CN111110469A - A multifunctional nursing bed - Google Patents

Abstract

The invention provides a multifunctional nursing bed, comprising: the lifting device comprises a base support, a front support plate, a rear support plate and a lifting device; the back support plate is fixed on the base support, the lifting beam of the front support plate is connected with the lifting device, and the front support plate can rotate around a hinge point of the front support plate and the back support plate along with the pushing of the lifting device to realize the back lifting of the front support plate; the side of the base support is provided with a connecting device, the tail end of the connecting device is provided with an interactive terminal, and the interactive terminal is provided with a voice acquisition device which is used for acquiring voice signals of a user; the base support is internally provided with a main control terminal, the main control terminal is respectively connected with the interactive terminal and the lifting device and used for sending the received user voice signals to the cloud computing platform, the cloud computing platform identifies the user voice signals and returns corresponding control instructions to the main control terminal, and the main control terminal drives the lifting device to execute corresponding operations according to the received control instructions. The invention is convenient to operate and improves the user experience.

Description

A multifunctional nursing bed

technical field

The invention relates to the technical field of medical nursing equipment, in particular to a multifunctional nursing bed.

Background technique

Nursing beds are widely used, and a large number of nursing beds have been set up in hospitals and nursing homes. Currently, nursing beds are developing towards improving the experience of users (such as the elderly and rehabilitation patients).

However, at present, users still have some inconvenient operations when using the nursing bed. For example, the nursing bed usually needs to operate under the bed when adjusting the back-up mechanism, and the user needs to get out of bed and stir the operating lever to adjust the angle of the back-up mechanism. , it is very inconvenient to use; there are also some nursing beds in the prior art that adjust the carrying mechanism through a remote control, but for some users with inconvenient hands and feet, there is still the defect of inconvenience.

SUMMARY OF THE INVENTION

In view of the above problems, the present invention aims to provide a multifunctional nursing bed.

The object of the present invention adopts the following technical solutions to realize:

A multifunctional nursing bed, comprising: a base bracket, a front support plate, a rear support plate and a lifting device; wherein,

The rear support plate is fixed on the base bracket, and the lifting beam of the front support plate is connected with the landing gear. rise;

A connecting device is arranged on the side of the base bracket, an interactive terminal is arranged at the end of the connecting device, and a voice collecting device is arranged on the interactive terminal, and the voice collecting device is used to collect user voice signals;

A main control terminal is arranged in the base bracket, and the main control terminal is respectively connected with the interactive terminal and the landing gear, and is used to send the received user voice signal to the cloud computing platform, and the cloud computing platform will recognize the user voice signal and return to the corresponding control The command is sent to the main control terminal, and the main control terminal drives the landing gear to perform the corresponding operation according to the received control command.

In one embodiment, the landing gear includes an electric push rod and a landing link, the landing link is an L-shaped structure, wherein the corner of the landing link is hinged with the base bracket through bolts, and the long end of the landing link is connected to the front support plate. The lifting beam is fixedly connected, wherein the lifting beam can be slidably connected with the bottom of the front support plate;

One end of the electric push rod is hinged on the base bracket, and the other end is hinged on the short end of the landing link;

The push of the electric push rod makes the lifting link rotate around its hinge point with the base bracket, so as to realize the back-up motion of the front support plate.

In one embodiment, the front end of the base bracket is further provided with a backrest.

In one embodiment, the connecting device is a robotic arm, and the interactive terminal is detachably and fixedly connected to the end of the robotic arm.

In one embodiment, the interactive terminal is provided with a display screen and a control panel.

In one embodiment, the main control terminal includes a processing module, a control module and a communication module;

The processing module is used to receive the user voice signal collected by the interactive terminal;

The communication module is wirelessly connected to the cloud computing platform, and is used for sending the user voice signal to the cloud computing platform, and receiving the control command returned by the cloud computing platform after recognizing the user voice signal;

The control module is used for controlling the electric push rod to perform corresponding push-pull motion according to the received control command.

In one embodiment, the cloud computing platform includes a communication system and a speech recognition system,

The communication system is used to receive the user voice signal transmitted by the main control terminal of one or more nursing beds;

The speech recognition system includes a preprocessing unit, an endpoint detection unit, a feature extraction unit and a recognition unit,

The preprocessing unit is used to enhance and denoise the received user voice signal, and output the preprocessed user voice signal;

The endpoint detection unit is used to perform endpoint detection on the preprocessed user voice signal, and segment the voice part in the voice signal;

The feature extraction unit is used to perform feature extraction processing on the voice part, obtain the logarithmic energy spectrum of the voice part, and obtain the voice feature parameter of the voice part according to the energy distribution information of the logarithmic energy general;

The identification unit is used to compare the acquired characteristic parameter with the characteristic parameter pre-stored in the database, identify the control instruction corresponding to the characteristic parameter, and return the acquired control instruction to the main control terminal specifying the bed management through the communication system.

