Reconfigurable microphone array system and control method thereof

Abstract

The invention relates to the technical field of voice recognition and sound source localization, and discloses a reconfigurable microphone array system and a control method thereof.The reconfigurable microphone array system comprises a flexible substrate used for bearing a microphone array, a microphone array module comprises at least two fixed microphones and four movable microphones, and the movable microphones are connected with the flexible substrate through magnetic attraction sliding blocks; the motor driving module is used for controlling the movable microphone to move along a preset track; the multi-mode audio processor integrates a sound source localization algorithm and a 3A audio processing algorithm and is used for analyzing a user instruction, generating a motor control signal and processing audio data; the closed-loop feedback module comprises a plurality of travel switches, detects the position of the motor in real time and feeds back the position to the audio processor; the man-machine interaction module is used for selecting an array form mode; according to the system, by adjusting the spatial layout of the movable microphones, the dynamic switching of the array form among a spherical form, a circular form and a linear form is realized. According to the invention, the flexibility and adaptability of the microphone array are improved.

Description

Technical Field

[0001] This application relates to the technical fields of speech recognition and sound source localization, for example, relates to a reconfigurable microphone array system and its control method. Background Art

[0002] The microphone array works through the cooperation of multiple microphones, and uses the time difference of arrival (TDOA) and phase difference of arrival (PDOA) of sound waves to achieve three-dimensional space sound source localization, beamforming (directional sound pickup), and adaptive noise suppression.

[0003] The current technology presents two major characteristics: highly integrated hardware and mature algorithm systems.

[0004] In terms of hardware, the mainstream solutions adopt the deep integration of MEMS microphone arrays and dedicated audio processing chips (such as XMOS XVF series, Allwinner R328, Rockchip RK3308), support multi-channel synchronous sampling and low-power operation, and meet the requirements of real-time signal processing.

[0005] In terms of algorithms, from the perspective of traditional signal processing, for example, sound source localization based on GCC-PHAT, echo cancellation based on LMS adaptive filtering, and noise suppression based on Wiener filtering have formed a standardized framework.

[0006] From the perspective of deep learning empowerment, complex scene sound source separation (such as the cocktail party problem) and end-to-end speech enhancement (such as the WaveNet noise reduction model) are realized through CNN / RNN networks, significantly improving the non-stationary noise suppression ability.

[0007] Although the algorithms and hardware have been highly mature, the fixed array structure has become a performance bottleneck:

[0008] 1. Rigid spatial layout

[0009] The microphone spacing and array topology (linear / ring / plane) are fixed during the device design stage and are difficult to match the requirements of different acoustic environments. For example: for far-field sound pickup, it is necessary to expand the array aperture (increase the microphone spacing), while for near-field interaction, a high-density microphone layout is required; in a reverberant environment, a differential array is needed to suppress reflected waves, and in an open space, a wideband response array is required.

[0010] 2. Static scene adaptation

[0011] A single array configuration cannot dynamically balance the beam width (directivity) and frequency response. For example:

[0012] In a conference scene, a narrow beam is required to suppress sidelobe interference, while in a smart home scene, a wide beam is required to cover multiple sound sources.

[0013] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present application, and thus may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0014] To have a basic understanding of some aspects of the disclosed embodiments, a simple summary is given below. This summary is not a general review, nor is it intended to identify key / important elements or delineate the scope of protection of these embodiments, but rather serves as a preface to the subsequent detailed description.

[0015] Embodiments of the present disclosure provide a reconfigurable microphone array system and its control method to solve the problem of poor flexibility and adaptability of the microphone array.

[0016] In some embodiments, the reconfigurable microphone array system includes:

[0017] A flexible substrate for carrying the microphone array;

[0018] A microphone array module including at least 2 fixed microphones and 4 movable microphones, where the movable microphones are connected to the flexible substrate through magnetic attraction sliders;

[0019] A motor drive module including multiple micro servo motors and drive circuits for controlling the movement of the movable microphones along a preset trajectory;

[0020] A multimodal audio processor integrating a sound source localization algorithm and a 3A audio processing algorithm for parsing user instructions, generating motor control signals, and processing audio data;

[0021] A closed-loop feedback module including multiple travel switches for real-time detecting the motor position and feeding it back to the audio processor;

[0022] A human-computer interaction module supporting at least one of a host computer graphical interface, physical buttons, and voice instruction input for selecting an array form mode;

[0023] The system realizes dynamic switching of the array form among spherical, circular, and linear by adjusting the spatial layout of the movable microphones.

