Radio station control system and radio station audio noise reduction method
By employing variational mode decomposition and wavelet denoising methods in the radio control system, the noise interference problem of radio audio signals in complex battlefield environments was solved, enabling efficient audio communication in complex environments.
Patent Information
- Application Number
- CN202410957197.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-08
AI Technical Summary
Existing radio noise reduction methods are ineffective in handling various noise interferences in complex battlefield environments, leading to a decrease in audio signal clarity and reliability, and affecting communication accuracy.
The radio control system, consisting of an audio acquisition module, a core processor, and a display control module, processes noisy frequency signals through variational mode decomposition and wavelet denoising, and combines automatic gain control and an improved threshold function for noise reduction.
It improves the clarity and reliability of audio communication in complex environments, ensures that the signal is within the appropriate range, reduces computational complexity and the impact on signal quality, and enhances the noise reduction effect of military radios.
Smart Images

Figure CN118921595A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of radio station control, and in particular to a radio station control system and a radio station audio noise reduction method. Background Art
[0002] With the rapid development of digital audio technology and the popularization of mobile smart devices, digital audio information has penetrated into all aspects of social life, providing great convenience for people. In this context, the widespread application of audio communication equipment is particularly important. They have become indispensable equipment in daily life, industrial production, agriculture and other fields. These devices not only effectively make up for the shortcomings of traditional communication equipment in stability and transmission speed, but also significantly improve the efficiency and quality of information transmission.
[0003] Especially in the military field, audio communication technology is crucial for real-time understanding of the frontline situation and making quick and accurate decisions. It plays an irreplaceable role in intelligence collection, combat command and communication coordination. Advanced audio communication technology makes military operations more flexible and agile, significantly improving combat efficiency.
[0004] With the continuous updating and iteration of radio technology, audio noise reduction methods are also constantly developing. However, different types of noise pose different challenges to noise reduction performance. The noise reduction methods of ordinary radio stations are mainly aimed at stable signals. However, in the actual battlefield environment, noise pollution is a common and unavoidable problem. Various weapon shooting sounds, explosions, human conversations, and even natural sounds such as wind and water can seriously affect the clarity and reliability of audio signals. This interference can cause audio signal distortion, which in turn affects the accuracy of communication and may even lead to the failure of combat missions. In some special application areas, non-stationary noise and random noise are more prominent. Summary of the invention
[0005] The purpose of the present invention is to provide a radio station control system and a radio station audio noise reduction method, so that the radio station can perform audio communication more effectively in various complex environments.
[0006] To achieve the above object, the present invention provides the following solutions:
[0007] A radio station control system, comprising:
[0008] An audio acquisition module, used to acquire audio signals containing noise;
[0009] A core processor connected to the audio acquisition module, used for performing variational mode decomposition on the noisy audio signal and performing noise reduction using a wavelet noise reduction method;
[0010] A display main control module, connected to the core processor, for performing signal conversion on the noisy audio signal and the noise-reduced audio signal, and converting the HDMI signal into an RGB signal;
[0011] The core processor includes:
[0012] An automatic gain control unit, used for adjusting the gain of the noisy audio signal and the noise-reduced audio signal;
[0013] The noise reduction unit is connected to the automatic gain control unit and is used to perform variational mode decomposition on the noisy audio signal after gain adjustment and to perform noise reduction by using a wavelet noise reduction method.
[0014] To achieve the above object, the present invention also provides the following solution:
[0015] A radio audio noise reduction method is applied to the above-mentioned radio control system, and the radio audio noise reduction method comprises:
[0016] The audio signal containing noise is collected through the audio collection module;
[0017] Performing automatic gain control on the noisy audio signal;
[0018] Perform variational mode decomposition on the gain-adjusted noisy audio signal to obtain multiple IMF components;
[0019] Based on multiple IMF components, the wavelet denoising method is used to perform denoising processing to obtain the denoised audio signal.
