Adaptive ultrasonic anti-recording interference method based on local AI human voice feature analysis

Abstract

This invention belongs to the field of audio signal processing and privacy protection technology, specifically relating to an adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis. The method involves: acquiring environmental audio signals through an audio acquisition module; performing voice feature analysis on the pre-processed audio signals; establishing a dual-microcontroller collaborative working mode; controlling the ultrasonic emission module to generate and drive the ultrasonic transducer array to emit ultrasonic interference signals; the second microcontroller dynamically adjusting interference parameters based on a closed-loop feedback compensation mechanism; and a power management module switching system power consumption modes according to the voice detection status to achieve multi-mode power management. This invention relies on a local lightweight AI model to complete voice feature extraction, achieving real-time voice detection and interference parameter generation without relying on a network. This eliminates network latency and avoids the privacy leakage risks of uploading audio data to the cloud, making it suitable for offline use scenarios of portable devices.

Description

Technical Field

[0001] This invention belongs to the field of audio signal processing and privacy protection technology, specifically relating to an adaptive ultrasonic anti-recording interference method based on local AI human voice feature analysis. Background Technology

[0002] With the widespread adoption and miniaturization of recording devices such as smartphones, voice recorders, and hidden recording devices, the demand for privacy protection in scenarios such as confidential meetings, business negotiations, and private conversations is increasing. Ultrasonic anti-recording interference technology, as an effective anti-recording method, works by emitting ultrasonic signals (frequency higher than 20kHz) that are inaudible to the human ear. Utilizing the non-linear distortion characteristics of the microphone in the recording device, it generates audible interference noise within the device, thereby compromising recording quality without affecting normal human communication.

[0003] Currently, there are various ultrasonic anti-recording interference technologies available both domestically and internationally, mainly including the following categories: 1. Fixed-frequency ultrasonic interference technology According to patent document CN220325622U, this utility model patent provides a circuit for preventing ultrasonic interference during recording, including a start / stop module, a power supply module, a control module, a signal generation module, and an ultrasonic sensing module. This technical solution uses an STM32F103C8T6 microcontroller as the core controller. The signal generation module generates an ultrasonic signal of a fixed frequency (e.g., 40kHz), which is then amplified and emitted by the ultrasonic transducer.

[0004] According to patent document CN119921894A, this invention patent proposes an anti-recording interference system that interferes with voice signals using random sound wave signals, causing the recording device to only record noise. The system includes components such as a controller, an ultrasonic probe array, a battery, and a power button.

[0005] 2. Round-robin / cyclic interference technology According to patent document CN121415795A, this invention patent proposes a method for preventing recording interference by outputting ultrasonic interference signals through an ultrasonic cyclic interference strategy. This cyclic interference strategy includes cyclically outputting a preset first mixed interference signal, a preset second mixed interference signal, and a preset third mixed interference signal according to a preset period, to cover more types of recording devices.

[0006] 3. Device identification interference technology According to patent document CN120074737A, this invention patent proposes a recording jamming method based on ultrasonic interference, which includes detecting and identifying recording devices in a target area and obtaining the type information of the recording devices; configuring a first ultrasonic interference signal and a second ultrasonic interference signal based on the type information; and performing recording interference on the target area according to the configured signals.

[0007] Academic research technology The academic research, titled "Research and Implementation of Ultrasonic Anti-Recording Shielding Based on Acoustic Parametric Array Theory," proposes an ultrasonic anti-recording shielding scheme based on acoustic parametric array theory. It utilizes the principle that multiple high-frequency ultrasonic waves generate low-frequency sound signals through a nonlinear system to effectively shield against recording and eavesdropping without interfering with normal conversation. Test results show that the maximum shielding distance is approximately 1.5 meters.

