Satellite portable station capable of intelligent voice control and broadcasting

Abstract

The invention discloses a satellite portable station capable of intelligent voice control and broadcast, and relates to the technical field of voice control, and the satellite portable station comprises a voice controller module which is used for carrying out the voice pickup and voice conversion of an operator, and is in functional butt joint with an antenna control unit ACU; the satellite portable station main body module is used for converting the acquired voice of the operator into a corresponding control instruction; the voice broadcast loudspeaker module broadcasts the current working state of the satellite portable station; the antenna control module is used for receiving the control instruction provided by the voice controller and realizing antenna control through a three-axis transmission mechanism; and the three-axis transmission module executes orientation, pitching and polarization adjustment according to the control instruction. According to the satellite portable station, the functions of intelligent voice control, efficient noise processing, real-time state broadcasting, multi-language supporting and the like are integrated, so that the operation convenience, the intelligent level and the adaptability of the satellite portable station are remarkably improved, and more efficient, reliable and friendly use experience is provided for a user.

Description

Technical Field

[0001] The invention relates to the technical field of voice control, and in particular to a satellite portable station capable of intelligent voice control and broadcasting. Background Art

[0002] Traditional portable satellite stations mainly rely on manual control in their design and operation. This operation mode increases the complexity of use to a certain extent and may lead to operational errors and inefficiency. Especially in scenarios such as field operations, emergency communications or military operations, operators often need to quickly complete antenna adjustment and control in complex and harsh environments. However, manual operation is not only time-consuming and labor-intensive, but also easily limited by the operator's experience and technical level, thus affecting the timeliness and stability of communications.

[0003] In addition, the existing voice control system still has great limitations in complex noisy environments. Although voice recognition technology has made some progress in recent years, in practical applications, especially in outdoor or noisy environments, the accuracy and response speed of voice recognition still cannot meet the needs of real-time control. At the same time, traditional satellite portable stations usually lack intelligent broadcasting functions. Operators need to use display screens, indicator lights or other visual feedback devices to understand the working status of the equipment. This method is not intuitive and convenient in some scenarios, especially in low light or when the operator cannot directly observe the equipment. For example, at night or in an emergency, the operator may not be able to quickly obtain the status information of the equipment, which affects the efficiency and safety of the operation. Summary of the invention

[0004] In order to solve the above technical problems, a satellite portable station capable of intelligent voice control and broadcasting is provided. This technical solution solves the above problems.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: A portable satellite station capable of intelligent voice control and broadcasting, comprising: Voice controller module: used for operator voice pickup and voice conversion, and docking with the antenna control unit ACU function; Satellite portable station main body module: The satellite portable station main body module is electrically connected to the voice controller module, and the satellite portable station main body module is used to convert the acquired operator voice into corresponding control instructions, and push the control instructions to the azimuth, pitch, and polarization three-axis transmission mechanism through the antenna ACU; Voice broadcast speaker module: The voice broadcast speaker module is electrically connected to the satellite portable station main module. The voice broadcast speaker module is used to broadcast the current working status of the satellite portable station; Antenna control module: The antenna control module is typically connected to the voice broadcast speaker module, and is used to receive control instructions provided by the voice controller and realize antenna control through a three-axis transmission mechanism; Three-axis transmission module: The three-axis transmission module is electrically connected to the antenna control module, and the three-axis transmission module is used to perform azimuth, pitch and polarization adjustments according to control instructions.

[0006] Preferably, the voice controller module specifically includes: Voice pickup unit: used for 360-degree omnidirectional sound pickup; Speech processing unit: used for speech recognition, voiceprint recognition, speech enhancement, speech detection, and intelligent noise reduction; Model training unit: used to train and write command models for starting satellite search, satellite search status report, network access status report, antenna AGC value report, antenna retraction and antenna shutdown; The control instruction conversion unit is used to convert the operator's voice into corresponding control instructions.