[0020] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0021] The radio control system and radio audio noise reduction method provided by the present invention, in the audio noise reduction processing, the automatic gain control unit in the core processor adjusts the audio signal to control the signal within a certain range, avoiding the problem of signal distortion and loss due to being too low or too high, and then uses a noise reduction unit to decompose the signal into multiple modal components (IMF), and uses a wavelet threshold method to achieve audio noise reduction. The present invention takes practicality and computational complexity as the starting point, while ensuring the noise reduction effect, controls the computational complexity and reduces the impact on the signal quality, can achieve good results in military radio audio noise reduction, and solves the problems faced by radio stations and audio processing. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0023] Figure 1 A module block diagram of the radio station control system provided by the present invention;
[0024] Figure 2 The module block diagram of the core processor;
[0025] Figure 3 It is a schematic diagram of the structure of the core processor;
[0026] Figure 4 This is the working flow diagram of the core processor;
[0027] Figure 5 To display the workflow diagram of the main control module;
[0028] Figure 6 This is the power supply block diagram of the power module;
[0029] Figure 7 is a schematic diagram of an audio circuit;
[0030] Figure 8 It is a schematic diagram of the hardware security interface circuit;
[0031] Fig. 9 This is a schematic diagram of the power on / off logic control circuit;
[0032] Fig.10 The heating function schematic diagram for pure hardware heating;
[0033] Fig.11 Thermal function schematic for software controlled heating;
[0034] Fig.12 is a schematic diagram of a USB functional circuit;
[0035] Fig.13 is a schematic diagram of an LVDS communication circuit;
[0036] Fig.14 is a schematic diagram of a network functional circuit;
[0037] Fig.15 is a schematic diagram of the key circuit;
[0038] Fig.16 This is a flowchart of signal noise reduction processing;
[0039] Fig.17is the automatic gain control flow chart;
[0040] Fig.18 It is the PKO processing flow chart;
[0041] Fig.19 is a comparison curve diagram of the soft and hard threshold functions and the improved threshold function;
[0042] Fig. 20 It is the PKO iteration curve diagram;
[0043] Fig.21 Schematic diagram of each IMF component after VMD decomposition;
[0044] Fig. 22 It is a schematic diagram of the spectrum distribution of each IMF component;
[0045] Fig.23 Schematic diagram of the traditional hard threshold for noise reduction analysis of noisy signals;
[0046] Fig.24 Schematic diagram of noise reduction analysis of noisy signals using soft threshold;
[0047] Fig.25 Schematic diagram of noise reduction analysis for a noisy signal to improve the threshold function. DETAILED DESCRIPTION
[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0049] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0050] Embodiment 1
[0051] This embodiment provides a radio station control system, such as Figure 1 As shown, the radio station control system includes: an audio acquisition module 1, a core processor 2 and a display main control module 3.
[0052] The audio acquisition module 1 is used to acquire audio signals containing noise.
[0053] The core processor 2 is connected to the audio acquisition module 1 and is used for performing variational mode decomposition on the noisy audio signal and performing noise reduction using a wavelet denoising method.
[0054] The display main control module 3 is connected to the core processor 2 and is used for performing signal conversion on the noisy audio signal and the noise-reduced audio signal, and converting the HDMI signal into an RGB signal.
[0055] Furthermore, in this embodiment, the communication noise comes from three aspects: first, it is generated at the signal sending end. During the audio communication process, it will be affected by the steady-state noise, non-steady-state noise and pulse noise in the environment. After being superimposed with the useful signal, they are transformed into electrical signals through the audio acquisition interface and transmitted to the receiving end; second, it is the electrical noise generated by the signal processing equipment and the electrical noise in the transmission channel; third, it is generated by the signal receiving end. When receiving the signal, it will also be affected by the noise of the environment. The above three situations will cause the useful signal to be interfered with to varying degrees, resulting in low accuracy of audio communication and poor communication effect.
[0056] In this embodiment, earphones and microphones are used as the collection ends of the audio signals containing noise.
[0057] Furthermore, if Figure 2 As shown, the core processor 2 includes an audio power amplifier unit 21, an audio decoding unit 22, an A / D conversion unit 23, an automatic gain control unit 24, a noise reduction unit 25, a D / A conversion unit 26 and an audio encoding unit 27.
[0058] The audio power amplifier unit 21 is connected to the audio acquisition module 1 and is used to amplify the audio signal containing noise.
[0059] The audio power amplifier unit 21 adopts a mono audio power amplifier to amplify the audio input signal.
[0060] The audio decoding unit 22 is connected to the audio power amplifier unit 21 and is used to decode the amplified audio signal containing noise.
[0061] The A / D conversion unit 23 is connected to the audio decoding unit 22 and is used to perform analog-to-digital conversion on the decoded audio signal containing noise.
[0062] The automatic gain control unit 24 is used to adjust the gain of the noisy audio signal and the noise-reduced audio signal.
[0063] The noise reduction unit 25 is connected to the automatic gain control unit 24 and is used to perform variational mode decomposition on the noisy audio signal after gain adjustment and to perform noise reduction using a wavelet noise reduction method.
[0064] A D / A conversion unit 26, connected to the automatic gain control unit 25, for performing digital-to-analog conversion on the gain-adjusted and noise-reduced audio signal;
[0065] An audio encoding unit 27, connected to the D / A conversion unit 26, for encoding the audio signal after the digital-to-analog conversion and the noise reduction;
[0066] The audio power amplifier unit 21 is also connected to the audio encoding unit 27 for amplifying the encoded and noise-reduced audio signal.
[0067] In an exemplary embodiment, Figure 3 As shown, the core processor 2 has an HDMI display output signal, and GW2AR is an FPGA chip, the purpose of which is to convert the HDMI signal output by the core board into an RGB signal. The storage uses the core board's onboard 64GB eMMC and two 1GB memories. It is equipped with an audio interface, an Ethernet port, and a serial port. The Ethernet port is used to realize network communication and software upgrades, and the audio interface includes a headphone interface, a speaker interface, and a microphone interface. The serial port realizes the software upgrade of the processor inside the system. The MCU is used to control power on and off and the working status. The core processor 2 can also be connected to a mouse, keyboard, and buttons. The buttons and USB interface realize human-computer interaction on the display screen.