[0008] To address the shortcomings of the existing technologies, this invention aims to provide an adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis, dual microcontroller collaboration, low power consumption management, and closed-loop feedback compensation. This system solves problems such as poor interference adaptability, AI dependence on the cloud, insufficient real-time performance, excessive power consumption, and unstable interference effects in the existing technologies. Summary of the Invention

[0009] To address the technical problems existing in current ultrasonic anti-recording interference methods, this invention provides an adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis includes the following steps: S1. The ambient audio signal is acquired through the audio acquisition module, preprocessed and filtered by the signal processing module, and then transmitted to the first microcontroller. S2. The first microcontroller calls the local lightweight AI model to perform human voice feature analysis on the preprocessed audio signal, extract human voice gender, voice decibel, frequency distribution features, and generate adaptive interference parameters that match human voice features. S3. The first microcontroller sends the adaptive interference parameters to the second microcontroller through the communication interface, forming a dual-microcontroller collaborative working mode. S4. The second microcontroller controls the ultrasonic transmitting module to generate and drive the ultrasonic transducer array to transmit ultrasonic interference signals based on the adaptive interference parameters. S5. The monitoring microphone module collects the audio signal after interference, the effect evaluation module calculates the interference effect index, and the second microcontroller dynamically adjusts the interference parameters based on the closed-loop feedback compensation mechanism. S6. The power management module switches the system power consumption mode according to the human voice detection status to achieve multi-mode power consumption management.

[0011] The local lightweight AI model mentioned in step S2 is an embedded AI model with a quantized size of less than 1MB, which can complete voice detection, gender recognition, decibel classification, and frequency distribution analysis without a network connection.

[0012] The dual-microcontroller collaborative working mode described in step S3 is as follows: the first microcontroller focuses on the AI ​​analysis task, the second microcontroller focuses on the interference execution task, and the two controllers communicate through a communication interface using a UART with a baud rate of ≥1Mbps.

[0013] The frequency range of the ultrasonic interference signal in step S4 is 20kHz-33kHz, the frequency resolution is ≤1Hz, and the modulation method includes FM, AM and hybrid modulation.

[0014] The closed-loop feedback compensation mechanism in step S5 is as follows: the effect evaluation module calculates the audio signal-to-noise ratio and speech intelligibility index, and the second microcontroller dynamically adjusts the frequency, power and modulation method of the ultrasonic interference signal through the PID algorithm.

[0015] The multi-mode power management in step S6 includes four modes: sleep, standby, monitoring, and working. The power management module automatically switches power consumption states based on the voice detection results. The standby power consumption is lower than... .

[0016] An adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis includes a first microcontroller, a second microcontroller, an audio acquisition module, a monitoring microphone module, an ultrasonic transmission module, a communication interface, a power management module, a signal processing module, an effect evaluation module, and an ultrasonic transducer array. The power management module is electrically connected to the first microcontroller, the second microcontroller, and the ultrasonic transmitting module, respectively, and is used to provide power supply and multi-mode power consumption management for each module of the system. The first microcontroller is an AI analysis unit, which is electrically connected to the audio acquisition module, the monitoring microphone module, and the communication interface, and is used for human voice detection, AI human voice feature extraction, and generation of adaptive interference parameters. The second microcontroller is an interference execution unit, which is electrically connected to the communication interface and the signal processing module, respectively. It is used to receive interference parameters sent by the first microcontroller through the communication interface and control the generation and transmission of interference signals. The ultrasonic transmitting module is electrically connected to the ultrasonic transducer array and is used to drive the ultrasonic transducer array to transmit ultrasonic interference signals. The audio acquisition module, monitoring microphone module, and communication interface are electrically connected to the effect evaluation module to transmit audio signals and interactive data to the effect evaluation module in order to complete the interference effect evaluation. The signal processing module is electrically connected to the second microcontroller and is used to preprocess and filter interference-related signals.

[0017] Both the first and second microcontrollers use Xtensa LX7 dual-core 32-bit RISC architecture microprocessors. The first microcontroller natively supports AI inference, while the second microcontroller has a built-in DDS signal generator and PWM control unit. The ultrasonic transmitting module uses a Class D power amplifier with a power conversion efficiency of ≥85% and an adjustable output power of 0-10W, and is equipped with an impedance matching circuit.

[0018] The power management module uses a DC-DC step-down converter with a conversion efficiency of ≥90% and features multiple independent voltage regulation outputs to reduce electromagnetic interference between modules.