The beneficial effects of the present invention are as follows: in order to improve the comfort of the user, the multifunctional nursing bed proposed by the present invention can adjust the rising and falling back angle of the front bed by setting the adjustable front bed board and the coordination of the lifting device, so that the user can Adjust the angle of the front bed in the most comfortable state for yourself, with strong comfort.

At the same time, a new type of lifting device support structure is proposed. The back-up motion of the front support plate is realized by using an L-shaped lifting rod and an electric push rod. The L-shaped lifting rod can enhance the stability during lifting and improve the nursing bed comfort.

At the same time, an interactive terminal is set on the side of the nursing bed through a connecting device such as a robotic arm, so that the interactive terminal is maintained and fixed in a suitable position on the side or above the nursing bed, which can facilitate the user to use the interactive terminal when lying on the nursing bed. The interactive terminal watches the multimedia information locally or downloaded from the cloud computing platform.

Further, a voice collection device is also provided on the interactive terminal to collect the user's voice control instructions. The user can send the voice control instructions by issuing the voice control instructions, which are received by the interactive device and sent to the main control terminal, and then the main control terminal sends the voice signal to the main control terminal. The cloud computing platform performs speech recognition processing, and receives the corresponding control instructions returned by the cloud computing platform, and the main control terminal drives or controls the corresponding mechanism (such as a landing gear) to perform corresponding actions; Part of it is completed through the set cloud computing platform, which can save the input cost of the local master control terminal. Through the cloud computing platform, the information processing tasks of multiple nursing beds can be centrally processed at the same time, reducing the difficulty of setting nursing beds.

In addition, when the cloud computing platform recognizes and processes the received voice signal, it specifically extracts the features of the received voice signal through modular design, and accurately identifies the corresponding control command according to the characteristic parameters of the voice signal and returns to the Nursing beds perform corresponding operations with strong reliability and high accuracy.

Description of drawings

The present invention will be further described by using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, under the premise of no creative work, other Attached.

Fig. 1 is the structure diagram of the nursing bed of the present invention;

Fig. 2 is the structure diagram of the landing gear of the present invention;

Fig. 3 is the working state diagram of the landing gear of the present invention;

FIG. 4 is a frame structure diagram of the main control terminal of the present invention

FIG. 5 is a frame structure diagram of the speech recognition system in the cloud computing platform of the present invention.

Reference number:

1- base bracket, 2- front support plate, 3- rear support plate, 4- electric push rod, 5- lift link, 6- interactive terminal, 7- main control terminal, 8-, 9- backrest, 21- lift Beam, 71-processing module, 72-control module, 73-communication module, 80-communication system, 81-speech recognition system, 811-preprocessing unit, 812-endpoint detection unit, 813-feature extraction unit, 814-recognition unit , 815-Database.

Detailed ways

The present invention will be further described with reference to the following application scenarios.

1-3, it shows a multifunctional nursing bed, including: a base bracket 1, a front support plate 2, a rear support plate 3 and a lifting device;

The rear support plate 3 is fixed on the base bracket 1, and the landing beam 21 of the front support plate 2 is connected with the landing gear. The front support plate 2 can rotate around the hinge point of the front support plate 2 and the rear support plate 3 with the push of the landing gear. , to realize the back up of the front support plate 2;

The side of the base bracket 1 is provided with a connection device, the end of the connection device is provided with an interactive terminal 6, and a voice acquisition device is provided on the interactive terminal 6, and the voice acquisition device is used to collect user voice signals;

The base bracket 1 is provided with a main control terminal 7, and the main control terminal 7 is respectively connected with the interactive terminal 6 and the landing gear, and is used for sending the received user voice signal to the cloud computing platform 8, and the cloud computing platform 8 performs the user voice signal. After identification, the corresponding control command is returned to the main control terminal 7, and the main control terminal 7 drives the landing gear to perform the corresponding operation according to the received control command.