[0068] The processing flow of the core processor 2 is as follows: Figure 4 As shown, the audio acquisition module 1 captures the noisy audio signal, and the acquired signal is processed by the audio power amplifier (audio power amplifier unit 21), and then the analog signal is converted into a digital signal by the audio decoder (audio decoding unit 22) and the audio decoding process is completed, and the gain control is performed by the AGC controller (automatic gain control unit 24), and the ARM controls the noise reduction processing unit 25 to perform audio noise reduction processing on the decoded noisy audio signal. The processed audio signal is encoded by the audio encoder (audio encoding unit 27) to encode the audio signal converted into an analog signal, and then amplified by the audio power amplifier, and finally the noise reduction audio signal is played according to the specified path.
[0069] Furthermore, the display main control module 3 mainly completes information status display, whole machine control, human-machine interface interaction and log recording, and the main functions realized include:
[0070] (1) Display function: collect the work of the equipment through the internal network port and other communication interfaces, including the normal fault status of the whole machine, the power module status, the network status and the network status, etc., and display them through the display interface (including the display screen and the front panel indicator light display);
[0071] (2) Control function: Control of the equipment is completed through the human-machine interface, such as equipment self-test and other functions;
[0072] (3) Human-computer interaction function: Human-computer interaction processing is completed through the front panel display, buttons, USB mouse, USB keyboard, etc., and terminal data interaction processing is completed through display interface, control interface and DT device communication;
[0073] (4) Log recording and storage function: The working log of each module is received through the internal network port and recorded in the storage unit of the module. The log can also be sent to an external device through the network port.
[0074] The key peripheral interface driver of the display main control module 3 is a binary file written in C language, which runs under the domestic operating system to realize the corresponding key function response and display function. This function is realized by pressing the 0~9 numeric keys, F1~F5 keys and ENTER, ESE keys on the device. Through this peripheral interface driver, accurate response of key operation and correct display on the display screen can be ensured. The corresponding key function is realized.
[0075] The signal conversion of the display main control module 3 is processed by FPGA technology. The core chip processes the completed data and displays an HDMI signal output, which is converted into RGB signal after FPGA processing and connected to the LCD screen. In addition to completing the HDMI signal to RGB signal processing, the display main control module also preheats the LED screen. The LED screen displays the audio signal with noise and the audio signal after noise reduction.
[0076] Since the system is used in special fields and faces complex and changeable environmental and climate changes, when the device is used at low temperatures, the display screen will be affected by the low temperature and become black and cannot be used normally. Therefore, it is necessary to design a preheating function for the display screen to ensure that the display screen can start working normally at low temperatures. The temperature sensor of the hardware circuit senses the external temperature change, and then processes the digital-to-analog conversion (ADC). The data is compared with a given threshold. When it is lower than the threshold, the display screen preheating function is turned on. The functional structure diagram is shown in the figure below. Figure 5 shown.
[0077] Furthermore, the radio station control system further comprises: a power management module 4, which is used to convert the input voltage and provide power for the radio station control system.
[0078] The power management unit is mainly composed of DC-DC, LDO, battery, etc. Its main function is to realize voltage conversion, converting the input DC5V and DC12V power into the power supply required by the functional module. The power supply block diagram of the power management module 4 is as follows Figure 6 shown.
[0079] Furthermore, the station control system further comprises: a security interface for connecting with other security devices;
[0080] The confidentiality interface includes a key injection interface, a parameter injection interface and a key destruction interface.
[0081] A confidentiality interface has been added to the military radio system, through which it can be connected to other confidential devices, thus increasing security. The network port solves the problem of software upgrade and adopts a domestically produced design.
[0082] Furthermore, the radio station control system also includes: an audio output module for outputting the noise-reduced audio.
[0083] The earphone and microphone serve as both the collection end of the noisy audio signal and the output end of the noise-reduced audio signal; the speaker only serves as the output end of the audio signal after noise reduction.
[0084] The main circuit designs of the radio control system provided in this embodiment include: audio circuit design, hardware security interface circuit design, power on / off logic control circuit design, USB function circuit design, LVDS communication circuit design, network function circuit design and button circuit design.
[0085] (1) Audio circuit design
[0086] like Figure 7 As shown, the audio codec uses the digital decoding chip U10, the chip model is ES8388; the audio decoding is processed by software and then digitally decoded, and then output through the power amplifier.
[0087] The audio interface supports one headphone interface J5 input and output and a microphone input interface MIC. When the audio signal is input, it is amplified by the audio power amplifiers U11 and U12, and then digitally decoded by ES8388. The signal is input by the front panel aviation plug. The carrier unit is designed with an access connector. After the audio noise reduction software AGC processing and software noise reduction processing are performed in ARM, it is encoded by ES8388 and output from the headphone interface, microphone interface and speaker interface J6 through the power amplifier to the aviation plug board output, and externally connected in series with magnetic beads and capacitors for filtering.
[0088] (2) Hardware security interface circuit design
[0089] Figure 8 It is a hardware security circuit, which includes a key injection interface, a parameter injection interface and a key destruction port. The key injection interface J7 is a high-speed key injection interface. The signal is connected from the front panel, directly transferred to the motherboard interface through the carrier interface, and finally connected to the motherboard module and transferred to the security machine module. The parameter injection interface J9 is connected from the front panel. The signal is directly transferred to the motherboard interface through the carrier interface, and finally connected to the motherboard module and transferred to the IO module. The signal is connected from the front panel key destruction port J11, directly transferred to the motherboard interface through the carrier interface, and finally connected to the motherboard module.
