from the optimization problem.

[0013] In a preferred embodiment, in step S1, the directional pattern of each microphone in a three-dimensional spherical coordinate system is as follows: ,in, The directivity coefficient for each microphone, It is the angle of incidence of sound. For the m-th microphone, This refers to the microphone's facing angle.

[0014] In a preferred embodiment, in step S2, the steering vector of the microphone array is represented as:

[0015] ;

[0016] in, ,

[0017] , M is the number of microphones in the microphone array, and k is the wave number. Where f is the frequency and c is the speed of sound. This indicates the position coordinates of each microphone. This represents the transpose operation of a vector or matrix.

[0018] In a preferred embodiment, in step S3, the system of linear equations is as follows:

[0019]

[0020] When the minimum requirement M=2N+1 is met and When it is reversible,

[0021] When the minimum requirement M>2N+1 is met, the maximum beamformer for WNG is described as follows: : The filter that yields the maximum white noise gain The maximum WNG obtained under the constraints of the desired direction and zero point is:

[0022] in, , The main beam direction of the beam. Indicates transpose conjugate. For an Nth-order ideal beam in There are N distinct zeros within the range. This is the weight vector in the actual beam. .

[0023] In a preferred embodiment, in step S5, the white noise gain constraint condition is expressed as:

[0024] ;

[0025] in, It is a number greater than or equal to 0, measured in decibels. This is the white noise gain value that is less than the maximum white noise gain.

[0026] In a preferred embodiment, in step S5, the error between the actual beam and the ideal beam is expressed as:

[0027] ;

[0028] in, This is the actual beam; For an ideal beam,

[0029] In a preferred embodiment, the optimization problem is expressed as:

[0030] ;

[0031] in, ;

[0032] ;

[0033] ,yes A square formation.

[0034] On the other hand, the present invention proposes a linear array differential beamforming system based on non-uniform directional microphone orientation optimization, comprising:

[0035] An array building module is used to build a linear microphone array, the microphone array including a plurality of omnidirectional microphones and directional microphones arranged at intervals, the directional microphones having a variable microphone orientation angle;

[0036] A steering vector module is used to calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array;

[0037] The constraint equations acquisition module is used to obtain a linear equation system based on the guide vector, the distortion-free constraint conditions, and the zero-point constraint conditions.

[0038] The orientation angle optimization module is used to optimize the orientation angle of the microphones pointing to the microphones in the microphone array using a grid search strategy, and finds the orientation angle corresponding to the white noise gain with the largest average value in the broadband range as the optimal microphone orientation angle;

[0039] The weight acquisition module is used to construct an optimization problem that minimizes the error between the actual beam and the ideal beam based on the linear equation system and the white noise gain constraint, and obtain the weight vector in the actual beam from the optimization problem.

[0040] In another aspect, the present invention proposes a readable storage medium storing a computer program, which, when run, executes the steps in the beamforming method described above.

[0041] In another aspect, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and the computer program, when run by the processor, performs the steps in the above-described beamforming method.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] This invention employs a linear array of omnidirectional and directional microphones arranged alternately, and optimizes the orientation of the directional microphones through grid search. The optimal microphone orientation angle is found by identifying the orientation angle corresponding to the white noise gain with the largest average value over a wide bandwidth, achieving a higher white noise gain than traditional designs. Furthermore, the method of this invention supports arbitrary beam deflection and maintains superior performance at all deflection angles. Compared with the zero-point constraint method, this invention shows significant improvements in WNG-error equalization, beam approximation, and frequency consistency. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the linear array differential beamforming method based on non-uniform directional microphone orientation optimization according to the present invention.

[0045] Figure 2 This is a schematic diagram of the microphone array in Embodiment 1 of the present invention;

[0046] Figure 3 The orientation angle searched in Embodiment 1 of the present invention and Heatmap of the maximum WNG with the corresponding average value within the broadband;

[0047] Figures 4a-4d The beam diagrams show the beams formed by the existing zero-point constraint method, the beams formed by the method of the present invention, and the ideal beams at different frequencies (2000Hz, 4000Hz, 6000Hz, and 8000Hz) in Embodiment 1 of the present invention.