[0007] Preferably, the speech processing unit specifically includes: Convert the input voice signal into corresponding text and commands, obtain the voice signal through the microphone, pre-process the voice signal, remove background noise, equalize the signal, and remove reverberation; The speech signal is converted from the time domain to the frequency domain based on the short-time Fourier transform algorithm, and the Mel-frequency cepstral coefficients are used to extract features related to human auditory perception; Detect valid speech segments in speech signals and remove silent or non-speech parts; Use a convolutional neural network model to model features, map features to corresponding phonemes or words, and optimize the output of the acoustic model in combination with a language model to generate the final text result; Extract features related to the speaker's identity, including formant frequency, vocal tract characteristics, and phonation style, use a deep learning model to build the speaker's voiceprint template, and train the model to distinguish the characteristics of different speakers; Compare the features of the speech to be recognized with the pre-stored voiceprint template, calculate the similarity score, and determine whether they are the same speaker; By analyzing the silent segment to estimate the noise characteristics, signal processing is performed, including frequency domain processing and time domain processing; a neural network model is used to directly recover clean speech from the noisy signal; Use the machine learning model to classify each signal segment, determine whether the signal is a speech segment, and output the detection result; Detect the type and intensity of environmental noise in real time, use adaptive filters or deep learning models to dynamically track noise characteristics, use neural networks to directly separate clean speech from noise signals, smooth the noise-reduced signals, and adjust noise reduction parameters according to actual scenarios.

[0008] Preferably, the method of converting the speech signal from the time domain to the frequency domain based on the short-time Fourier transform algorithm and extracting features related to human auditory perception using Mel-frequency cepstral coefficients specifically includes: Among them, the short-time Fourier transform algorithm formula is: ; In the formula, Indicated in Frames and The complex spectrum value on the frequency components, is the frequency index, is the time index, represents the original signal, is the sample index in the current frame, is the windowing function, is a complex exponential term.

[0009] Preferably, the estimating noise characteristics by analyzing the silent segment and performing signal processing, including frequency domain processing and time domain processing; and using a neural network model to directly recover clean speech from the noise signal specifically include: In speech signals, silence segments refer to the parts without speech content. Silence segments are identified by energy detection, zero-crossing rate detection or machine learning-based methods. By analyzing the spectrum characteristics of the silent segment, the power spectrum density of the noise is estimated and the spectrum characteristics of the noise are obtained; Collect clean speech signals and corresponding noise signals, and simulate real-world scenarios by adding different types of noise; Frequency domain or time domain features are extracted based on convolutional neural networks, the model optimization goal is defined based on the loss function, and the model parameters are updated using the Adam optimization algorithm.

[0010] Preferably, the real-time detection of the type and intensity of environmental noise, the use of an adaptive filter or a deep learning model to dynamically track noise characteristics, the use of a neural network to directly separate clean speech from the noise signal, the smoothing of the noise-reduced signal, and the adjustment of noise reduction parameters according to the actual scenario specifically include: Detect distortion and artifacts introduced during the noise reduction process, use soft thresholding to further optimize signal quality, and dynamically adjust the degree of smoothing based on signal characteristics; The accelerometer senses the current environment and uses the device's location and time context to determine the current scene. Dynamically adjust filter parameters based on noise type.

[0011] Preferably, the voice broadcast speaker module specifically includes: A working status broadcasting unit is used to broadcast the current working status of the satellite portable station; A volume adjustment unit, used to automatically adjust the broadcast volume according to the ambient noise; The multi-language support unit is used to support language broadcasting in Chinese, English and Japanese.

[0012] Preferably, the satellite portable station uses the ASRPRO chip as the core, supports offline voice recognition, is independent of the network, has a delay of less than 0.1 seconds, a recognition rate of more than 98%, and supports ultra-long-distance recognition of 10 meters; The ASRPRO chip uses DNN, TDNN and RNN neural networks and convolution operations to perform speech recognition tasks. The calculation formula is: ; In the formula, is the input feature vector, is the weight matrix, is the bias vector, is the activation function, is the hidden layer state, is the current input feature, and are the weight matrices from input to hidden layer and from hidden layer to hidden layer, is the activation function.