[0019] The dual-microcontroller communication protocol transmitted through the communication interface includes a frame header, command type, data length, interference frequency, transmission power, modulation method, human voice decibel, human voice gender, interference effect, checksum, and frame tail field.

[0020] Compared with the prior art, the beneficial effects of this invention are: 1. This invention adopts a dual-architecture design of a first microcontroller and a second microcontroller to decouple AI voice analysis from ultrasonic interference execution tasks, avoiding the latency bottleneck caused by single-controller task coupling; at the same time, it relies on a local lightweight AI model to complete voice feature extraction, and can realize real-time voice detection and interference parameter generation without relying on the network, which not only eliminates network latency, but also avoids the privacy leakage risk of uploading audio data to the cloud, and is suitable for offline use scenarios of portable devices.

[0021] 2. This invention generates adaptive ultrasonic interference parameters based on characteristics such as gender, decibel level, and frequency distribution of human voice, which can be specifically matched to the high-frequency response characteristics of different recording devices. At the same time, by monitoring the microphone module to collect the audio signal after interference, and combining it with the PID algorithm to dynamically adjust the frequency, power and modulation mode of the interference signal, a closed-loop feedback compensation mechanism is formed to ensure that the interference effect is continuously stable and meets the standard, thereby improving the anti-interference reliability in complex environments.

[0022] 3. This invention achieves automatic switching between four power consumption modes—sleep, standby, monitoring, and working—through a power management module. It only enters working mode when human voice is detected, and its standby power consumption is lower than [previous level]. This significantly reduces system energy consumption and extends equipment battery life; at the same time, the use of a 360° circular ultrasonic transducer array achieves omnidirectional ultrasonic interference coverage without blind spots, balancing interference range and energy utilization efficiency. Attached Figure Description

[0023] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0024] The structures, proportions, sizes, etc. illustrated in this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0025] Figure 1 This is a hardware connection diagram of the system of the present invention; Figure 2 This is a schematic diagram of the first microcontroller circuit of the present invention; Figure 3 This is a schematic diagram of the second microcontroller circuit of the present invention; Figure 4 This is a circuit diagram of the ultrasonic transmitting module of the present invention; Figure 5 This is a circuit schematic diagram of the power management module of the present invention; Figure 6 This is a flowchart of the software program for the first microcontroller of the present invention; Figure 7 This is a flowchart of the software program for the second microcontroller of the present invention; Figure 8 This is a complete flowchart of the system workflow of the present invention.

[0026] Wherein: 1 is the first microcontroller, 2 is the second microcontroller, 3 is the audio acquisition module, 4 is the monitoring microphone module, 5 is the ultrasonic transmitting module, 6 is the communication interface, 7 is the power management module, 8 is the signal processing module, 9 is the effect evaluation module, and 10 is the ultrasonic transducer array. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. These descriptions are only for further illustrating the features and advantages of the present invention, and not for limiting the claims of the present invention. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0029] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0030] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0031] I. System Hardware Composition and Electrical Connections This embodiment discloses an adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis, such as... Figure 1-5 As shown, it consists of a first microcontroller 1, a second microcontroller 2, an audio acquisition module 3, a monitoring microphone module 4, an ultrasonic transmitting module 5, a communication interface 6, a power management module 7, a signal processing module 8, an effect evaluation module 9, and an ultrasonic transducer array 10. The electrical connections and spatial layout of each component are as follows: 1. Electrical connection The power management module 7 is electrically connected to the first microcontroller 1, the second microcontroller 2, and the ultrasonic transmitting module 5, respectively, to provide power supply and power consumption control for all modules of the system; the first microcontroller 1 is electrically connected to the audio acquisition module 3 and the signal processing module 8 in sequence, and also communicates bidirectionally with the second microcontroller 2 through the communication interface 6; the second microcontroller 2 is electrically connected to the ultrasonic transmitting module 5, and the ultrasonic transmitting module 5 drives the ultrasonic transducer array 10; the monitoring microphone module 4 is electrically connected to the effect evaluation module 9, and the effect evaluation module 9 feeds back to the second microcontroller 2, forming a closed-loop feedback link.