[0048] Figure 5a and Figure 5b These are, respectively, the beam diagrams of beamformers designed using existing zero-point constraint methods and the beam diagrams of beamformers designed in Embodiment 1 of the present invention;

[0049] Figures 6a-6c These are, respectively, a comparative diagram of the WNG performance of the beamformer designed in Embodiment 1 of the present invention and the beamformer designed by the existing zero-point constraint method, a comparative diagram of the DF performance of the beamformer designed in Embodiment 1 of the present invention and the beamformer designed by the existing zero-point constraint method, and a comparative diagram of the error (ERROR) of the beamformer designed in Embodiment 1 of the present invention and the beamformer designed by the existing zero-point constraint method.

[0050] Figure 7 This is a schematic diagram of the microphone array in Embodiment 2 of the present invention;

[0051] Figure 8 The orientation angle searched in Embodiment 2 of the present invention and Heatmap of the maximum WNG with the corresponding average value within the broadband;

[0052] Figures 9a-9d The beam diagrams show the beams formed by the zero-point constraint method and the beams formed by the method of the present invention at different frequencies (2000Hz, 4000Hz, 6000Hz, and 8000Hz) in Embodiment 2 of the present invention, and are compared with the ideal beams.

[0053] Figure 10a and Figure 10b These are, respectively, the beam diagrams of beamformers designed using existing zero-point constraint methods and the beam diagrams of beamformers designed in Embodiment 2 of the present invention;

[0054] Figures 11a-11c These are, respectively, a comparative diagram of the WNG performance of the beamformer designed in Embodiment 2 of the present invention and the beamformer designed by the existing zero-point constraint method, a comparative diagram of the DF performance of the beamformer designed in Embodiment 2 of the present invention and the beamformer designed by the existing zero-point constraint method, and a comparative diagram of the error (ERROR) of the beamformer designed in Embodiment 2 of the present invention and the beamformer designed by the existing zero-point constraint method. Detailed Implementation

[0055] The specific embodiments of the present invention will be described in detail below, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0056] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprising" shall be understood to include the stated elements or components without excluding other elements or other components.

[0057] It should be noted that beamforming is achieved by using a linear spatial filter. It is applied to the observation of signals by a microphone, and the ultimate goal of beamforming is to determine the optimal beamformer. This is done so that the signal from the microphone array matches the desired signal.

[0058] like Figure 1 As shown, the present invention discloses a linear array differential beamforming method based on non-uniform directional microphone orientation optimization, belonging to the field of microphone array beamforming technology, and applicable to scenarios requiring directional sound pickup such as voice communication, video conferencing, and smart terminals. The method specifically includes the following steps:

[0059] S1, Construct a linear microphone array comprising a plurality of omnidirectional microphones and directional microphones arranged at intervals, wherein the directional microphones have a variable microphone orientation angle.

[0060] Specifically, the microphone array is a linear differential microphone array formed by combining various types of microphones. A linear differential microphone array is defined as having M microphone units, specifically including multiple omnidirectional and directional microphones arranged at intervals, with one omnidirectional microphone spaced apart from one directional microphone. The directional microphones have variable directional angles. The microphone types can be either omnidirectional or directional; the directional microphones can be cardioid, supercardioid, or dipole microphones, etc. Figure 2 and Figure 7 As shown, in both Examples 1 and 2, the directional microphone is a dipole microphone.

[0061] The directional pattern of each microphone in a microphone array can be represented as:

[0062] ;

[0063] in, For each microphone, the directivity coefficient, such as when When = 0.5, it indicates that the microphone is in cardioid beam pattern. When =0, it indicates that the microphone is in dipole beam mode. It is the angle of incidence of sound. Let M be the m-th microphone, and M be the number of microphones in the microphone array. The microphone orientation angle is denoted by directional microphones. Unlike existing microphone arrays where all microphones are of the same type and the directional microphones only have an end-fire direction, the microphone array of this invention is formed by alternating omnidirectional and directional microphones, and the directional microphones have variable microphone orientation angles. By optimizing the orientation angles of each directional microphone in the hybrid array, higher white noise gain is achieved compared to traditional designs.

[0064] S2, Calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array.