[0013] Preferably, the antenna control module specifically includes: An instruction receiving unit, used for receiving a control instruction sent by a voice controller; An instruction parsing unit, used for parsing control instructions and generating corresponding execution signals; The transmission control unit is used to push the execution signal to the azimuth, pitch and polarization three-axis transmission mechanism to realize the conversion of control instructions into execution actions.

[0014] Preferably, the only manual operation that the operator of the satellite portable station needs to perform is the one-button power-on, and the remaining functions can be achieved through voice commands. The satellite portable station will trigger a voice broadcast function when starting and ending an action, prompting the current working status of the portable station, including but not limited to the start of satellite search, successful satellite search, successful network access, normal antenna AGC value, completion of antenna retraction and extension, and completion of antenna shutdown status information.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes to significantly improve the operational convenience, intelligence level and adaptability of the satellite portable station by integrating functions such as intelligent voice control, efficient noise processing, real-time status broadcast and multi-language support, providing users with a more efficient, reliable and friendly user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0017] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0018] Reference Figure 1 As shown, a satellite portable station capable of intelligent voice control and broadcasting includes: a voice controller module: used for operator voice pickup, voice conversion, and functional docking with the antenna control unit ACU; Satellite portable station main body module: The satellite portable station main body module is electrically connected to the voice controller module, and the satellite portable station main body module is used to convert the acquired operator voice into corresponding control instructions, and push the control instructions to the azimuth, pitch, and polarization three-axis transmission mechanism through the antenna ACU; Voice broadcast speaker module: The voice broadcast speaker module is electrically connected to the satellite portable station main module. The voice broadcast speaker module is used to broadcast the current working status of the satellite portable station; Antenna control module: The antenna control module is typically connected to the voice broadcast speaker module, and is used to receive control instructions provided by the voice controller and realize antenna control through a three-axis transmission mechanism; Three-axis transmission module: The three-axis transmission module is electrically connected to the antenna control module, and the three-axis transmission module is used to perform azimuth, pitch and polarization adjustments according to control instructions.

[0019] The voice controller module specifically includes: Voice pickup unit: used for 360-degree omnidirectional sound pickup; Speech processing unit: used for speech recognition, voiceprint recognition, speech enhancement, speech detection, and intelligent noise reduction; Model training unit: used to train and write command models for starting satellite search, satellite search status report, network access status report, antenna AGC value report, antenna retraction and antenna shutdown; A control instruction conversion unit, used to convert the operator's voice into corresponding control instructions; By integrating advanced voice recognition and processing technologies, precise voice control is achieved, significantly improving operational efficiency and accuracy.

[0020] The speech processing unit specifically includes: Convert the input voice signal into corresponding text and commands, obtain the voice signal through the microphone, pre-process the voice signal, remove background noise, equalize the signal, and remove reverberation; The speech signal is converted from the time domain to the frequency domain based on the short-time Fourier transform algorithm, and the Mel-frequency cepstral coefficients are used to extract features related to human auditory perception; Detect valid speech segments in speech signals and remove silent or non-speech parts; Use a convolutional neural network model to model features, map features to corresponding phonemes or words, and optimize the output of the acoustic model in combination with a language model to generate the final text result; Extract features related to the speaker's identity, including formant frequency, vocal tract characteristics, and phonation style, use a deep learning model to build the speaker's voiceprint template, and train the model to distinguish the characteristics of different speakers; Compare the features of the speech to be recognized with the pre-stored voiceprint template, calculate the similarity score, and determine whether they are the same speaker; By analyzing the silent segment to estimate the noise characteristics, signal processing is performed, including frequency domain processing and time domain processing; a neural network model is used to directly recover clean speech from the noisy signal; Use the machine learning model to classify each signal segment, determine whether the signal is a speech segment, and output the detection result; Detect the type and intensity of ambient noise in real time, use adaptive filters or deep learning models to dynamically track noise characteristics, use neural networks to directly separate clean speech from noise signals, smooth the noise-reduced signals, and adjust noise reduction parameters according to actual scenarios; It uses a variety of signal processing technologies and deep learning models to effectively suppress background noise, improve voice quality, and ensure stable operation even in complex environments.