[0032] 2. Spatial Layout The audio acquisition module 3 is located directly above the device and behind the interference area. It is treated with acoustic foam for noise reduction to prevent its own interference signals from entering. The monitoring microphone module 4 is located on the side of the device, at a 90° angle to the ultrasonic transmitting module 5, to avoid directly receiving the transmitted signal and only to collect the reflected audio after interference. The ultrasonic transducer array 10 is arranged in a 360° ring around the device to achieve interference-free coverage. The first microcontroller 1 is located on the left side of the motherboard, close to the audio acquisition module 3, to reduce signal transmission loss. The second microcontroller 2 is located on the right side of the motherboard, close to the ultrasonic transmitting module 5, to improve control response speed. The power management module 7 is located in the independent power supply area at the bottom of the motherboard to reduce electromagnetic interference.

[0033] 3. Hardware parameter selection First microcontroller 1 and second microcontroller 2: Both adopt Xtensa LX7 dual-core 32-bit RISC architecture microprocessors; First microcontroller 1 natively supports AI inference, is equipped with a 24-bit high-precision ADC, a 48kHz sampling interface, and external Flash storage for lightweight AI models less than 1MB after quantization; Second microcontroller 2 has a built-in DDS signal generator, PWM control unit and monitoring ADC.

[0034] Ultrasonic transmitting module 5: adopts Class D power amplifier, power conversion efficiency ≥85%, output power 0-10W continuously adjustable, integrated impedance matching circuit; signal frequency range 20kHz-33kHz, frequency resolution ≤1Hz.

[0035] Power Management Module 7: Employs a DC-DC step-down converter with a conversion efficiency ≥90%, featuring multiple independent regulated outputs and supporting power consumption switching between four modes: sleep, standby, monitoring, and operating. Standby power consumption... .

[0036] Communication Interface 6: UART communication is used, with a baud rate of ≥1Mbps. The transmission protocol includes frame header, command type, data length, interference frequency, transmission power, modulation method, human voice decibel, human voice gender, interference effect, checksum, and frame tail field.

[0037] II. Specific Implementation of Methods and Steps This embodiment fully executes the adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to steps S1-S6, such as... Figure 6-8 As shown, the specific implementation details are as follows: S1 Audio Acquisition and Preprocessing The audio acquisition module 3 acquires environmental audio signals from confidential scenarios such as conference rooms and office areas in real time and transmits them to the signal processing module 8. The signal processing module 8 performs high-pass filtering, noise reduction, and adaptive gain adjustment preprocessing on the audio signals to remove environmental noise and non-target frequency band interference, and transmits the clean audio signals to the ADC interface of the first microcontroller 1.

[0038] S2 Local AI Voice Feature Analysis and Interference Parameter Generation The first microcontroller 1 calls a lightweight, embedded local AI model with a quantized size of <1MB, without requiring a network connection, to perform AI analysis on the preprocessed audio signal: Voice detection: Distinguishes human voices from environmental noise and filters out invalid signals; Feature extraction: Identify speaker gender (male / female / unknown), speech decibel (30-120dB), and human voice frequency distribution characteristics; Parameter generation: Based on the extracted human voice features, adaptive interference parameters are generated, including interference frequency, transmission power level, and modulation method (FM / AM / hybrid modulation).

[0039] S3 Dual Microcontroller Collaborative Communication The first microcontroller 1 serves as the AI ​​analysis unit, focusing on human voice feature analysis and parameter generation; the second microcontroller 2 serves as the interference execution unit, focusing on signal generation and transmission control; the two controllers transmit data through the communication interface 6 at a UART baud rate of ≥1Mbps, and the first microcontroller 1 sends adaptive interference parameters to the second microcontroller 2 to achieve task separation and collaborative work.

[0040] S4 Ultrasonic Interference Signal Generation and Transmission After receiving the adaptive interference parameters, the second microcontroller 2 generates a high-precision ultrasonic signal through the built-in DDS signal generator. The transmission power is controlled by the PWM unit, which drives the ultrasonic transmission module 5 to amplify the signal power. The ultrasonic transmission module 5 drives the 360° ultrasonic transducer array 10 to transmit ultrasonic interference signals through the impedance matching circuit. The signal frequency is 20kHz-33kHz, and the modulation mode is switched as needed to achieve precise interference that matches the characteristics of human voice.