[0065] Specifically, the steering vector of the microphone array is obtained based on the directivity and position of each microphone in the array. In this embodiment, the steering vector of the microphone array... Specifically, it is expressed as follows:

[0066] ;

[0067] in, Let be the microphone response matrix composed of the directivity of the aforementioned M microphones. , is the steering vector of the microphone array. In the formula, k is the wave number. Where f is the frequency and c is the speed of sound. This indicates the position coordinates of each microphone. , , The spacing between microphones This represents the transpose operation of a vector or matrix. It is the imaginary unit.

[0068] S3. Based on the guiding vector, the distortion-free constraint, and the zero-point constraint, a system of linear equations is obtained.

[0069] Specifically, in this embodiment, the problem of designing an Nth-order differential beamforming system for a microphone array is described as a system of linear equations. This system of linear equations is specifically derived based on the aforementioned steering vector and distortion-free constraint conditions (i.e., ensuring distortion-free passage of the signal in the desired direction) and null constraint conditions (i.e., forming nulls in the interference direction). That is, through a combined design method of null constraint and distortion-free constraint, multiple null positions of the beam are subjected to distortion-free constraint. In this embodiment, the system of linear equations is as follows:

[0070]

[0071] When the minimum requirement M=2N+1 is met and When it is reversible,

[0072] When the minimum requirement M>2N+1 is met, the maximum beamformer for WNG is described as follows: : The filter that yields the maximum white noise gain The maximum WNG obtained under the constraints of the desired direction and zero point is:

[0073] in, , The main beam direction of the beam. Indicates transpose conjugate. For an Nth-order ideal beam in There are N distinct zeros within the range. This is the weight vector in the actual beam. .

[0074] S4. A grid search strategy is used to optimize the microphone orientation angles pointing to the microphones in the microphone array. The orientation angle corresponding to the white noise gain with the largest average value in the broadband range is searched as the optimal microphone orientation angle.

[0075] Specifically, in this embodiment, the search range of the grid search strategy is: frequency from 100Hz to 8kHz, and orientation angle. From 0 to Search step size: frequency step size is 50Hz, orientation angle The step size is 1°; its optimization objective is to find the orientation angle corresponding to the white noise gain with the largest average value within the broadband range. The optimal microphone orientation angle is determined by a specific search process, such as traversing the orientation angle using a two-dimensional grid. (0-180°) and orientation angle All combinations of (0-180°) are determined by symmetry. and For each angle combination, its white noise gain (WNG) at multiple frequency points is calculated, and then the average value across the frequency band is taken. Finally, the angle combination with the largest average WNG is selected as the optimal solution. This is a broadband optimization strategy that ensures the array has good robustness and efficiency across the entire frequency band, rather than being optimal only at a specific frequency.

[0076] The search table is shown below:

[0077]

[0078] S5. Based on the linear equations and white noise gain constraints, an optimization problem is constructed to minimize the error between the actual beam and the ideal beam. The weight vector in the actual beam is obtained from the optimization problem.

[0079] Preferably, the present invention adds a white noise gain constraint to the optimization problem. In this embodiment, the white noise gain constraint is specifically designed as follows:

[0080] .

[0081] in, It is a number greater than or equal to 0, measured in decibels. This is the white noise gain value that is less than the maximum white noise gain.

[0082] Furthermore, this invention uses the error between the actual beam and the ideal beam as the objective function of the optimization problem. In this embodiment, the error between the actual beam and the ideal beam is specifically expressed as:

[0083] ;

[0084] In the formula, This is the actual beam; , for the sound pressure field along An ideal beam of order N in the direction can be written in vector mode as follows: In the formula, .

[0085] in, ; .

[0086] error In , It can then be calculated using the following formula:

[0087] ;

[0088] ;

[0089] ,yes A square formation.

[0090] Therefore, the present invention designs the optimization problem as an optimization problem with minimizing the error between the actual beam and the ideal beam as the objective function, and combining the above-mentioned zero-point constraints, distortion-free constraints and white noise gain constraints to form an optimization problem with constraints.

[0091] In this embodiment, the optimization problem is specifically expressed as:

[0092] .

[0093] By solving the above optimization problem, the weight vector in the actual beam can be calculated. This results in the final beamformer structure.

