A voice input system for intelligently controlling infusion equipment
Through the voice input system of intelligently controlling the infusion device, the voice recognition module is used to process the infusion solution for medical infusion solutions, the problems of infusion parameters and improper selection of consumables caused by manual input are solved, and the infusion accuracy and efficiency are achieved, which reduces the medical staff's Work burden.
Patent Information
- Application Number
- CN201711048282.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2017-10-31
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2037-10-31
AI Technical Summary
The infusion parameter setting and consumable selection of existing medical pump products mainly rely on manual input, which can easily lead to parameter errors and improper selection of consumables, resulting in inaccurate infusion accuracy, and the risk of excessive or too little medicine may occur, and the setting time will be long, which will increase The risk of delayed infusion.
Design a voice input system that intelligently controls infusion equipment, enters the medical infusion scheme through a handheld terminal, and uses the voice recognition module to recognize the voice signal as an electrical signal, and transmits it to the infusion equipment through the medium to realize the download of text information and infusion operations, reducing the The probability of manual input errors improves infusion accuracy and efficiency.
It reduces the chance of infusion parameters errors and inappropriate consumable selection, reduces the risk of accidents caused by inaccurate infusion accuracy, saves time to set infusion parameters, improves the work efficiency of medical staff, and reduces their burden.
Smart Images
Figure CN107742519B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a medical auxiliary control system, in particular to a voice input system for intelligently controlling an infusion device. Background Art
[0002] At present, the infusion parameter settings of most medical pump products rely solely on manual input by clinical medical staff. Manual input may cause errors in the infusion parameters, resulting in excessive or insufficient infusion. It may also cause the infusion parameter setting time to be too long, resulting in the risk of delayed infusion. On the other hand, the brand and specifications of consumables are also mainly selected manually by clinical medical staff, which has certain errors. Once the selection is inappropriate, it will cause inaccurate infusion accuracy, resulting in the risk of accidental overdose of drug or too little drug. Summary of the invention
[0003] In order to solve the above problems, the purpose of the present invention is to provide a voice input system for intelligently controlling infusion equipment. Compared with manual input, it reduces the probability of infusion parameter errors and inappropriate selection of consumables brands and specifications, and saves the time for setting infusion parameters.
[0004] The technical solution adopted by the present invention to solve the problem is:
[0005] A voice input system for intelligently controlling an infusion device comprises an infusion device, a handheld terminal and a voice recognition module. The handheld terminal voice-enters a doctor's infusion plan and transmits the doctor's infusion plan to the voice recognition module. The doctor's infusion plan is recognized and processed by the voice recognition module and output in the form of an electrical signal. The electrical signal is transmitted to the infusion device through a medium and downloaded to the infusion device in the form of text information. The infusion device inputs a drug solution or a nutrient solution to a clinical patient according to the text information.
[0006] Furthermore, the speech recognition module includes a front-end module and a back-end module. The front-end module is used to preprocess the speech signal and establish a model library, the model library includes a sound model library and a language model library, and the back-end module is used to recognize the preprocessed speech signal and match it with the model library.
[0007] Furthermore, the front-end module pre-processes the speech signal including sampling, filtering to overcome aliasing, endpoint detection, removing pronunciation differences and reducing environmental noise.
[0008] Furthermore, the steps for the front-end module to establish the model library are: first, extract features from the preprocessed speech signal, then train the extracted features, and finally establish the model library according to predetermined rules.
[0009] Furthermore, the back-end module recognizes the pre-processed speech signal, including identifying the language information encoding and using distinguishable discrete symbols to represent the sound signal, as well as analyzing the grammatical semantics and terminology norms.
[0010] Further, matching with the model library includes comparing the similarity between the extracted features of the preprocessed speech signal and the stored features in the model library to match the recognition result of the speech output signal.
[0011] Furthermore, the training of the extracted features is done offline.
[0012] Preferably, the handheld terminal is a mobile phone, a tablet or a computer.
[0013] Preferably, the infusion equipment includes an infusion pump, a syringe pump, a nutrition pump, a blood transfusion pump and a target-controlled pump.
[0014] The beneficial effect of the present invention is as follows: the present invention provides a voice input system for intelligently controlling infusion equipment, and medical staff only need to input the doctor's infusion plan through a handheld terminal. The voice signal is recognized by the voice recognition module and output as an electrical signal, and finally transmitted to the infusion equipment in the form of text information. The infusion equipment inputs liquid medicine or nutrient solution for clinical patients according to the text information. Therefore, compared with manual input, the present invention reduces the probability of errors in infusion parameters and inappropriate selection of brands and specifications of consumables, reduces the risk of accidental overdose of liquid medicine or too little liquid medicine caused by inaccurate infusion accuracy, and saves the time for setting infusion parameters, avoids the risk of delayed infusion, and can save time and reduce the burden on medical staff as a whole, thereby improving the efficiency of medical staff. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Preferred embodiments of the present invention are given below in conjunction with the accompanying drawings to illustrate the implementation scheme of the present invention in detail.
[0016] Figure 1 It is a structural block diagram of the speech input system of the present invention;
[0017] Figure 2 It is a work flow chart of the speech recognition module of the present invention. DETAILED DESCRIPTION
[0018] Reference Figure 1 A voice input system for intelligently controlling an infusion device includes an infusion device 3, a handheld terminal 1 and a voice recognition module 2. The handheld terminal 1 voice-records a doctor's infusion plan and transmits the doctor's infusion plan to the voice recognition module 2. The doctor's infusion plan is recognized and processed by the voice recognition module 2 and output in the form of an electrical signal. The electrical signal is transmitted to the infusion device 3 through a medium and downloaded into the infusion device 3 in the form of text information. The infusion device 3 inputs a drug solution or a nutrient solution for a clinical patient according to the text information.