[0090] (3) Design of power on / off logic control circuit
[0091] according to Fig. 9 It can be seen that in the power-on state, the panel power-on control switch button is pressed once, and the internal circuit of the display main control module generates a pulse signal (5.2V when not pressed, pulled down to 0V when pressed, and restored to 5.2V after popping up), which is transmitted to the interface module through the OC_DC_IO_PowerCtrl pin. The interface module generates a display main control module operating system power-off instruction (through the network port) to the display main control module. After the display main control module OS is shut down, the screen is turned off, and the display main control module OC_DC_IO_OSShutdown pin output is grounded to respond to the interface module that the display control OS has been powered off. After receiving the response, the IO generates an OC gate power-off state (OC_IO_MPM1_PowerCtrl and OC_IO_MPM2_PowerCtrl are suspended) to the power module, and the power module is powered off. At this time, all normal power is powered off, and only the normal power supply is normally powered on. This state is called the soft shutdown state.
[0092] In the soft shutdown state, if the front panel power-on control switch is pressed once, the internal circuit of the display main control module generates a pulse signal (5.2V when not pressed, pulled down to 0V when pressed, and restored to 5.2V after popping up), which is transmitted to the interface module through the OC_DC_IO_PowerCtrl pin. The interface module generates the OC gate power-on state (OC_IO_MPM1_PowerCtrl and OC_IO_MPM2_PowerCtrl pins are grounded) to the power module. The power module is powered on and all modules are powered normally. All modules start the operating system, bottom layer and application normally.
[0093] (4) Display screen preheating circuit design
[0094] In order to ensure that the display screen can work normally at low temperatures, the mainboard is reserved for the display screen heating function. The hardware uses the sensor NTC to sense the external temperature, thereby changing its own resistance. The resistance value changes in a linear curve, and the heating function is switched on and off through the comparator threshold comparison; the software uses the MCU to collect the sensor NTC resistance value, judge the external temperature by the resistance value, and heat the LCD screen by controlling the IO port switch heating MOS tube. The heating function automatically turns on the heating when the external temperature is lower than -30℃. The temperature can be set through hardware parameters. The heating power supply is 12V and the power consumption is less than 3W; the pure hardware heating principle is as follows Fig.10 As shown, software controls heating Fig.11 shown.
[0095] (5) USB functional circuit design
[0096] Complete the function of transferring the USB keyboard signal and USB mouse signal of the front panel to the front panel interface. TVS and ESD protection are added externally, and the power supply is designed with overcurrent protection function. The principle is as follows Fig.12 shown.
[0097] (6) LVDS communication circuit design
[0098] LVDS communication is realized through serial port conversion, model AT3010; AT3010 is a high-speed, low-power chip designed for backplane bus interface. The chip uses a single voltage of 3.3V or 5.0V for power supply, has a built-in differential transmitter and a receiver, and shorts the output of the transmitter and the receiving end of the receiver together to reduce the load on the bus. The chip provides 4 independent TTL level ports: RE, DE, DIN and ROUT to increase the flexibility of customer configuration.
[0099] The transmitter receives single-ended TTL level and outputs low voltage differential signal (LVDS). LVDS can transmit at high speed and low power consumption to reduce electromagnetic interference and suppress ±1V common mode noise. The chip's drive current reaches 10mA and can drive a backplane with a large load. The backplane impedance is ≥27 ohms.
[0100] The receiver receives LVDS level and outputs single-ended CMOS / TTL level with a transmission rate of ≥100Mbps. The receiving threshold is ±100mV and the common mode range is ±1V. Fig.13 shown.
[0101] (7) Network function circuit design
[0102] The LAN port function is designed to be 100M / 1000M adaptive, which is led out from the core of the core board, output to the carrier unit through the PHY interface chip, and then connected to the network transformer to output to the aviation plug-in board. The principle is as follows Fig.14 shown.
[0103] (8)Key circuit design
[0104] Complete the front panel key detection function, can detect the pressing and releasing of keys, and complete the signal debounce; quantity: 10+5+2; key value: 0-9, F1-F5, confirm, cancel. Fig.15 As shown, the chip model used for the key is: BC6301; the principle is to convert the physical key into a UART signal for the system to process, and through driver writing, the corresponding key is consistent with the corresponding key on the keyboard.
[0105] The radio control system provided in this implementation is designed with a self-test function. The self-test function is designed to automatically detect when the software is powered on. After power-on, the system will print self-test information through the debugging serial port, and read the working status of each unit function through the self-test information.
[0106] The radio control system provided in this implementation executes noise reduction code in Ubuntu and can provide two noise reduction methods with certain flexibility. It can collect audio signals on-site and reduce noise, or select audio signals transmitted from the host to the Linux system, and choose to process them according to its own needs. The file display screen of the noise reduction results will promptly feedback the current processing results.