[0024] Optionally, the flexible substrate is made of a bend-resistant polyimide material, supports dynamic deformation, has a thickness of 0.1 - 0.3 mm, a bending radius ≥ 5 mm, and can withstand more than 10 5 bending cycle times.

[0025] Optionally, the movement trajectories of the movable microphones include:

[0026] Linear motion: Displacement along the Z-axis direction, with a maximum stroke ≥ N millimeters, N ≥ 30;

[0027] Arc movement: Rotate around a preset center in the XY plane, with the rotation angle range of 0° to 180°.

[0028] Optionally, the microphone array includes 2 fixed microphones and 4 movable microphones;

[0029] The microphone array supports three preset forms:

[0030] Spherical mode: 6 microphones are distributed on the spherical surface with a radius of N millimeters, and the adjacent azimuth angles are separated by 90°. Among them, 2 fixed microphones are located on the same diameter of the spherical surface;

[0031] Circular mode: 2 fixed microphones and 2 movable microphones form a circular array in the XY plane with a radius of N millimeters, and the other 2 movable microphones are located on the diameter where the two fixed microphones are located;

[0032] Linear mode: 2 movable microphones are located on the straight line where the two fixed microphones are located, and the other two movable microphones are located at the midpoint of the connection line where the two fixed microphones are located.

[0033] Optionally, the travel switch is installed at the limit position of the motor movement, and sends a TTL level signal to the audio processor after being triggered.

[0034] Optionally, the micro servo motor is of the four-phase five-wire type 28BYJ-48 model, with a step angle of 5.625° / pulse, and the drive circuit uses a ULN2003 Darlington tube array;

[0035] The motor control signal is a PWM waveform, with an adjustable frequency of 1 kHz - 5 kHz and a duty cycle of 10% - 90%.

[0036] Optionally, the upper computer interface of the human-computer interaction module supports three-dimensional array form visualization and real-time displays the sound source azimuth angle and signal-to-noise ratio index;

[0037] The physical button uses a three-state anti-shake switch, with a hardware filtering time constant of 10 ms and a software anti-shake sampling interval of 5 ms.

[0038] Optionally, the sound source localization algorithm uses an improved time difference of arrival method, combined with generalized cross-correlation delay estimation;

[0039] The 3A audio processing algorithm includes automatic gain control, adaptive noise suppression, and echo cancellation.

[0040] In some embodiments, the control method for the aforementioned reconfigurable microphone array system includes the following steps:

[0041] Receive the mode selection instruction input by the user;

[0042] Parse the instruction and generate the motor motion sequence, and synchronously trigger the voice broadcast;

[0043] Drive the micro servo motor to move to the target position, and verify the in-place state through the limit switch;

[0044] Activate the sound source localization and beamforming algorithms to dynamically optimize the array directivity.

[0045] Optionally, when the sound source is detected to move, automatically trigger the array morphology adjustment; if the limit switch is not triggered for 3 consecutive motion timeouts, enter the fault lock state and turn on the alarm LED.

[0046] The reconfigurable microphone array system and its control method provided by the embodiments of the present disclosure can achieve the following technical effects:

[0047] Through the deep cooperation of the reconfigurable hardware architecture and intelligent algorithms, the present invention realizes a comprehensive improvement in acoustic performance: the deformable microphone array based on the flexible substrate, driven by the micro servo motor, dynamically adjusts the array geometry through closed-loop feedback, making the main lobe width of beamforming continuously adjustable from 15° (circular directional mode) to 60° (linear wide-area mode) according to the scene requirements, directly resulting in a 10 - 12 dB increase in the signal-to-noise ratio; combined with the 3A algorithm (AGC dynamic range 90 dB, ANS noise reduction 15 dB) and MVDR beamforming (sidelobe suppression -35 dB), the word error rate (WER) of speech recognition in an environment with a reverberation time of 1.2 s is optimized from 22% to 6%; at the hardware level, a hybrid layout of fixed / movable microphones (2 fixed + 4 movable) is adopted, reducing the system cost by 60% through dynamic hardware reuse. At the same time, relying on impedance control wiring (differential line 100Ω ± 10%) and low-power strategy (standby 0.5W), 8-meter far-field sound pickup and a morphology switching life of >50,000 times are achieved. A single device can cover multi-modal acoustic scenarios from personal near-field (within 1m + 15 dB gain) to large conference rooms (8m omnidirectional coverage). The measured speech recognition accuracy in meetings is increased by 21%, reaching the performance of professional-level systems.