In one embodiment, referring to FIGS. 2 and 3 , the landing gear includes an electric push rod 4 and a landing link 5 . The landing link 5 is an L-shaped structure, and the corner of the landing link 5 is connected to the base bracket 1 through bolts. Hinged, the long end of the lift link 5 is fixedly connected with the lift beam 21 of the front support plate 2, wherein the lift beam 21 is slidably connected to the bottom of the front support plate 2;

One end of the electric push rod 4 is hinged on the base bracket 1, and the other end is hinged on the short end of the landing link 5;

The push of the electric push rod 4 makes the lift link 5 rotate around its hinge point with the base bracket 1 , so as to realize the back-up motion of the front support plate 2 .

In one embodiment, the bottom of the front support plate 2 is provided with a chute along the front-rear direction, and the lifting beam 21 is clamped in the chute and can slide along the front-rear direction at the bottom of the front support plate 2 .

In one scenario, since the front support plate 2 and the landing rod each rotate around its own hinge point during the rotation, the relative distance between the front support plate 2 and the landing rod will change, so that the front support plate 2 can After the rotation is completed, the lifting beam 21 has a movable gap in the upper and lower positions of the chute, or the lifting beam 21 is connected to the bottom of the front support plate 2 and the landing rod respectively using a retractable elastic member that can be stretched up and down.

Preferably, the angle at the L-shaped corner of the landing rod is 75-105 degrees.

The above-mentioned embodiment of the present application proposes a new type of support structure for the landing gear. By using an L-shaped landing rod to cooperate with an electric push rod, the back-up motion of the front support plate is realized, so as to drive the short arm to drive the long arm to rotate around the hinge point, thereby pushing the The front support plate has high stability, which improves the comfort of the nursing bed during the lifting and lowering of the front support plate, and further improves the user experience.

In one embodiment, the front end of the base bracket 1 is further provided with a backrest 9 .

In one embodiment, the connecting device is a robotic arm, and the interactive terminal 6 is detachably and fixedly connected to the end of the robotic arm.

In one embodiment, the interactive terminal 6 is provided with a display screen and a control panel.

In one embodiment, the main control device has a built-in multimedia application program, and the interactive terminal 6 acts as a display device of the main control device to play the interface or screen when the main control device runs the multimedia application program.

In one scenario, the interactive terminal 6 is connected to the main control terminal 7, an application program is installed in the main control terminal 7, the interactive terminal 6 serves as a display device of the main control terminal 7, and the user can use different application programs through the interactive terminal 6, For example, playing video, text content, or just making a video call with another terminal, by setting the interactive terminal 6 on the side of the nursing bed to meet the user's needs for entertainment, social interaction, health care, calling, and adjusting the nursing bed, etc., provide users with rich functional support .

In one scenario, the interactive terminal 6 is provided with a microphone for receiving the user's voice signal, wherein the interactive terminal 6 can be awakened by the user's voice, for example, when the user sends out a specified voice signal, such as "start control"; the microphone receives After the voice signal is received, the main control device sends it to the cloud computing platform 8 for recognition. When the input voice signal is recognized as "start control", the command to wake up the interactive terminal 6 is returned.

In one embodiment, the main control terminal 7 includes a processing module 71, a control module 72 and a communication module 73;

The processing module 71 is used to receive the user voice signal collected by the interactive terminal 6;

The communication module 73 is connected in wireless communication with the cloud computing platform 8, and is used for sending the user's voice signal to the cloud computing platform 8, and receiving the control instruction returned after the cloud computing platform 8 recognizes the user's voice signal;

The control module 72 is configured to control the electric push rod 4 to perform corresponding push-pull movements according to the received control instructions.

In one scenario, the control module 72 can also be connected to other external devices through a dedicated interface, and the control module 72 sends control instructions to other external devices (such as temperature adjustment devices, alarm devices, etc.), and drives the external devices to execute corresponding Actions. Through the main control terminal 7, different external devices can be controlled by voice, and the expansibility is strong.

In the multifunctional nursing bed proposed by the above-mentioned embodiments of the present invention, in order to improve the user's comfort, by setting the adjustable and movable front bed board and the cooperation of the lifting device, the lifting and falling back angle of the front bed board can be adjusted, so that the user can adjust the lifting and falling back angle of the front bed board according to the most comfortable one. The state adjusts the angle of the front bed board, and the comfort is strong.

At the same time, an interactive terminal is set on the side of the nursing bed through a connecting device such as a robotic arm, so that the interactive terminal is maintained and fixed in a suitable position on the side or above the nursing bed, which can facilitate the user to use the interactive terminal when lying on the nursing bed. The interactive terminal watches the multimedia information locally or downloaded from the cloud computing platform.