[0107] The radio station control system provided in this embodiment has the following advantages:
[0108] (1) Confidentiality: The hardware circuit is designed with three confidentiality interfaces: key injection interface, parameter injection interface and key destruction interface, which increases the security and confidentiality of the system equipment in the military field.
[0109] (2) Noise reduction: Comprehensively consider the noise reduction effect of the algorithm, the control of music noise, the complexity of the algorithm, the real-time performance and the impact on the sound quality. The audio noise of the radio station is filtered out by optimizing the variational mode decomposition and wavelet threshold of the kingfisher.
[0110] (3) Low energy consumption design: The main energy consumption units are the CPU and memory of the motherboard. Select low-power components. The solid-state hard disk and memory used in this device are all low-power devices. DDR4 low-voltage memory is used for memory to reduce memory power consumption and achieve energy saving. Increase the resistance value of the collected ground resistance to reduce the current flowing to the ground and reduce power consumption.
[0111] In terms of software, the system sleep mode is used to wake up when working and keep in sleep mode when not working. At the same time, the display screen is turned off. For some unused IOs, the MCU can be configured as analog input state by default to achieve the purpose of reducing power consumption.
[0112] Embodiment 2
[0113] This embodiment provides a radio audio noise reduction method, such as Fig.16 As shown, the method includes:
[0114] S1: Collect the noisy audio signal through the audio acquisition module.
[0115] S2: Performing automatic gain control on the noisy audio signal.
[0116] S3: performing variational mode decomposition on the gain-adjusted noisy audio signal to obtain a plurality of IMF components.
[0117] S4: Based on the multiple IMF components, a wavelet denoising method is used to perform denoising processing to obtain a denoised audio signal.
[0118] Furthermore, step S2 specifically includes:
[0119] Automatic Gain Control (AGC) is a commonly used technology in signal processing, which is used to adjust the gain of the signal to keep the signal level within an appropriate range. AGC plays an important role in the fields of communication, audio processing, and image processing. Its main function is to automatically adjust the gain of the amplifier or receiver according to the strength of the input signal to ensure that the amplitude of the output signal is within a stable range. This can avoid information loss or distortion caused by too strong or too weak signals. In this embodiment, software is used to implement AGC control. The processing process is as follows: Fig.17 shown.
[0120] S21: Input signal strength detection: The strength information of the input signal is obtained by measuring the power or amplitude of the decoded noisy signal.
[0121] S22: Determine whether the signal strength exceeds the threshold: According to the preset threshold, determine whether the strength of the input signal exceeds the set range. When the signal noise reduction is lower than the threshold, the noisy signal is output for the next step of processing (noise reduction processing); when the signal strength noise reduction is higher than the threshold, the automatic gain adjustment processing is performed.
[0122] S23: Adjust gain: If the input signal strength is too strong, AGC will reduce the signal gain; if the input signal strength is too weak, AGC will increase the signal gain to keep the output signal amplitude within an appropriate range.
[0123] S24: Output signal processing: The signal after gain adjustment is amplified, filtered, and processed before output.
[0124] Furthermore, step S3 specifically includes:
[0125] S31: The Pied Kingfisher optimization algorithm is used to determine the modal decomposition number and penalty factor.
[0126] S32: performing variational modal decomposition on the gain-adjusted noisy audio signal based on the modal decomposition number and the penalty factor to obtain a plurality of IMF components.
[0127] Specifically, VMD decomposes a time domain original signal x(t) into K intrinsic mode functions u with frequency modulation and amplitude modulation characteristics and sparse characteristics. k (t), each u k (t) have different center frequencies of AM / FM signals, u k (t) can be expressed as:
[0128] u k (t) = A k (t)cos(θ k (t))(1)
[0129] Where: θ k (t) Kth instantaneous phase; A k (t) is the envelope amplitude. The core idea of the VMD algorithm is to construct the signal as a variational problem model, and then search for the optimal solution of the model through iterative updates. That is, the signal x can be decomposed into K center frequencies ω k and the IMF component with limited bandwidth. The corresponding variational constraint model is:
[0130]
[0131] In the above formula, {u k}={u1,...,u k} are the K IMF components obtained after decomposition, {ω k}={ω 1, ...,ω k} is the center frequency of each component.
[0132] In order to solve the above variational model, the constrained variational model is converted into an unconstrained variational problem. Therefore, the quadratic penalty factor α and the Lagrange multiplier λ(t) are introduced to obtain the augmented Lagrangian function:
[0133]
[0134] Where α is the quadratic penalty factor and λ(t) is the Lagrange multiplier.
[0135] The alternating direction multiplier algorithm is used to solve the above variational problem and the modal components are updated iteratively. Update center frequency and update the Lagrange multiplier Calculate (3) as the "saddle point", and then use the Parseval / Plancherel Fourier isometry transform to and Mapped to the frequency domain, the expressions for the modal component, center frequency and Lagrange multiplier update are shown in equations (4), (5) and (6), respectively:
[0136]
[0137] In the above formula, They are f(t), λ n (t) is the Fourier transform of
[0138] Given a convergence tolerance ε>0, perform iterative updates until the convergence condition is met. The iterative stop condition is:
[0139]
[0140] The convergence tolerance ε is generally 10 -6 .