[0041] S5 Closed-Loop Feedback Compensation Adjustment The monitoring microphone module 4 collects the audio signal after interference in the target area in real time and transmits it to the effect evaluation module 9. The effect evaluation module 9 calculates two core interference effect indicators, namely audio signal-to-noise ratio and speech intelligibility, and feeds the indicator data back to the second microcontroller 2. The second microcontroller 2 dynamically adjusts the frequency, power and modulation mode of the ultrasonic interference signal through the PID algorithm to complete the closed-loop feedback compensation and ensure the stability of the interference effect in complex acoustic environments.

[0042] S6 Multi-Mode Power Management Power management module 7 has a built-in power state machine that automatically switches the system power consumption mode based on human voice detection results. 1. Sleep Mode: Triggered when there is no noise, the system enters a low-power hibernation state, reducing power consumption. ; 2. Standby mode: Periodically wakes up to listen, with extremely low power consumption; 3. Monitoring Mode: Continuously acquires audio and performs AI voice detection without emitting interference signals; 4. Working mode: Triggered when human voice is detected, it operates at full capacity and emits interference signals.

[0043] Through four-mode collaborative management, refined power consumption control is achieved, extending device battery life.

[0044] III. Overall System Workflow Upon system power-on initialization, the power management module 7 completes its power supply self-test and enters sleep mode by default. The audio acquisition module 3 periodically wakes up to acquire and preprocess ambient audio. The first microcontroller 1 performs local AI voice detection; if no voice is detected, it maintains low-power mode; if a voice is detected, it switches to monitoring mode. The first microcontroller 1 performs voice feature analysis, generates adaptive interference parameters, and sends them to the second microcontroller 2. The second microcontroller 2 controls the transmission module to generate and transmit ultrasonic interference signals, and the system enters working mode. The monitoring microphone and effect evaluation module provide real-time feedback on the interference effect, and the second microcontroller 2 dynamically adjusts the interference parameters using a PID algorithm. After the voice disappears, the system automatically switches back to standby / sleep mode, completing one working cycle.

[0045] IV. Verification of the Effects of the Examples In this embodiment, the system is deployed in a closed conference room setting to capture male voices (65dB, core frequency 2kHz-5kHz). A local AI model generates a 28kHz FM modulated ultrasonic interference signal with a transmission power of 30%. After closed-loop feedback adjustment, the audio signal-to-noise ratio captured by the recording device drops below -15dB, making the speech completely unrecognizable. The system's standby power consumption... The battery life is 4 times longer than that of traditional devices, and the interference coverage of mainstream devices such as mobile phones, voice recorders, and hidden recording devices is >95%, fully realizing the technical effects of this invention.

[0046] The above description only illustrates the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention, and all such changes should be included within the protection scope of the present invention.

Claims

1. An adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis, characterized in that, Includes the following steps: S1. The ambient audio signal is acquired by the audio acquisition module (3), preprocessed and filtered by the signal processing module (8), and then transmitted to the first microcontroller (1). S2. The first microcontroller (1) calls the local lightweight AI model to perform human voice feature analysis on the preprocessed audio signal, extract the human voice gender, voice decibel, frequency distribution features, and generate adaptive interference parameters that match the human voice features. S3. The first microcontroller (1) sends the adaptive interference parameters to the second microcontroller (2) through the communication interface (6) to form a dual microcontroller collaborative working mode. S4. The second microcontroller (2) controls the ultrasonic transmitting module (5) to generate and drive the ultrasonic transducer array (10) to transmit ultrasonic interference signals according to the adaptive interference parameters. S5. The audio signal after interference is collected by the monitoring microphone module (4), the effect evaluation module (9) calculates the interference effect index, and the second microcontroller (2) dynamically adjusts the interference parameters based on the closed-loop feedback compensation mechanism. S6. Power management module (7) switches the system power consumption mode according to the human voice detection status to realize multi-mode power consumption management.