Further, a voice collection device is also provided on the interactive terminal to collect the user's voice control instructions. The user can send the voice control instructions by issuing the voice control instructions, which are received by the interactive device and sent to the main control terminal, and then the main control terminal sends the voice signal to the main control terminal. The cloud computing platform performs speech recognition processing, and receives the corresponding control instructions returned by the cloud computing platform, and the main control terminal drives or controls the corresponding mechanism (such as a landing gear) to perform corresponding actions; Part of it is completed by the set cloud computing platform 8, which can save the input cost of the local master control terminal. The cloud computing platform 8 can simultaneously process the information processing tasks of multiple nursing beds, reducing the difficulty of setting nursing beds.

In one embodiment, referring to FIG. 4, the cloud computing platform 8 includes a communication system 80 and a speech recognition system 81,

The communication system 80 is used to receive the user voice signal transmitted by the main control terminal 7 of one or more nursing beds;

Referring to FIG. 5, the speech recognition system 81 includes a preprocessing unit 811, an endpoint detection unit 812, a feature extraction unit 813 and an identification unit 814,

The preprocessing unit 811 is configured to perform enhanced denoising processing on the received user voice signal, and output the preprocessed user voice signal;

The endpoint detection unit 812 is used to perform endpoint detection on the preprocessed user voice signal, and segment the voice part in the voice signal;

The feature extraction unit 813 is used to perform feature extraction processing on the voice part, obtain the logarithmic energy spectrum of the voice part, and obtain the voice feature parameter of the voice part according to the energy distribution information of the logarithmic energy common;

The identification unit 814 is used to compare the acquired characteristic parameters with the pre-stored characteristic parameters in the database 815, identify the control instructions corresponding to the characteristic parameters, and return the acquired control instructions to the main control terminal of the designated bed management through the communication system 80. 7.

Wherein the database 815 pre-stores user voice signal characteristic parameters and corresponding control instruction information;

In the above-mentioned embodiments of the present invention, a communication system is provided in the cloud computing platform to be connected to the main control terminals in different nursing beds, and can simultaneously receive, process, and feed back information (such as user voice signals) sent by different nursing beds; realize the processing system in the cloud Centralized processing reduces the information processing pressure of the main control terminal in the nursing bed. It is helpful for information interaction between nursing beds or between nursing beds and cloud computing platforms.

In the above-mentioned embodiments of the present invention, when the cloud computing platform recognizes and processes the received voice signal, it specifically adopts a modular design to accurately extract the features of the received voice signal, and accurately identify its corresponding control according to the characteristic parameters of the voice signal. Instruct and return to the nursing bed to perform the corresponding operation, with strong reliability and high accuracy.

In one embodiment, the preprocessing unit 811 performs enhanced denoising processing on the received user speech signal, specifically including:

1) Fast Fourier transform is performed on the obtained user voice signal y(t), and the amplitude spectrum |Y(w)| of the user voice signal is obtained;

其中该谱减函数为：2) Use the improved spectral subtraction function to perform spectral subtraction processing on the acquired user voice signal amplitude spectrum |Y(w)|, and output the processed user voice signal amplitude spectrum

where the spectral reduction function is:

表示谱减处理后的用户语音信号幅度谱中第w个频谱分量的幅值，|Y(w)|表示谱减处理前的用户语音信号幅度谱，|N(w)|表示噪声幅度谱估计，根据获取的用户语音信号的前若干帧语音信号作为噪声估计所得，|Y_max(w)|表示谱减处理前的用户语音信号的幅度谱中各频谱分量的幅值的最大值，P表示设定的谱减调节因子，Q噪声权重比因调节因子；In the formula,

represents the amplitude of the wth spectral component in the amplitude spectrum of the user speech signal after spectral subtraction processing, |Y(w)| represents the amplitude spectrum of the user speech signal before spectral subtraction processing, and |N(w)| represents the noise amplitude spectrum estimation , according to the obtained speech signals of the first several frames of the user speech signal as noise estimation, |Y _max (w)| represents the maximum value of the amplitude of each spectral component in the amplitude spectrum of the user speech signal before spectral subtraction processing, and P represents The set spectral reduction adjustment factor, the Q noise weight ratio is due to the adjustment factor;

3) reconstructing the amplitude spectrum of the processed user voice signal to obtain the user voice signal after spectral subtraction processing;

4) Perform wavelet decomposition on the user speech signal after spectral subtraction processing according to the set wavelet base and decomposition scale, and obtain low-frequency wavelet coefficients and high-frequency wavelet coefficients of different scales;

5) Thresholding the high-frequency wavelet coefficients of each scale, and outputting the high-frequency wavelet coefficients after thresholding;

6) Reconstruct according to the low-frequency wavelet coefficients and the high-frequency wavelet coefficients of each decomposition scale, and output the preprocessed user speech signal.