[0094] The performance of the linear array differential beamforming method designed in this invention is illustrated below with a specific embodiment. In this embodiment, as... Figure 2 As shown, the linear microphone array is depicted as a 7-unit microphone array, specifically consisting of 3 omnidirectional microphones and 4 bidirectional microphones, with corresponding response matrices. These microphone units are arranged linearly with a spacing of 0.005m, operating in the frequency range of 100Hz to 8kHz, with a search step size of 50Hz. The desired beam pattern in this example is a second-order (N=2) sharp cardioid beam pattern (this beam pattern has two zeros, namely...). ), The coefficient is Optimization parameters: White noise gain The maximum gain of white noise Maximum white noise gain (Right now White noise gain WNG, three-dimensional directivity factor DF, and error at different values. (Selection) The optimal orientation angle for the bidirectional microphone end-fire direction was determined by beam comparison and grid search. For ease of description, this embodiment is defined as Embodiment 1, combined with Figure 2 and Figure 3 As shown; the optimal orientation angle in the 30° direction is , combined Figure 7 and Figure 8 As shown, this embodiment is defined as Embodiment 2.

[0095] Figures 4a-4d The diagrams illustrate the beamformations of the zero-point constraint method and the method of the present invention at different frequencies (2000Hz, 4000Hz, 6000Hz, and 8000Hz) compared to the ideal beam. As can be seen from the diagrams, the beamformer designed in Example 1 produces an actual beam that completely or nearly completely overlaps with the ideal beam at different frequencies (i.e., the error between the two is minimized), and is closer to the ideal beam than the beam formed by the existing zero-point constraint method.

[0096] like Figure 5a and Figure 5b The figures shown are beam diagrams of beamformers designed using existing zero-point constraint methods and beam diagrams of beamformers designed according to the present invention. Figure 5aIt can be seen that although the existing zero-point constraint method can maintain a constant beam pattern below 6000 Hz, the actual beam will deviate from the ideal beam above 6000 Hz. However, the actual beam pattern designed by the method of the present invention can remain basically constant throughout the entire evaluation frequency band (i.e., 100Hz~8kHz), indicating that the beam designed by the present invention has the characteristic of frequency invariance.

[0097] In addition, metrics for evaluating beamformer performance typically include white noise gain (WNG), beammap, and directivity factor (DF). WNG demonstrates the beamformer's ability to suppress spatially uncorrelated noise and is also the most convenient method for assessing the beamformer's sensitivity to certain defects (such as sensor noise, position errors, etc.). Specifically, it is expressed as follows: The beam diagram shows the directional sensitivity of the beamformer to plane waves incident on the array from an incident angle θ, specifically expressed as follows: The directivity factor (DF) is defined as the ratio between the array output response power in the desired steering direction and the average power in the 0° to 360° direction, and is specifically expressed as: .

[0098] like Figures 6a-6c The figures shown are, respectively, a WNG performance comparison diagram between the beamformer designed in Embodiment 1 of the present invention and a beamformer designed using the existing zero-point constraint method; a DF performance comparison diagram between the beamformer designed in Embodiment 1 of the present invention and a beamformer designed using the existing zero-point constraint method; and an error comparison diagram between the beamformer designed in Embodiment 1 of the present invention and a beamformer designed using the existing zero-point constraint method. Figure 6b It can be seen that the DF value of the method of the present invention remains approximately constant throughout the entire evaluation frequency band (i.e., 100Hz~8kHz), unlike the DF of the zero-point constraint method which exhibits fluctuations. The results indicate that the beam pattern formed by the design of the present invention is more stable. Figure 6a and 6c It can be seen that although the WNG of the method of this invention is slightly lower than that of the zero-point constraint method (about 10dB lower) above 300 Hz, the error between the beamformer and the ideal beam (ERROR) is more than 13dB lower than that of the zero-point constraint method. The comparison results show that, unlike the existing zero-point constraint method which maximizes white noise gain, the constraint condition designed in this invention can make fuller use of the array's degrees of freedom, so that the system can achieve a better balance between robustness (WNG) and beam pattern approximation (ERROR). Furthermore, after confirming the deflection angle through grid search, the average maximum WNG is obtained, and there are more degrees of freedom to achieve a balance between the system and the error.