[0141] VMD is affected by the modal decomposition number K and the penalty factor α. If the K value is too small, modal aliasing will occur; if the K value is too large, false components will be generated; the smaller α is, the larger the bandwidth of the IMF component after decomposition; the larger α is, the smaller the bandwidth of the IMF component after decomposition; therefore, before VMD, select the appropriate [K, α].
[0142] Pied Kingfisher Optimizer (PKO) is a swarm intelligence optimization algorithm inspired by nature. The algorithm is inspired by the unique hunting strategy and symbiotic relationship of Pied Kingfisher in nature, which mainly includes perching and hovering, diving strategy, and symbiotic stage.
[0143] 1) Population initialization:
[0144] Similar to many other population-based methods, PKO starts the search process by randomly generating a set of starting solutions from the search space as the first trial. The equation used to generate the initial population is as follows:
[0145]
[0146] In formula (8), F x,y is the position of the xth individual in the yth dimension, rand is a random value from 0 to 1, U B To search for the upper bound, L B To search for a lower bound.
[0147] Once the initial population is generated, the fitness function is used to evaluate the fitness value of each individual based on its ability to solve the problem. The individuals with the best fitness value are then selected to create the new generation.
[0148] 2) Perching and hovering strategies (exploration phase)
[0149] The exploration phase of PKO is inspired by the perching and hovering behavior of kingfishers. Observations of kingfishers in their natural habitats show that they alternate between perching and hovering positions for attack based on various factors. In PKO, the position of the search agent is determined based on the foraging activity of the kingfisher. The position of the kingfisher is updated according to the following equation:
[0150]
[0151] During the iteration process, the solution of the next iteration is calculated using Fx (n+1) indicates that the position of the current iteration is represented by F x (n). The parameter α is calculated as 2×randn(1,Dim)-1, where randn represents a random number from a normal distribution. N represents the population size, and Dim represents the dimension of the problem under consideration.
[0152] The parameter T plays a key role in this approach, and its value is determined dynamically based on the current strategy, which can be either "perched" or "hovering". The calculation of T is tailored for each strategy, ensuring the best performance in different operating modes.
[0153] When searching for prey in the water, Pied Kingfishers typically hold their crests at a vertical angle. This helps expand their field of vision, allowing them to detect movement from greater distances, helping them find fish even if they are hiding in vegetation or debris in the water. Once the Pied Kingfisher has spotted potential prey, it lowers its crest, uses binocular vision to focus on the prey, and makes precise movements to capture it. The perching strategy can be mathematically modeled using the previous formula, and the T parameter is calculated as follows:
[0154]
[0155] number=2*π*rand(11)
[0156] Maxiter is the maximum number of iterations, the constant value BF is set to 8, and randn represents a random number between 0 and 1. The mathematical model of the hovering strategy is shown below, and the calculation formula of the T parameter is:
[0157]
[0158] The fitness of the xth and yth kingfishers are respectively expressed as PKO fitness(x) and PKO fitness(y) In addition, a constant value, BF is the beating factor.
[0159] The Pied Kingfisher is known for its rapid flapping of its wings while hovering in the air. Its wings are relatively short and wide, allowing it to flap efficiently and quickly. By flapping its wings at a high frequency, this bird can maintain a stable position in the air. The Pied Kingfisher's wings are particularly well suited for hovering because their wings are long and pointed, which creates lift that allows them to maintain their position. The wings are relatively stiff, allowing the bird to flap its wings quickly and maintain a stable hover. The image below demonstrates the Pied Kingfisher's ability to hover and maintain a stable position while hunting.
[0160] 3) Diving strategy (development phase)
[0161] As mentioned earlier, the Pied Kingfisher is well known for its diving behavior, which it uses to catch fish. Typically, the bird perches from a high point overlooking the water, such as a branch or rock, to observe the presence of fish. Once a fish is spotted, the Pied Kingfisher quickly dives into the water from a height of several meters and captures its prey with its beak, with quick and precise movements. The bird can reach depths of several meters before returning to its perch with the fish firmly in its beak.
[0162] The Pied Kingfisher's diving behavior makes it an efficient hunter, which is one of the reasons why the species is so successful in its habitat. The bird's ability to dive quickly and accurately, combined with its sharp beak and excellent eyesight, make it an effective predator in aquatic environments.
[0163] F x (n+1)=F x (n)+HA*o*α*(bF best (n)),(x=1,2,…,N) (14)
[0164] Among them, the fitness value of the x-th Pied Kingfisher is recorded as PKO fitness(x) , the best fitness represents the best fitness value obtained from all iterations. α is the control parameter, HA and o represent the hunting ability, which can be calculated by the following formula:
[0165]
[0166] b=F x (n)+o 2 *rand*F best (n)(17)
[0167] 4) The accompanying stage (local escape stage)
[0168] The hunting efficiency of the Pied Kingfisher can be affected by a variety of factors, such as the availability of prey, the hunting skills of individual birds, and the presence of competitors or predators in the area. This behavior is mathematically represented as follows:
[0169]
[0170] α=2*randn(1,Dim)-1 (19)
[0171]
[0172] Where PE is the predation efficiency of the Pied Kingfisher, PE max Maximum predation efficiency, PE min Minimum predation efficiency.