2. The adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to claim 1, characterized in that: The local lightweight AI model mentioned in step S2 is an embedded AI model with a quantized size of less than 1MB, which can complete voice detection, gender recognition, decibel classification, and frequency distribution analysis without a network connection.

3. The adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to claim 1, characterized in that: The dual-microcontroller collaborative working mode described in step S3 is as follows: the first microcontroller (1) focuses on the AI ​​analysis task, the second microcontroller (2) focuses on the interference execution task, and the two controllers communicate through the communication interface (6) with a UART baud rate ≥1Mbps.

4. The adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to claim 1, characterized in that: The frequency range of the ultrasonic interference signal in step S4 is 20kHz-33kHz, the frequency resolution is ≤1Hz, and the modulation method includes FM, AM and hybrid modulation.

5. The adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to claim 1, characterized in that: The closed-loop feedback compensation mechanism in step S5 is as follows: the effect evaluation module (9) calculates the audio signal-to-noise ratio and speech intelligibility index, and the second microcontroller (2) dynamically adjusts the frequency, power and modulation method of the ultrasonic interference signal through the PID algorithm.

6. The adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis according to claim 1, characterized in that: The multi-mode power management in step S6 includes four modes: sleep, standby, monitoring, and working. The power management module (7) automatically switches the power consumption state according to the human voice detection result. The standby power consumption is lower than that of the human voice detection result. .

7. An adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis, wherein the system is used in the adaptive ultrasonic anti-recording interference method based on local AI voice feature analysis as described in any one of claims 1-6, characterized in that: It includes a first microcontroller (1), a second microcontroller (2), an audio acquisition module (3), a monitoring microphone module (4), an ultrasonic transmitting module (5), a communication interface (6), a power management module (7), a signal processing module (8), an effect evaluation module (9), and an ultrasonic transducer array (10). The power management module (7) is electrically connected to the first microcontroller (1), the second microcontroller (2), and the ultrasonic transmitting module (5) respectively, and is used to provide power supply and multi-mode power consumption management for each module of the system; The first microcontroller (1) is an AI analysis unit, which is electrically connected to the audio acquisition module (3), the monitoring microphone module (4), and the communication interface (6) respectively, and is used for human voice detection, AI human voice feature extraction and generation of adaptive interference parameters; The second microcontroller (2) is an interference execution unit, which is electrically connected to the communication interface (6) and the signal processing module (8) respectively. It is used to receive the interference parameters sent by the first microcontroller (1) through the communication interface (6) and control the generation and transmission of interference signals. The ultrasonic transmitting module (5) is electrically connected to the ultrasonic transducer array (10) and is used to drive the ultrasonic transducer array (10) to emit ultrasonic interference signals. The audio acquisition module (3), monitoring microphone module (4), and communication interface (6) are electrically connected to the effect evaluation module (9) respectively, and are used to transmit audio signals and interactive data to the effect evaluation module (9) to complete the interference effect evaluation; The signal processing module (8) is electrically connected to the second microcontroller (2) and is used to preprocess and filter interference-related signals.

8. The adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis according to claim 7, characterized in that: Both the first microcontroller (1) and the second microcontroller (2) adopt the Xtensa LX7 dual-core 32-bit RISC architecture microprocessor. The first microcontroller (1) natively supports AI inference, and the second microcontroller (2) has a built-in DDS signal generator and PWM control unit. The ultrasonic transmitting module (5) adopts a Class D power amplifier with a power conversion efficiency of ≥85% and an adjustable output power of 0-10W. It has an impedance matching circuit.

9. The adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis according to claim 7, characterized in that: The power management module (7) adopts a DC-DC step-down converter with a conversion efficiency of ≥90% and has multiple independent voltage regulation outputs to reduce electromagnetic interference between modules.

10. The adaptive ultrasonic anti-recording interference system based on local AI voice feature analysis according to claim 7, characterized in that: The communication interface (6) transmits a dual microcontroller communication protocol that includes a frame header, command type, data length, interference frequency, transmission power, modulation method, human voice decibel, human voice gender, interference effect, checksum, and frame tail field.

Images