[0048] The above general description and the following description are only exemplary and explanatory, and are not used to limit this application. Description of the Drawings

[0049] One or more embodiments are exemplarily illustrated by the corresponding drawings. These exemplary illustrations and the drawings do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are shown as similar elements. The drawings do not constitute a scale limitation, and among them:

[0050] Att Figure 1 : Block diagram of the reconfigurable microphone array system of the present invention;

[0051] AttFigure 2 : Spherical state diagram of the reconfigurable microphone array of the present invention;

[0052] Appendix Figure 3 : State diagram of the conversion of the spherical shape of the reconfigurable microphone array of the present invention into a circular shape;

[0053] Appendix Figure 4 : State diagram of the conversion of the spherical shape of the reconfigurable microphone array of the present invention into a linear shape. Detailed implementation manners

[0054] In order to more comprehensively understand the features and technical content of the embodiments of the present disclosure, the implementation of the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. The accompanying drawings are only for reference and explanation purposes and are not used to limit the embodiments of the present disclosure. In the following technical description, for the convenience of explanation, a sufficient understanding of the disclosed embodiments is provided through multiple details. However, one or more embodiments can still be implemented without these details. In other cases, well-known structures and devices can be shown in a simplified manner to simplify the drawings.

[0055] The terms "first", "second", etc. in the embodiments of the present disclosure are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so as to implement the embodiments of the present disclosure described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion.

[0056] In the embodiments of the present disclosure, the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "middle", "outer", "front", "rear", etc. is based on the orientation or positional relationship shown in the accompanying drawings. These terms are mainly used to better describe the embodiments of the present disclosure and their embodiments, and are not used to limit that the indicated devices, elements or components must have a specific orientation or be constructed and operated in a specific orientation. And, in addition to being able to represent the orientation or positional relationship, some of the above terms may also be used to represent other meanings. For example, the term "upper" may also be used to represent a certain attachment relationship or connection relationship in some cases. For those of ordinary skill in the art, the specific meanings of these terms in the embodiments of the present disclosure can be understood according to specific circumstances.

[0057] In addition, the terms "arrange", "connect", "fix" should be understood in a broad sense. For example, "connect" can be a fixed connection, a detachable connection, or an integral structure; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or there can also be internal communication between two devices, elements or components. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present disclosure can be understood according to specific circumstances.

[0058] Unless otherwise specified, the term "plurality" means two or more.

[0059] In the embodiments of the present disclosure, the character " / " indicates that the objects before and after are in an "or" relationship. For example, A / B means: A or B.

[0060] The term "and / or" is an associative relationship describing objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, the three relationships of A and B.

[0061] It should be noted that, without conflict, the embodiments in the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0062] Combined Figures 1 - 4 As shown, the embodiments of the present disclosure provide a reconfigurable microphone array system, including:

[0063] A flexible substrate for carrying the microphone array;

[0064] A microphone array module including at least 2 fixed microphones and 4 movable microphones, and the movable microphones are connected to the flexible substrate through magnetic attraction sliders;

[0065] A motor drive module including a plurality of micro servo motors and drive circuits for controlling the movable microphones to move along a preset trajectory;

[0066] A multimodal audio processor integrating a sound source localization algorithm and a 3A audio processing algorithm for parsing user instructions, generating motor control signals, and processing audio data;

[0067] A closed-loop feedback module including a plurality of travel switches for real-time detecting the motor position and feeding it back to the audio processor;

[0068] A human-computer interaction module supporting at least one of a host computer graphical interface, physical buttons, and voice instruction input for selecting an array form mode;

[0069] The system realizes dynamic switching of the array form among spherical, circular, and linear by adjusting the spatial layout of the movable microphones.

[0070] It can be understood that the reconfigurable microphone array system adopts a mechatronic design, integrates multimodal perception and dynamic reconfiguration capabilities, and the core modules include a microphone array module, a motion execution mechanism, an audio processor, a motor drive module, and an environment interaction interface.

[0071] The microphone array module is composed of 6 MEMS microphones and supports a three-dimensional space distributed layout.