Preferably, the first 3-5 frames of the acquired user speech signal are used as the noise estimate, and Fourier transform is performed on the noise estimate to obtain the noise magnitude spectrum estimate.

The above-mentioned embodiments of the present invention are aimed at the situation that there are various uncertain noises in the environment where the nursing bed is used, resulting in the uneven quality of the user's voice signal received by the cloud computing platform. The above-mentioned embodiments also provide targeted A voice signal preprocessing method preprocesses the received user voice signal. First, the ambient noise existing in the voice signal is estimated according to the ambient sound end in the voice signal, and the adaptive adjustment of the amplitude spectrum of the voice signal is performed. The actual situation of the sound part The improved spectral subtraction function performs spectral subtraction processing on the amplitude spectrum of the speech signal to remove the environmental noise interference in the speech signal; after that, the wavelet threshold denoising method is used to remove the received noise in the speech signal due to the transmission process. The random noise further improves the preprocessing effect of the speech signal, and lays a foundation for the accurate recognition of the speech signal in the future.

In one embodiment, in the preprocessing unit 811, threshold processing is performed on the high-frequency wavelet coefficients of each scale, which specifically includes:

The high-frequency wavelet coefficients are thresholded by an improved threshold function, where the threshold function is:

表示阈值处理后的第a层第b个高频小波系数，r_a,b表示阈值处理前的第a层第b个高频小波系数，Z表示设定的阈值，g表示设定的门闸调节因子，h表示变形调节因子。In the formula,

Represents the b-th high-frequency wavelet coefficient of the a-th layer after thresholding, r _a,b represents the b-th high-frequency wavelet coefficient of the a-th layer before thresholding, Z represents the set threshold, and g represents the set gate adjustment factor, h represents the deformation adjustment factor.

The above-mentioned embodiment of the present invention further proposes an improved wavelet threshold processing function. The threshold processing function is divided into three parts in total. To avoid the traditional wavelet coefficients, the wavelet coefficients in the middle section are directly set to zero to deal with the distortion effect. At the same time, it avoids the phenomenon of "overprocessing" of the threshold function, and ensures the denoising effect of thresholding the high-frequency coefficients.

In one embodiment, in the preprocessing unit 811, the above-mentioned threshold processing is performed on the high-frequency wavelet coefficients, wherein the thresholds used for different decomposition scales are obtained by the following functions:

表示第A层高频小波系数的均值，L表示该高频小波系数的个数。In the formula, Z _a represents the set threshold of the a-th layer, r _{A, b} represents the b-th high-frequency wavelet coefficient of the A-th layer, where A represents the maximum number of decomposition layers,

represents the mean value of the high-frequency wavelet coefficients of the A-th layer, and L represents the number of the high-frequency wavelet coefficients.

Since the noise information reflected in the high-frequency coefficients can be easily distinguished when the number of decomposition layers is larger, the above-mentioned embodiments of the present invention adaptively set the high-frequency coefficients according to the distribution of the high-frequency coefficients in the maximum decomposition layers in the wavelet decomposition. The threshold value is determined, and the improved threshold value processing function is used to further improve the processing effect of the speech signal on random noise.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, not to limit the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should analyze the , the technical solutions of the present invention may be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims (6)