[0021] The speech signal is converted from the time domain to the frequency domain based on the short-time Fourier transform algorithm, and the Mel-frequency cepstral coefficients are used to extract the features related to human auditory perception, including: Among them, the short-time Fourier transform algorithm formula is: ; In the formula, Indicated in Frames and The complex spectrum value on the frequency components, is the frequency index, is the time index, represents the original signal, is the sample index in the current frame, is the windowing function, is a complex exponential term.

[0022] By analyzing the silent segment to estimate the noise characteristics, signal processing is performed, including frequency domain processing and time domain processing; using the neural network model to directly recover clean speech from the noisy signal, specifically including: In speech signals, silence segments refer to the parts without speech content. Silence segments are identified by energy detection, zero-crossing rate detection or machine learning-based methods. By analyzing the spectrum characteristics of the silent segment, the power spectrum density of the noise is estimated and the spectrum characteristics of the noise are obtained; Collect clean speech signals and corresponding noise signals, and simulate real-world scenarios by adding different types of noise; Frequency domain or time domain features are extracted based on convolutional neural networks, the model optimization goal is defined based on the loss function, and the model parameters are updated using the Adam optimization algorithm.

[0023] Detect the type and intensity of environmental noise in real time, use adaptive filters or deep learning models to dynamically track noise characteristics, use neural networks to directly separate clean speech from noise signals, smooth the noise-reduced signals, and adjust noise reduction parameters according to actual scenarios. Specifically, it includes: Detect distortion and artifacts introduced during the noise reduction process, use soft thresholding to further optimize signal quality, and dynamically adjust the degree of smoothing based on signal characteristics; The accelerometer senses the current environment and uses the device's location and time context to determine the current scene. Dynamically adjust filter parameters based on noise type.

[0024] The voice broadcast speaker module specifically includes: A working status broadcasting unit is used to broadcast the current working status of the satellite portable station; A volume adjustment unit, used to automatically adjust the broadcast volume according to the ambient noise; Multi-language support unit, used to support language broadcasting in Chinese, English and Japanese; Through the voice broadcast speaker module, the operator can be informed of the working status of the equipment in real time, improving the transparency and safety of operations.

[0025] The satellite portable station uses the ASRPRO chip as the core, supports offline voice recognition, is independent of the network, has a delay of less than 0.1 seconds, a recognition rate of more than 98%, and supports ultra-long-distance recognition of 10 meters; The ASRPRO chip uses DNN, TDNN and RNN neural networks and convolution operations to perform speech recognition tasks. The calculation formula is: ; In the formula, is the input feature vector, is the weight matrix, is the bias vector, is the activation function, is the hidden layer state, is the current input feature, and are the weight matrices from input to hidden layer and from hidden layer to hidden layer, is the activation function.

[0026] The antenna control module specifically includes: An instruction receiving unit, used for receiving a control instruction sent by a voice controller; An instruction parsing unit, used for parsing control instructions and generating corresponding execution signals; The transmission control unit is used to push the execution signal to the azimuth, pitch and polarization three-axis transmission mechanism to realize the conversion of control instructions into execution actions.

[0027] The only manual operation that the operator of the satellite portable station needs to perform is the one-button power-on. The rest of the functions can be achieved through voice commands. The satellite portable station will trigger the voice broadcast function when starting and ending an action, prompting the current working status of the portable station, including but not limited to the start of satellite search, successful satellite search, successful network access, normal antenna AGC value, completion of antenna retraction and extension, and completion of antenna shutdown status information.