[0072] The motion execution mechanism includes a micro servo motor and an intelligent control unit. The micro servo motor drives the movable microphones (Mic 1, Mic 3, Mic 5, Mic 6) to move along an arc / linear trajectory. At the same time, a magnetic adsorption slider is used to connect with the flexible substrate to ensure the positioning accuracy.

[0073] The audio processor (SoC) integrates a multi-core CPU and a dedicated DSP, which is responsible for signal processing and motion control. The audio processor (SoC) is a multi-modal audio processor that integrates a sound source localization algorithm and a 3A audio processing algorithm for parsing user instructions, generating motor control signals, and processing audio data.

[0074] The motor drive module adopts a closed-loop PID control algorithm to achieve precise adjustment of the motor speed and position.

[0075] The environmental interaction interface includes a wireless communication module and a voice interaction system. The wireless communication module supports dual-mode instruction transmission of Bluetooth 5.2 / BLE and USB3.0; the voice interaction system integrates a TTS engine and a far-field voice wake-up function.

[0076] Regarding the multi-modal human-computer interaction design, users can trigger the array reconstruction in two ways:

[0077] In the software interaction layer, the upper computer GUI provides a scene mode selection interface (round table meeting / long table meeting / personal mode), supporting drag-and-drop array preview.

[0078] In addition, it also supports natural language control (such as "switch to linear array") and combines voiceprint recognition to achieve permission management.

[0079] In the hardware control layer, a mode button matrix is set up to directly trigger the preset scene configuration through the GPIO interface, supporting blind operation design.

[0080] As an example, in the instruction parsing stage, the control unit receives a mode instruction (wireless / button), calls the preset kinematic model, and broadcasts the control parameters to each motor driver through the CAN bus.

[0081] In addition, a status feedback mechanism is set up. An optical encoder is used to monitor the motor rotation angle error (accuracy ±0.1°) in real time, and a Hall sensor array is used to detect the microphone in-place signal to form a closed-loop control.

[0082] On this basis, acoustic performance adaptive settings can also be carried out. For example: perform beamforming optimization, and automatically load the corresponding directivity algorithm according to the array form (such as the ring array adopts the super-directivity mode). Adopt a reverberation suppression strategy, enable sub-band adaptive filtering for the linear array, and start spatial difference processing for the ring array.

[0083] As another example, in a reconfigurable microphone array system, it mainly consists of a microphone array module, a motor and a drive module, an audio processor, an audio power amplifier, and a travel switch module. The specific system block diagram is as shown in Figure 1 shown.

[0084] In the human-machine interaction interface of the host computer, in the interface, the corresponding form of the microphone array can be selected according to the small conference round table, the conference rectangular table, and the personal meeting working mode. The selection instruction of the human-machine interaction interface is transmitted to the board through Bluetooth or the USB port. The audio processor on the board, which is also the CPU, receives the instruction transmitted through Bluetooth or the USB port and parses the instruction. After the audio processor parses the instruction, on the one hand, it performs voice broadcast to explain the current shape of the microphone array. The voice data of the voice broadcast is amplified by the audio power amplifier and then played through the speaker. After the audio processor parses the instruction, on the other hand, it controls the motor driver, and the motor driver controls the corresponding micro-motor to move. After the micro-motor moves to the specified position, the travel switch feeds back a digital signal with a high level to the audio processor. After the audio processor receives the signal fed back by the travel switch, it controls the motor driver, and the corresponding micro-motor stops moving.

[0085] Similarly, for the corresponding mode button, the working mode can also be selected. The mode button signal is connected to the GPIO pin of the audio processor. After the audio processor parses and processes it, on the one hand, it controls the micro-motor to move, and on the other hand, it controls the speaker to perform voice broadcast.

[0086] Optionally, the flexible substrate is made of a bend-resistant polyimide material, supports dynamic deformation, has a thickness of 0.1 - 0.3 mm, a bending radius ≥ 5 mm, and can withstand more than 10 5 bending cycles.

[0087] Optionally, the movement trajectory of the movable microphone includes:

[0088] Linear motion: Displacement along the Z-axis direction, with a maximum stroke ≥ N millimeters, N ≥ 30;

[0089] Arc motion: Rotation around a preset center in the XY plane, with a rotation angle range of 0° - 180°.