1. a multifunctional nursing bed, characterized in that, comprising: a base bracket, a front support plate, a rear support plate and a landing gear; The rear support plate is fixed on the base bracket, the landing beam of the front support plate is connected with the landing gear, and the front support plate can go around the hinge point of the front support plate and the rear support plate with the push of the landing gear Rotate to realize the back up of the front support plate; A connection device is provided on the side of the base bracket, an interactive terminal is provided at the end of the connection device, and a voice acquisition device is provided on the interactive terminal, and the voice acquisition device is used to collect user voice signals; A main control terminal is arranged in the base bracket, and the main control terminal is respectively connected with the interactive terminal and the landing gear, and is used for sending the received user voice signal to the cloud computing platform, and the cloud computing platform can monitor the The user's voice signal is recognized, and corresponding control commands are returned to the main control terminal, and the main control terminal drives the landing gear to perform corresponding operations according to the received control commands. 2 . The multifunctional nursing bed according to claim 1 , wherein the lifting device comprises an electric push rod and a lifting link, and the lifting link is an L-shaped structure, wherein the lifting link has an L-shaped structure. 3 . The corner is hinged with the base bracket through bolts, and the long end of the lift link is fixedly connected with the lift beam of the front support plate, wherein the lift beam is slidably connected to the bottom of the front support plate; One end of the electric push rod is hinged on the base bracket, and the other end is hinged on the short end of the lift link; The push of the electric push rod makes the lift link rotate around its hinge point with the base bracket, so as to realize the back-up motion of the front support plate. 3. The multifunctional nursing bed according to claim 1, wherein the front end of the base bracket is further provided with a backrest; and/or The connecting device is a robotic arm, and the interactive terminal is detachably and fixedly connected to the end of the robotic arm; and/or The interactive terminal is provided with a display screen and a control panel. 4. The multifunctional nursing bed according to claim 1, wherein the main control terminal comprises a processing module, a control module and a communication module; The processing module is used to receive the user voice signal collected by the interactive terminal; The communication module is wirelessly connected to the cloud computing platform, and is used for sending the user voice signal to the cloud computing platform, and receiving a control instruction returned by the cloud computing platform after recognizing the user voice signal; The control module is configured to control the electric push rod to perform corresponding push-pull movements according to the received control instructions. 5. The multifunctional nursing bed according to claim 4, wherein the cloud computing platform comprises a communication system and a voice recognition system, The communication system is configured to receive user voice signals transmitted by one or more main control terminals of the nursing bed; The speech recognition system includes a preprocessing unit, an endpoint detection unit, a feature extraction unit and a recognition unit, The preprocessing unit is configured to perform enhanced denoising processing on the received user voice signal, and output the preprocessed user voice signal; The endpoint detection unit is used to perform endpoint detection on the preprocessed user voice signal, and segment the voice part in the voice signal; The feature extraction unit is used to perform feature extraction processing on the voice part, obtain the logarithmic energy spectrum of the voice part, and obtain the voice feature parameter of the voice part according to the energy distribution information of the logarithmic energy spectrum; The identification unit is used to compare the acquired characteristic parameter with the characteristic parameter pre-stored in the database, identify the control instruction corresponding to the characteristic parameter, and return the acquired control instruction to the designated bed management through the communication system. master terminal. 6. The multifunctional nursing bed according to claim 5, wherein the preprocessing unit performs enhancement and denoising processing on the received user voice signal, specifically comprising: 1) Fast Fourier transform is performed on the obtained user voice signal y(t), and the amplitude spectrum |Y(w)| of the user voice signal is obtained; 2) Use the improved spectral subtraction function to perform spectral subtraction processing on the acquired user voice signal amplitude spectrum |Y(w)|, and output the processed user voice signal amplitude spectrum

where the spectral reduction function is:

In the formula,

represents the amplitude of the wth spectral component in the amplitude spectrum of the user speech signal after spectral subtraction processing, |Y(w)| represents the amplitude spectrum of the user speech signal before spectral subtraction processing, and |N(w)| represents the noise amplitude spectrum estimation , according to the obtained speech signals of the first several frames of the user speech signal as noise estimation, |Y _max (w)| represents the maximum value of the amplitude of each spectral component in the amplitude spectrum of the user speech signal before spectral subtraction processing, and P represents The set spectral reduction adjustment factor, the Q noise weight ratio is due to the adjustment factor; 3) reconstructing the amplitude spectrum of the processed user voice signal to obtain the user voice signal after spectral subtraction processing; 4) Perform wavelet decomposition on the user speech signal after spectral subtraction processing according to the set wavelet basis and decomposition scale, and obtain low-frequency wavelet coefficients and high-frequency wavelet coefficients of different scales; 5) Thresholding the high-frequency wavelet coefficients of each scale, and outputting the high-frequency wavelet coefficients after thresholding; 6) Reconstruct according to the low-frequency wavelet coefficients and the high-frequency wavelet coefficients of each decomposition scale, and output the preprocessed user speech signal. Preferably, the first 3-5 frames of the acquired user speech signal are used as the noise estimate, and Fourier transform is performed on the noise estimate to obtain the noise magnitude spectrum estimate.

Images