[0173] 5) Convergence and termination conditions: The algorithm gradually optimizes the nest solution through multiple iterations until the set number of iterations is reached or a certain convergence condition is met. Usually, the termination condition of the algorithm can be reaching the maximum number of iterations or the objective function value changes very little in several consecutive generations. The processing flow of PKO is as follows Fig.18 shown.
[0174] Furthermore, step S4 specifically includes:
[0175] S41: Calculate the kurtosis value of each IMF component.
[0176] S42: Using a wavelet denoising method to perform denoising on the IMF components whose kurtosis values are less than a preset value.
[0177] S43: reconstructing the IMF components after the noise reduction processing and the IMF components whose kurtosis values are greater than or equal to the preset value to obtain the audio signal after the noise reduction processing.
[0178] Specifically, the wavelet threshold is an estimate of the critical point of the noise and signal wavelet coefficients, which will deviate from the real critical value. The wavelet coefficients after threshold processing will still contain a small part of the noise wavelet coefficients, so a reasonable rule is needed to correct the processed wavelet coefficients to a certain extent. The wavelet threshold is divided into hard threshold and soft threshold.
[0179] Hard thresholding is a simple and direct threshold processing method. Its basic idea is to set the wavelet coefficients less than the threshold to zero, while keeping the wavelet coefficients greater than the threshold unchanged. The expression of the hard threshold function is:
[0180]
[0181] In the formula, w j,k is the original wavelet coefficient, To estimate the wavelet coefficients, T is the threshold. Hard thresholding can effectively remove noise wavelet coefficients that are less than the threshold, but it may cause mutations in the signal (because the wavelet coefficients of some valid signals are directly set to zero), so it is suitable for situations where the signal contains sparse noise. However, the hard threshold is discontinuous at the threshold point, resulting in undesirable oscillations in the reconstructed signal and loss of smoothness.
[0182] The soft threshold is smoother than the hard threshold. It softens the wavelet coefficients, shrinks the coefficients less than the threshold to zero, and shrinks the coefficients greater than the threshold. The expression of the soft threshold function is:
[0183]
[0184] Where sgn(·) is the sign function of the wavelet coefficient. Soft thresholding retains certain wavelet coefficients, so it can better maintain the smoothness of the signal and is suitable for situations where there is more noise and less mutation in the signal. The soft threshold is continuous at the threshold point, but the derivative is discontinuous, and the reconstructed signal has a permanent deviation, which affects the accuracy to a certain extent.
[0185] In view of the problems existing in traditional soft and hard thresholds, the improved threshold denoising function proposed by scholars is adopted, which combines the characteristics of soft and hard thresholds and solves the problem of discontinuity of soft and hard thresholds at the threshold point. The expression of the improved threshold function is:
[0186]
[0187] The above formula is ρ∈(0,T).
[0188] Fig.19 is the comparison curve between the soft and hard threshold function and the improved threshold function. As can be seen from the figure, the improved threshold function can overcome the discontinuity of the hard threshold function at ±T and the shortcomings of the soft threshold function with constant error. j,k | = T, the improved threshold function does not have abruptness like the hard threshold function, so no oscillation occurs when reconstructing the enhanced audio signal; at w j,k In the ≥T stage, the improved threshold function is closer to the hard threshold, avoiding With w j,k There is a constant difference between them, so the mutation information of useful audio signals will not be lost too much; the improved threshold function conforms to the law that the modulus value of wavelet transform decays exponentially, and its denoising effect is better.
[0189] Using envelope entropy as the fitness function, the Pied Kingfisher optimization algorithm is used to adaptively optimize the VMD parameters to determine the optimal VMD mode number k and quadratic penalty factor α; the Lagrange multiplier and quadratic penalty factor α are introduced to transform the original variational problem into an unconstrained variational problem. Through the alternating direction method, the optimal solution to the original problem is transformed into the saddle point of the Lagrange period equation, and finally n IMFs components are obtained; the kurtosis value of IMFs is calculated, and when the kurtosis value of IMFs is less than 3, improved threshold denoising is performed, because the smaller the IMFs kurtosis value, the more noise there is, and the larger the IMFs kurtosis value, the more noise there is; finally, the processed IMF components are reconstructed with the IMF components of the signal to obtain the denoised signal. The process based on variational mode decomposition combined with wavelet threshold denoising is as follows Fig.16 shown.
[0190] The signal-to-noise ratio (SNR), root mean square error (RMSE), and waveform similarity are used as evaluation indicators for signal noise reduction. The expression of the signal-to-noise ratio (SNR) is as follows:
[0191]
[0192] Where L is the signal length, X(t) is the original signal; It is the signal after noise reduction and reconstruction. The SNR is used to measure the relationship between the signals before and after filtering. The higher the SNR value, the better the filtering effect.
[0193] The root mean square error (RMSE) expression is as follows:
[0194]
[0195] In the formula, the closer the RMSE value is to 0, the better the decomposition effect is.