[0090] The microphone array includes 2 fixed microphones and 4 movable microphones;

[0091] The microphone array supports three preset forms:

[0092] Spherical mode: 6 microphones are distributed on a sphere with a radius of N millimeters, and the adjacent azimuth angles are separated by 90°. Among them, 2 fixed microphones are located on the same diameter of the sphere;

[0093] Circular pattern: Two fixed microphones and two movable microphones form a circular array in the XY plane with a radius of N millimeters. Another two movable microphones are located on the diameter where the two fixed microphones are located.

[0094] Linear pattern: Two movable microphones are located on the straight line where the two fixed microphones are located, and the other two movable microphones are located at the midpoint of the line connecting the two fixed microphones.

[0095] Optionally, the travel switch is installed at the limit position of the motor movement and sends a TTL level signal to the audio processor after being triggered.

[0096] It can be understood that the system supports millisecond-level switching among three basic array forms:

[0097] Array form Number of activated microphones Coverage range Applicable scenarios Reconstruction time Spherical array 6 microphones Omnidirectional 360° Complex environment with multiple sound sources - Ring array 4 microphones 180° plane Round table meeting / Vehicle-mounted 850 ms Linear array 4 microphones 120° fan Long table meeting / Directional sound pickup 1.2s

[0098] Specifically, as Figure 2 shown, the reconfigurable microphone array system consists of six microphones. The six microphones can form three forms: circular, spherical, and linear according to different environments.

[0099] Microphone 1 is located at point C1, microphone 2 is located at point E1, microphone 3 is located at point D1, microphone 4 is located at point F1, microphone 5 is located at point A1, and microphone 6 is located at point B1.

[0100] Microphone 2, microphone 4, microphone 5, and microphone 6 are located at the four rounded corners of the board. The four microphones are on the same circular cross-section and are symmetrically distributed about the center of the circle. The adjacent two microphones form a 90-degree angle.

[0101] Microphone 1, microphone 2, microphone 3, and microphone 4 are located on the same circular vertical cross-section and are symmetrically distributed about the center of the circle. The adjacent two microphones form a 90-degree angle.

[0102] Microphone 1, microphone 3, microphone 5, and microphone 6 are located on the same circular vertical cross-section and are symmetrically distributed about the center of the circle. The adjacent two microphones form a 90-degree angle.

[0103] In the reconfigurable microphone array system, microphone 2 and microphone 4 are fixed microphones and their positions on the board remain unchanged. Microphone 1, microphone 3, microphone 5, and microphone 6 are movable microphones and can be dragged to the specified positions by the motor according to the mode.

[0104] The default form of this reconfigurable microphone array system is spherical, as Figure 2 shown. The spherical form consists of six microphones. The six microphones are located on the spherical surface, and any adjacent two microphones form a 90-degree angle.

[0105] According to the working mode, if a circular microphone array is required, then the reconfigurable microphone array system needs to complete the conversion of the microphone array from spherical to circular.

[0106] The conversion of the microphone array from spherical to circular is as Figure 3 shown. After receiving the instruction, the CPU controls the motors on the sides of microphone 5 and microphone 6. Microphone 5 moves counterclockwise by 135 degrees with O2 as the center and the distance between point A1 and O2 as the radius, and stops moving at point A2. Microphone 6 moves counterclockwise by 135 degrees with O3 as the center and the distance between point B1 and O3 as the radius, and stops moving at point B2. That is, the circular microphone array is composed of microphone 1, microphone 3, microphone 2, and microphone 4. Microphone 5 and microphone 6 move to the specified positions and contract, and do not participate in the array composition.

[0107] According to the working mode, if a linear microphone array is required, then the reconfigurable microphone array system needs to complete the conversion of the microphone array from spherical to linear. The conversion of the microphone array from spherical to linear is as Figure 3 and Figure 4 shown. After receiving the instruction, the CPU controls the motors on the sides of microphone 1, microphone 3, microphone 5, and microphone 6. Among them, microphone 5 moves counterclockwise by 135 degrees with O2 as the center and the distance between point A1 and O2 as the radius, and stops moving at point A2. Microphone 6 moves counterclockwise by 135 degrees with O3 as the center and the distance between point B1 and O3 as the radius, and stops moving at point B2. Among them, microphone 1 moves from point C1 towards the center O1, and the movement trajectory is a straight line. It stops moving when it touches the travel switch above O1. Microphone 2 moves from point D1 towards the center O1, and the movement trajectory is a straight line. It stops moving when it touches the travel switch below O1.

[0108] That is, the basic layout principle of this embodiment is to use two fixed microphones (microphone 2 / microphone 4) as the X-axis to form a reference coordinate system to ensure the spatial reference during the form switching.