[0099] Figures 9a-9dThe diagrams illustrate the beamformations of the zero-point constraint method and the method of the present invention at different frequencies (2000Hz, 4000Hz, 6000Hz, and 8000Hz) compared to the ideal beam. As can be seen from the diagrams, the beamformer designed in Example 2 produces an actual beam that completely or nearly completely overlaps with the ideal beam at different frequencies (i.e., the error between the two is minimized), and is closer to the ideal beam than the beam formed by the existing zero-point constraint method.

[0100] like Figure 10a and Figure 10b The figures shown are beam diagrams of beamformers designed using existing zero-point constraint methods and beam diagrams of beamformers designed according to Embodiment 2 of the present invention. Figure 10a and Figure 10b It can be seen that, in the 4kHz and higher frequency bands, the beam pattern designed in Embodiment 2 of the present invention has a significantly better fit with the ideal beam than the traditional zero-point constraint method, which has a significant increase in sidelobe gain and deterioration in main lobe-sidelobe suppression performance at high frequencies.

[0101] In addition, such as Figures 11a-11c The figures shown are, respectively, a WNG performance comparison diagram between the beamformer designed in Embodiment 2 of the present invention and a beamformer designed using the existing zero-point constraint method; a DF performance comparison diagram between the beamformer designed in Embodiment 2 of the present invention and a beamformer designed using the existing zero-point constraint method; and an error comparison diagram between the beamformer designed in Embodiment 2 of the present invention and a beamformer designed using the existing zero-point constraint method. Figure 11b It can be seen that the DF value of the method of the present invention remains approximately constant throughout the entire evaluation frequency band (i.e., 100Hz~8kHz), unlike the DF of the zero-point constraint method which exhibits fluctuations. The results indicate that the beam pattern formed by the design of the present invention is more stable. Figure 10a and 10b It can be seen that although the WNG of the method of this invention is slightly lower than that of the zero-point constraint method (about 10dB lower) above 300Hz, the error (ERROR) between the beamformer and the ideal beam is more than 12dB higher than that of the zero-point constraint method at high frequencies. The comparison results show that, unlike the existing zero-point constraint method which maximizes white noise gain, the constraint condition designed in this invention can make fuller use of the array's degrees of freedom, so that the system can achieve a better balance between robustness (WNG) and beam pattern approximation (ERROR). Moreover, even at different deflection angles, the grid search is still effective, and while obtaining the maximum WNG, WNG and error can also be better balanced.

[0102] Furthermore, by providing Embodiments 1 and 2 above, it can be verified that the present invention achieves a comprehensive improvement in the all-angle performance of the beamformer through non-uniform directional microphone orientation optimization. Compared with the traditional zero-point constraint method, it not only has a significant advantage in the end-fire direction (WNG improvement of 10-15 dB), but more importantly, it maintains stable performance superiority at any deflection angle from 0° to 180°, including higher white noise gain, a beam pattern closer to the ideal, and better frequency consistency. This characteristic makes the present method more adaptable and reliable in practical applications.

[0103] The advantages of this invention are as follows: it constructs a linear array by alternating omnidirectional and directional microphones, and optimizes the orientation of the directional microphones through grid search. The orientation angle corresponding to the white noise gain with the largest average value in the broadband range is found as the optimal microphone orientation angle, achieving a higher white noise gain than traditional designs. Furthermore, the method of this invention supports arbitrary beam deflection and maintains superior performance at all deflection angles. Compared with the zero-point constraint method, this invention has significant improvements in WNG-error equalization, beam approximation, and frequency consistency.

[0104] Corresponding to the beamforming method described above, this invention also discloses a linear array differential beamforming system based on non-uniform directional microphone orientation optimization, specifically including:

[0105] An array building module is used to build a linear microphone array, the microphone array including a plurality of omnidirectional microphones and directional microphones arranged at intervals, the directional microphones having a variable microphone orientation angle;

[0106] A steering vector module is used to calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array;

[0107] The constraint equations acquisition module is used to obtain a linear equation system based on the guide vector, the distortion-free constraint conditions, and the zero-point constraint conditions.

[0108] The orientation angle optimization module is used to optimize the orientation angle of the microphones pointing to the microphones in the microphone array using a grid search strategy, and finds the orientation angle corresponding to the white noise gain with the largest average value in the broadband range as the optimal microphone orientation angle;

[0109] The weight acquisition module is used to construct an optimization problem that minimizes the error between the actual beam and the ideal beam based on the linear equation system and the white noise gain constraint, and obtain the weight vector in the actual beam from the optimization problem.