[0028] In summary, the advantages of the present invention are: By integrating advanced speech recognition and processing technologies, precise voice control is achieved, significantly improving operational efficiency and accuracy; Adopt a variety of signal processing technologies and deep learning models to effectively suppress background noise, improve voice quality, and ensure stable operation even in complex environments; Through the voice broadcast speaker module, the operator is informed of the working status of the equipment in real time, improving the transparency and safety of operation; Supports voice control and broadcasting in multiple languages, expanding the application scope of the device and improving the user experience; Adopt high-performance chips and optimized algorithms to achieve low-latency, high-recognition rate voice processing, ensuring the stability and reliability of the system at critical moments; In addition to one-button power-on, all functions can be realized through voice commands, which greatly reduces the complexity of operation and improves the user experience; Dynamically tracking noise characteristics and adaptive filtering technology enable the system to adapt to a variety of working environments and noise conditions, improving the versatility and adaptability of the equipment; Through functions such as automatic volume adjustment and multi-language support, we ensure a good user experience of the device in different environments and user groups.

[0029] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A portable satellite station capable of intelligent voice control and broadcasting, characterized in that: include: Voice controller module: used for operator voice pickup and voice conversion, and docking with the antenna control unit ACU function; Satellite portable station main body module: The satellite portable station main body module is electrically connected to the voice controller module, and the satellite portable station main body module is used to convert the acquired operator voice into corresponding control instructions, and push the control instructions to the azimuth, pitch, and polarization three-axis transmission mechanism through the antenna ACU; Voice broadcast speaker module: The voice broadcast speaker module is electrically connected to the satellite portable station main module. The voice broadcast speaker module is used to broadcast the current working status of the satellite portable station; Antenna control module: The antenna control module is typically connected to the voice broadcast speaker module, and is used to receive control instructions provided by the voice controller and realize antenna control through a three-axis transmission mechanism; Three-axis transmission module: The three-axis transmission module is electrically connected to the antenna control module, and the three-axis transmission module is used to perform azimuth, pitch and polarization adjustments according to control instructions.

2. A portable satellite station capable of intelligent voice control and broadcasting according to claim 1, characterized in that: The voice controller module specifically includes: Voice pickup unit: used for 360-degree omnidirectional sound pickup; Speech processing unit: used for speech recognition, voiceprint recognition, speech enhancement, speech detection, and intelligent noise reduction; Model training unit: used to train and write command models for starting satellite search, satellite search status report, network access status report, antenna AGC value report, antenna retraction and antenna shutdown; The control instruction conversion unit is used to convert the operator's voice into corresponding control instructions.

3. A portable satellite station capable of intelligent voice control and broadcasting according to claim 2, characterized in that: The speech processing unit specifically includes: Convert the input voice signal into corresponding text and commands, obtain the voice signal through the microphone, pre-process the voice signal, remove background noise, equalize the signal, and remove reverberation; The speech signal is converted from the time domain to the frequency domain based on the short-time Fourier transform algorithm, and the Mel-frequency cepstral coefficients are used to extract features related to human auditory perception; Detect valid speech segments in speech signals and remove silent or non-speech parts; Use a convolutional neural network model to model features, map features to corresponding phonemes or words, and optimize the output of the acoustic model in combination with a language model to generate the final text result; Extract features related to the speaker's identity, including formant frequency, vocal tract characteristics, and phonation style, use a deep learning model to build the speaker's voiceprint template, and train the model to distinguish the characteristics of different speakers; Compare the features of the speech to be recognized with the pre-stored voiceprint template, calculate the similarity score, and determine whether they are the same speaker; By analyzing the silent segment to estimate the noise characteristics, signal processing is performed, including frequency domain processing and time domain processing; a neural network model is used to directly recover clean speech from the noisy signal; Use the machine learning model to classify each signal segment, determine whether the signal is a speech segment, and output the detection result; Detect the type and intensity of environmental noise in real time, use adaptive filters or deep learning models to dynamically track noise characteristics, use neural networks to directly separate clean speech from noise signals, smooth the noise-reduced signals, and adjust noise reduction parameters according to actual scenarios.