[0109] The movement trajectories of the movable microphones are as follows:

[0110] Microphone 1 / microphone 3: Move linearly along the Z-axis direction (travel 0 - 50 mm).

[0111] Microphone 5 / microphone 6: Move along an arc in the XY plane (radius R = 30 mm, angle range 0 - 180°).

[0112] During the movement execution stage, in the circular mode: Microphone 5 / microphone 6 contract along a 135° arc trajectory with O2 / O3 as the center to point A2 / B2.

[0113] Linear mode: Wheat 1 / Wheat 3 converges towards the center O1 along a straight-line trajectory and stops after triggering the travel switch.

[0114] Optionally, the micro servo motor is of the four-phase five-wire type 28BYJ-48, with a step angle of 5.625° / pulse, and the drive circuit uses a ULN2003 Darlington transistor array;

[0115] The motor control signal is a PWM waveform, with an adjustable frequency of 1 kHz - 5 kHz and a duty cycle of 10% - 90%.

[0116] Optionally, the upper computer interface of the human-machine interaction module supports three-dimensional array form visualization, and real-time displays the sound source azimuth angle and signal-to-noise ratio index; the physical buttons use a three-state anti-shake switch, with a hardware filtering time constant of 10 ms and a software anti-shake sampling interval of 5 ms.

[0117] Optionally, the sound source localization algorithm uses an improved time difference of arrival method, combined with generalized cross-correlation delay estimation; the 3A audio processing algorithm includes automatic gain control, adaptive noise suppression, and echo cancellation.

[0118] In some embodiments, the control method for the aforementioned reconfigurable microphone array system includes the following steps:

[0119] Receive a mode selection instruction input by the user;

[0120] Parse the instruction and generate a motor motion sequence, and synchronously trigger voice broadcast;

[0121] Drive the micro servo motor to move to the target position, and verify the in-place state through the travel switch;

[0122] Activate the sound source localization and beamforming algorithms to dynamically optimize the array directivity.

[0123] Optionally, when it is detected that the sound source moves, the array form adjustment is automatically triggered; if the travel switch is not triggered after 3 consecutive motion timeouts, it enters the fault locking state and lights up the alarm LED.

[0124] Through experiments, the reconfigurable microphone array system of this application has the following advantages compared with the existing microphone arrays:

[0125] 1. Significantly improved acoustic performance

[0126] In terms of signal-to-noise ratio (SNR) optimization, the following measured data are available:

[0127] Scene SNR of traditional array SNR of this solution Improvement amplitude Round table meeting 58.2 dB 70.1 dB +11.9 dB Long table discussion 52.4 dB 63.8 dB +11.4 dB High reverberation environment 47.6 dB 59.3 dB +11.7 dB

[0128] During use, the array form dynamically matches the sound field distribution (e.g., the circular mode enhances the gain in the main direction by 10 dB); and the 3A algorithm and beamforming (MVDR) are used to jointly suppress noise and reverberation.

[0129] In terms of positioning accuracy and speed angle, the accuracy is controlled at: ±1.2° (at a distance of 1 m), which is a 65% improvement compared to the traditional fixed array (±3.5°); the update rate reaches 20 Hz for real-time positioning, supporting dynamic sound source tracking (such as a moving speaker); in a typical scenario: in a 5-person round-table meeting, the system can lock and switch the orientation of the speaker within 0.5 s.

[0130] 2. Hardware efficiency and cost advantages

[0131] Regarding the hardware reuse rate, currently a single device supports 3 array forms, replacing 3 sets of traditional independent devices, and the hardware cost is reduced by 62%.

[0132] Working mode Power consumption Standby power consumption Traditional array 5.2W 1.8W This solution 3.5W 0.5W

[0133] This application adopts energy-saving technologies, such as the motor starts and stops on demand (duty cycle < 15%), and the audio processor dynamically adjusts the voltage (0.9V - 1.2V), etc.

[0134] Regarding the reliability indicators as follows:

[0135] Project Parameter Motor life > 50,000 times of form switching Bending resistance of the substrate > 100,000 times (radius 5 mm) Failure rate < 0.1 times per thousand hours

[0136] 3. Breakthrough in scene adaptability

[0137] Regarding multi-environment compatibility as follows:

[0138]

[0139] Regarding the dynamic response ability, the form switching speed is as follows:

[0140] Conversion type Time consumption Positioning accuracy recovery time Spherical → Circular 180 ms 220 ms Spherical → Linear 210 ms 250 ms

[0141] This application supports adaptive triggering. When the sound source continuously moves > 30°, the system automatically switches to the spherical mode within 400 ms.