[0196] The expression of the waveform similarity coefficient (NCC) is as follows:
[0197]
[0198] The larger the NCC value, the better the noise reduction effect.
[0199] A noisy speech signal with a signal length of 48000 and a frequency of 48000 Hz was used. The VMD parameters were optimized based on the PKO optimization algorithm by using block division. PKO optimization VMD optimization is as follows Fig. 20 As shown in Figure 2, the IMFs after VMD decomposition are as follows: Fig.21 As shown in the figure. The IMF spectrum distribution is as follows Fig. 22 shown.
[0200] The traditional hard threshold, soft threshold and improved threshold function are used to perform noise reduction analysis on the noisy signal. DB3 is selected as the wavelet basis function for processing the noisy signal, and the number of decomposition layers is 4. The simulation diagram is shown in the figure below. Figure 23-Figure 25 shown.
[0201] According to the comparison of the above three denoising algorithms, the hard and soft threshold functions can achieve partial denoising, but the denoising effect is still not good for local details, with SNR, RMSE, and NCC of 9.89dB, 0.0283, and 0.9684, respectively. The soft threshold SNR, RMSE, and NCC are 9.93dB, 0.0282, and 0.9687, respectively. The PKO and improved wavelet threshold denoising methods proposed in this application can filter out noise while retaining the real components in the signal. The optimal modal component and quadratic penalty factor are K_opt: 6, alpha_opt: 20882, respectively. The SNR, RMSE, and NCC after filtering by this method are 14.6437dB, 0.0192, and 0.984, respectively.
[0202] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0203] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A radio station control system, characterized in that: include: An audio acquisition module, used to acquire audio signals containing noise; A core processor connected to the audio acquisition module, used for performing variational mode decomposition on the noisy audio signal and performing noise reduction using a wavelet noise reduction method; A display main control module, connected to the core processor, for performing signal conversion on the noisy audio signal and the noise-reduced audio signal, and converting the HDMI signal into an RGB signal; The core processor includes: An automatic gain control unit, used for adjusting the gain of the noisy audio signal and the noise-reduced audio signal; The noise reduction unit is connected to the automatic gain control unit and is used to perform variational mode decomposition on the noisy audio signal after gain adjustment and to perform noise reduction by using a wavelet noise reduction method.
2. The radio station control system according to claim 1, characterized in that: The core processor also includes: An audio power amplifier unit, connected to the audio acquisition module, for amplifying the audio signal containing noise; An audio decoding unit, connected to the audio power amplifier unit, for decoding the amplified audio signal containing noise; An A / D conversion unit, connected to the audio decoding unit, for performing analog-to-digital conversion on the decoded audio signal containing noise; A D / A conversion unit, connected to the automatic gain control unit, for performing digital-to-analog conversion on the gain-adjusted and noise-reduced audio signal; An audio encoding unit, connected to the D / A conversion unit, for encoding the audio signal after the digital-to-analog conversion and the noise reduction; The audio power amplifier unit is also connected to the audio encoding unit and is used to amplify the encoded and noise-reduced audio signal.
3. The radio station control system according to claim 1, characterized in that: The display main control module is also used to preheat the LED screen, and the LED screen is used to display the noisy audio signal and the noise-reduced audio signal.
4. The radio station control system according to claim 1, characterized in that: The station control system also includes: The power management module is used to convert the input voltage and provide power to the radio station control system.
5. The radio station control system according to claim 1, characterized in that: Also includes: Confidentiality interface, used to connect with other confidentiality equipment; The confidentiality interface includes a key injection interface, a parameter injection interface and a key destruction interface.
6. The station control system according to claim 1, characterized in that: Also includes: The audio output module is used to output the audio after noise reduction.
7. The station control system according to claim 6, characterized in that: The audio acquisition module is a headset and a microphone, and the audio output module is a headset, a microphone and a speaker.
8. A radio audio noise reduction method, characterized in that: The radio audio noise reduction method is applied to the radio control system described in claims 1-7, and the radio audio noise reduction method includes: The audio signal containing noise is collected through the audio collection module; Performing automatic gain control on the noisy audio signal; Perform variational mode decomposition on the gain-adjusted noisy audio signal to obtain multiple IMF components; Based on multiple IMF components, the wavelet denoising method is used to perform denoising processing to obtain the denoised audio signal.
9. The radio audio noise reduction method according to claim 8, characterized in that: The gain-adjusted noisy audio signal is subjected to variational mode decomposition to obtain multiple IMF components, including: The Pied Kingfisher optimization algorithm is used to determine the modal decomposition number and penalty factor; Based on the modal decomposition number and the penalty factor, variational modal decomposition is performed on the gain-adjusted noisy audio signal to obtain a plurality of IMF components.
10. The radio audio noise reduction method according to claim 8, characterized in that: Based on multiple IMF components, wavelet denoising method is used to perform denoising processing to obtain the denoised audio signal, which specifically includes: Calculate the kurtosis value of each IMF component; The wavelet denoising method is used to denoise the IMF components whose kurtosis value is less than the preset value; The IMF components after noise reduction processing and the IMF components with kurtosis values greater than or equal to a preset value are reconstructed to obtain a noise-reduced audio signal.