4. A portable satellite station capable of intelligent voice control and broadcasting according to claim 3, characterized in that: The method of converting the speech signal from the time domain to the frequency domain based on the short-time Fourier transform algorithm and extracting the features related to human auditory perception using the Mel-frequency cepstral coefficients specifically includes: Among them, the short-time Fourier transform algorithm formula is: ； In the formula, Indicated in Frames and The complex spectrum value at the frequency components, is the frequency index, is the time index, represents the original signal, is the sample index in the current frame, is the windowing function, is a complex exponential term.

5. A portable satellite station capable of intelligent voice control and broadcasting according to claim 4, characterized in that: The noise characteristics are estimated by analyzing the silent segment and performing signal processing, including frequency domain processing and time domain processing; Using a neural network model to directly recover clean speech from noisy signals specifically includes: In speech signals, silence segments refer to the parts without speech content. Silence segments are identified by energy detection, zero-crossing rate detection or machine learning-based methods. By analyzing the spectrum characteristics of the silent segment, the power spectrum density of the noise is estimated and the spectrum characteristics of the noise are obtained; Collect clean speech signals and corresponding noise signals, and simulate real-life scenarios by adding different types of noise; Frequency domain or time domain features are extracted based on convolutional neural networks, the model optimization goal is defined based on the loss function, and the model parameters are updated using the Adam optimization algorithm.

6. A portable satellite station capable of intelligent voice control and broadcasting according to claim 5, characterized in that: The real-time detection of the type and intensity of environmental noise, the use of an adaptive filter or a deep learning model to dynamically track noise characteristics, the use of a neural network to directly separate clean speech from noise signals, the smoothing of the noise-reduced signals, and the adjustment of noise reduction parameters according to actual scenarios specifically include: Detect distortion and artifacts introduced during the noise reduction process, use soft thresholding to further optimize signal quality, and dynamically adjust the degree of smoothing based on signal characteristics; The accelerometer senses the current environment and uses the device's location and time context to determine the current scene. Dynamically adjust filter parameters based on noise type.

7. A portable satellite station capable of intelligent voice control and broadcasting according to claim 6, characterized in that: The voice broadcast speaker module specifically includes: A working status broadcasting unit is used to broadcast the current working status of the satellite portable station; A volume adjustment unit, used to automatically adjust the broadcast volume according to the ambient noise; The multi-language support unit is used to support language broadcasting in Chinese, English and Japanese.

8. A portable satellite station capable of intelligent voice control and broadcasting according to claim 7, characterized in that: The satellite portable station uses the ASRPRO chip as the core, supports offline voice recognition, is independent of the network, has a delay of less than 0.1 seconds, a recognition rate of more than 98%, and supports ultra-long-distance recognition of 10 meters; The ASRPRO chip uses DNN, TDNN and RNN neural networks and convolution operations to perform speech recognition tasks. The calculation formula is: ； In the formula, is the input feature vector, is the weight matrix, is the bias vector, is the activation function, is the hidden layer state, is the current input feature, and are the weight matrices from input to hidden layer and from hidden layer to hidden layer, is the activation function.

9. A portable satellite station capable of intelligent voice control and broadcasting according to claim 8, characterized in that: The antenna control module specifically includes: An instruction receiving unit, used for receiving a control instruction sent by a voice controller; An instruction parsing unit, used for parsing control instructions and generating corresponding execution signals; The transmission control unit is used to push the execution signal to the azimuth, pitch and polarization three-axis transmission mechanism to realize the conversion of control instructions into execution actions.

10. A portable satellite station capable of intelligent voice control and broadcasting according to claim 9, characterized in that: The only manual operation that the operator of the satellite portable station needs to perform is the one-button power-on. The rest of the functions can be achieved through voice commands. The satellite portable station will trigger the voice broadcast function when starting and ending an action, prompting the current working status of the portable station, including but not limited to the start of satellite search, successful satellite search, successful network access, normal antenna AGC value, completion of antenna retraction and extension, and completion of antenna shutdown status information.

Images