[0142] 4. Scalability and intelligence

[0143] Regarding software-defined arrays, it supports adding preset forms (such as oval, cross) through the API without hardware modification, and opens the beamforming parameter interface, allowing developers to customize the directivity pattern.

[0144] Supports machine learning expansion, the voiceprint recognition module can be integrated to achieve speaker ID binding; adopts environmental noise classification (air conditioner / keyboard / traffic noise) to automatically match noise reduction strategies.

[0145] 5. Commercial verification data

[0146] Actual measurement case (a corporate meeting room)

[0147] Index Before transformation After transformation (this solution) Speech recognition accuracy 0.78 0.94 Manual correction rate of meeting records 0.35 0.08 Annual equipment maintenance cost 12000 3500

[0148] Through the collaboration of dynamic reconfigurable hardware and intelligent algorithms, this solution has achieved the following:

[0149] 1. Dramatic improvement in acoustic performance: SNR improvement > 10 dB, positioning accuracy < 1.5°, reaching the level of professional conference systems;

[0150] 2. Revolution in cost efficiency: Hardware cost reduced by 60%, energy consumption decreased by 35%, breaking through the cost-performance bottleneck of traditional arrays;

[0151] 3. Unlimited expansion of scenarios: From a 1m 2 personal space to a 100m 2 conference hall, a single device can cover the full-scenario voice interaction requirements;

[0152] 4. Intelligent evolution ability: The software programmable architecture provides an open hardware platform for AI voice processing.

[0153] In summary, the present invention proposes a reconfigurable microphone array system based on dynamic deformation and intelligent collaboration. The core lies in deeply collaborating the hardware programmable architecture with algorithm adaptive optimization to break through the physical limitations of traditional fixed arrays.

[0154] The system uses a polyimide flexible substrate as the deformation carrier. This material has both a high elastic modulus (2.5 GPa) and bend resistance characteristics (> 10 5 cycles), allowing 4 movable microphones to perform three-dimensional space reconstruction along a preset arc or straight-line trajectory under the precise drive of a micro servo motor. When the user selects the working mode through the host computer interface or physical buttons, the audio processor parses the instructions and generates a motor control sequence to drive the microphones to move with an accuracy of a step angle of 5.625°. At the same time, the in-place state is detected in real time through a travel switch to form a position-current double closed-loop control, ultimately achieving millimeter-level precise switching of the array geometry (positioning error < 0.5 mm).

[0155] This dynamic reconfiguration ability enables the array aperture to be intelligently adjusted within the range of 50 - 150 mm. For example, in the round-table meeting mode, the 6 microphones expand into a circular layout with a diameter of 100 mm. By optimizing the main lobe width (15°) and sidelobe suppression (-35 dB) of the beamforming algorithm, the speech signal-to-noise ratio in a complex environment is increased by 10 - 12 dB. When switched to the linear mode, the array shrinks into an axial linear structure. Combining the dynamic gain control (30 - 90 dB) and spectral subtraction noise reduction (15 dB) of the 3A algorithm, a clear voice pickup effect above 55 dB can be maintained within a distance of 8 meters.

[0156] The working principle of the system forms a complete technical closed-loop: the deformation and reconstruction at the physical layer directly change the spatial sampling characteristics of the array, enabling the sound source localization algorithm (based on generalized cross-correlation time-delay estimation) at the signal layer to lock the target direction with an accuracy of ±1.2°. At the same time, the dynamically updated array geometric parameters are input into the MVDR beamforming algorithm in real time to generate a spatial filter that matches the environmental noise characteristics, effectively suppressing reverberation and interfering sound sources.

[0157] This two-way empowerment of hardware and algorithm improves the accuracy of conference room speech recognition from 78% to 94% in engineering field tests, and at the same time reduces the hardware cost by 60% through modular design.

[0158] The system also innovatively introduces an energy efficiency management strategy, such as dynamic voltage regulation (0.9 - 1.2 V) and intelligent motor start-stop mechanism, reducing the standby power consumption to 0.5 W. While ensuring a reconstruction life of 50,000 times, it realizes multi-scenario adaptation from personal near-field directional enhancement (+15 dB gain within 1 meter) to large-space omnidirectional coverage, providing a hardware platform with both high performance and economy for intelligent voice interaction.