[0110] The working principles of these five modules can be referred to in the descriptions of steps S1 to S5 above, and will not be repeated here.

[0111] On the other hand, the present invention also provides a readable storage medium having a computer program stored thereon, which, when run, implements the steps in the beamforming method provided in the above embodiments.

[0112] In another aspect, the present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, performs the steps in the above-described beamforming method.

[0113] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0114] More specific examples (a non-exhaustive list) of readable storage media include: electrical connections (electronic devices) with one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0115] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0116] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A linear array differential beamforming method based on non-uniform directional microphone orientation optimization, characterized in that, The method includes: S1, construct a linear microphone array, the microphone array comprising a plurality of omnidirectional microphones and directional microphones arranged at intervals, each of the directional microphones having a variable microphone orientation angle; S2, Calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array; S3. Based on the aforementioned guiding vector, distortion-free constraint conditions, and zero-point constraint conditions, a system of linear equations is obtained. S4. A grid search strategy is used to optimize the microphone orientation angle pointing to the microphone in the microphone array. The orientation angle corresponding to the white noise gain with the largest average value in the broadband range is searched as the optimal microphone orientation angle. S5. Based on the linear equations and white noise gain constraints, an optimization problem is constructed to minimize the error between the actual beam and the ideal beam. The weight vector in the actual beam is obtained from the optimization problem.

2. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 1, characterized in that, In S1, in the three-dimensional spherical coordinate system, the directional pattern of each microphone is as follows: ,in, The directivity coefficient for each microphone, It is the angle of incidence of sound. For the m-th microphone, This refers to the microphone's facing angle.

3. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 2, characterized in that, In step S2, the steering vector of the microphone array is represented as follows: ； in, , , M is the number of microphones in the microphone array, and k is the wave number. Where f is the frequency and c is the speed of sound. This indicates the position coordinates of each microphone. This represents the transpose operation of a vector or matrix.

4. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 3, characterized in that, In S3, the system of linear equations is as follows: ； When the minimum requirement M=2N+1 is met and When it is reversible, ; When the minimum requirement M>2N+1 is met, the maximum beamformer for WNG is described as follows: : The filter that yields the maximum white noise gain ; The maximum WNG obtained under the constraints of the desired direction and zero point is: ; in, , The main beam direction of the beam. Indicates transpose conjugate. For an Nth-order ideal beam in There are N distinct zeros within the range. This is the weight vector in the actual beam. .

5. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 4, characterized in that, In S5, the white noise gain constraint condition is expressed as follows: ； in, It is a number greater than or equal to 0, measured in decibels. This is the white noise gain value that is less than the maximum white noise gain.

6. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 5, characterized in that, In S5, the error between the actual beam and the ideal beam is expressed as: ； in, This is the actual beam; For an ideal beam, 7. The linear array differential beamforming method based on non-uniform directional microphone orientation optimization as described in claim 6, characterized in that, The optimization problem is expressed as: ； in, ; ； ,yes A square formation.

8. A linear array differential beamforming system based on non-uniform directional microphone orientation optimization, characterized in that, The system includes: An array building module is used to build a linear microphone array, the microphone array including a plurality of omnidirectional microphones and directional microphones arranged at intervals, the directional microphones having a variable microphone orientation angle; A steering vector module is used to calculate the steering vector of the microphone array based on the directivity and position of each microphone in the microphone array; The constraint equations acquisition module is used to obtain a linear equation system based on the guide vector, the distortion-free constraint conditions, and the zero-point constraint conditions. The orientation angle optimization module is used to optimize the orientation angle of the microphones pointing to the microphones in the microphone array using a grid search strategy, and finds the orientation angle corresponding to the white noise gain with the largest average value in the broadband range as the optimal microphone orientation angle; The weight acquisition module is used to construct an optimization problem that minimizes the error between the actual beam and the ideal beam based on the linear equation system and the white noise gain constraint, and obtain the weight vector in the actual beam from the optimization problem.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed, performs the steps of the beamforming method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, performs the steps of the beamforming method according to any one of claims 1 to 7.

Images