[0159] The above description and the drawings fully illustrate the embodiments of the present disclosure, enabling those skilled in the art to practice them. Other embodiments may include structural and other changes. The embodiments only represent possible variations. Unless explicitly required, the individual components and functions are optional, and the order of operations can vary. Some parts and features of some embodiments can be included in or replace those of other embodiments. The embodiments of the present disclosure are not limited to the structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A reconfigurable microphone array system, characterized in that, Comprising: A flexible substrate for carrying a microphone array; A microphone array module including at least two fixed microphones and four movable microphones, wherein the movable microphones are connected to the flexible substrate through magnetic attraction sliders; A motor drive module including a plurality of micro servo motors and drive circuits for controlling the movement of the movable microphones along a preset trajectory; A multi-modal audio processor integrating a sound source localization algorithm and a 3A audio processing algorithm for parsing user instructions, generating motor control signals, and processing audio data; A closed-loop feedback module including a plurality of travel switches for real-time detecting the motor position and feeding it back to the audio processor; A human-computer interaction module supporting at least one of a host computer graphical interface, physical buttons, and voice instruction input for selecting an array form mode; The system realizes dynamic switching of the array form among spherical, circular, and linear by adjusting the spatial layout of the movable microphones.

2. The system according to claim 1, wherein: The flexible substrate is made of bend-resistant polyimide material, supports dynamic deformation, has a thickness of 0.1 - 0.3 mm, a bending radius ≥ 5 mm, and can withstand more than 10 5 times of bending cycles.

3. The system according to claim 1, characterized in that: The movement trajectories of the movable microphones include: Linear movement: Displacement along the Z-axis direction with a maximum stroke ≥ N millimeters, N ≥ 30; Arc movement: Rotation around a preset center in the XY plane with a rotation angle range of 0° to 180°.

4. The system according to claim 3, wherein The microphone array includes two fixed microphones and four movable microphones; The microphone array supports three preset forms: Spherical mode: Six microphones are distributed on a spherical surface with a radius of N millimeters, and the adjacent azimuth angles are spaced 90°. Among them, two fixed microphones are located on the same diameter of the spherical surface; Circular mode: Two fixed microphones and two movable microphones form a circular array in the XY plane with a radius of N millimeters, and the other two movable microphones are located on the diameter where the two fixed microphones are located; Linear mode: Two movable microphones are located on the straight line where the two fixed microphones are located, and the other two movable microphones are located at the midpoint position of the connection line between the two fixed microphones.

5. The system according to any one of claims 1 to 4, wherein: The travel switches are installed at the limit positions of the motor movement and send TTL level signals to the audio processor after being triggered.

6. The system according to any one of claims 1 to 4, wherein: The micro servo motor is of the four-phase five-wire type 28BYJ-48 model with a step angle of 5.625° / pulse, and the drive circuit uses a ULN2003 Darlington tube array; The motor control signal is a PWM waveform with an adjustable frequency of 1 kHz - 5 kHz and a duty cycle of 10% - 90%.

7. The system according to any one of claims 1 to 4, wherein: The host computer interface of the human-computer interaction module supports three-dimensional visualization of the array form and real-time displays the sound source azimuth angle and signal-to-noise ratio index; The physical buttons use three-state anti-shake switches with a hardware filtering time constant of 10 ms and a software anti-shake sampling interval of 5 ms.

8. The system according to any one of claims 1 to 4, wherein: The sound source localization algorithm uses an improved time difference of arrival method combined with generalized cross-correlation time delay estimation; The 3A audio processing algorithm includes automatic gain control, adaptive noise suppression, and echo cancellation.

9. A control method for the reconfigurable microphone array system according to any one of claims 1 to 8, characterized in that, Including the following steps: Receive the mode selection instruction input by the user; Parse the instruction and generate a motor motion sequence, and simultaneously trigger voice broadcast; Drive the micro servo motor to move to the target position, and verify the in-place state through the limit switch; Activate the sound source localization and beamforming algorithms to dynamically optimize the array directivity.

10. The control method according to claim 9, wherein: When it is detected that the sound source moves, automatically trigger the adjustment of the array form; If the limit switch is not triggered due to motion timeout for three consecutive times, enter the fault lock state and turn on the alarm LED.

Images