[0019] Specifically, medical staff can download the APP software of the voice input system of the present invention through the handheld terminal 1, and then log in to the APP software, and directly communicate with the handheld terminal 1 through voice to enter the doctor's order infusion plan. The APP software supports different languages and dialects. The doctor's order infusion plan includes the infusion device 3 number, patient information, infusion parameters, consumables brand and specifications, etc., and then is converted into an electrical signal through the voice recognition module 2. It is transmitted to the infusion device 3 through a medium, which can be Wi-Fi or broadband network, and finally downloaded to the corresponding infusion device 3 in the form of text information. At this time, the medical staff can confirm the doctor's order infusion plan again on the handheld terminal 1. If it is correct, press the start button, and the infusion device 3 starts to infuse the patient; therefore, compared with manual input, the present invention reduces the probability of infusion parameter errors and inappropriate brands and specifications of consumables, reduces the risk of accidental excessive drug solution shock or too little drug solution caused by inaccurate infusion accuracy, and saves the time for setting infusion parameters, avoids the risk of delayed infusion, and can save medical staff time and reduce the burden on medical staff as a whole, thereby improving the efficiency of medical staff.
[0020] Among them, refer to Figure 1 and Figure 2 The speech recognition module 2 includes a front-end module 21 and a back-end module 22. The front-end module 21 is used to preprocess the speech signal and establish a model library, which includes a sound model library and a language model library. The back-end module 22 is used to recognize the preprocessed speech signal and match it with the model library; the back-end module 22 also includes an adaptive feedback module that can self-learn the user's speech, thereby making necessary corrections to the sound model and language model to further improve the recognition accuracy.
[0021] Among them, the preprocessing of the voice signal by the front-end module 21 includes sampling, overcoming aliasing filtering, endpoint detection, removing pronunciation differences and reducing environmental noise. Sampling refers to extracting discrete components from a continuous voice signal for separate analysis. Aliasing refers to the phenomenon that the high-frequency signal in the voice signal is folded into the low-frequency band, resulting in the appearance of false frequency components. Overcoming aliasing filtering can further reduce the aliasing frequency components. The purpose of endpoint detection is to find suitable starting and ending points of the voice unit to obtain a clear sampling signal. The purpose of removing pronunciation differences is to minimize the errors caused by differences in user accents or speaking habits. The main purpose of reducing environmental noise is to remove unnecessary silence and non-speaking sounds. The essential characteristics of the voice signal can be more completely reflected through preprocessing.
[0022] Among them, the steps for the front-end module 21 to establish a model library are: first, extract features from the preprocessed speech signal, then train the extracted features, and finally establish the model library according to predetermined rules. The training of the extracted features is completed offline; the specific training method is: by collecting the features of a large number of speech signal samples, comparing to obtain different types of basic features, which correspond to different speech signals, and then sorting the data of these basic features to generate a model library; the predetermined rules are: classify speech signals with the same or similar basic features into the same model class, continue to compare speech signals that are not within this range, and so on, classify all speech signal samples.
[0023] Among them, the back-end module 22 recognizes the pre-processed voice signal, including identifying the language information encoding and using distinguishable discrete symbols to represent the sound signal, as well as analyzing grammatical semantics and terminology standards; the language information encoding in the voice signal is carried out according to the time change of the amplitude spectrum, and because the voice is readable, that is to say, the sound signal can be represented by multiple distinctive, discrete symbols without considering the information content conveyed by the speaker, and the interaction of voice is a cognitive process, which cannot be separated from grammar, semantics and terminology standards, so it is necessary to analyze grammatical semantics and terminology standards.
[0024] Among them, matching with the model library includes comparing the similarity between the extracted features of the preprocessed speech signal and the stored features in the model library to match the recognition results of the speech output signal. This recognition process is completed online and can match a comparison signal that is the same as or most similar to the features of the preprocessed speech signal, and output it to the next level in the form of an electrical signal.
[0025] Preferably, the handheld terminal 1 is a mobile phone, a tablet or a computer; the handheld terminal 1 is generally carried by medical staff, has a Wi-Fi wireless network card, and can remotely access the central monitoring server through the hospital Wi-Fi or 4G. It can display the basic status of any treatment unit in real time, receive alarm information and ring or / and vibrate prompts. At the same time, medical staff can observe the patient's infusion process and physiological monitoring information through the handheld terminal 1, so as to understand the treatment situation at any time.
[0026] Preferably, the infusion device 3 includes an infusion pump, a syringe pump, a nutrition pump, a blood transfusion pump and a target-controlled pump.
[0027] The above content has discussed in detail the preferred embodiments and basic principles of the present invention, but the present invention is not limited to the above-mentioned implementation modes. Those skilled in the art should understand that there may be various equivalent deformations and substitutions without violating the spirit of the present invention, and these equivalent deformations and substitutions all fall within the scope of the present invention to be protected.
Claims
1. A voice input system for intelligently controlling infusion equipment, Features: The invention comprises an infusion device (3), a handheld terminal (1) and a voice recognition module (2); the handheld terminal (1) voice-records a doctor's infusion plan and transmits the doctor's infusion plan to the voice recognition module (2); the doctor's infusion plan is recognized and processed by the voice recognition module (2) and then output in the form of an electrical signal; the electrical signal is transmitted to the infusion device (3) through a medium and downloaded into the infusion device (3) in the form of text information; the infusion device (3) implements the infusion of a drug solution or a nutrient solution for a clinical patient according to the text information; The speech recognition module (2) comprises a front-end module (21) and a back-end module (22), wherein the front-end module (21) is used to pre-process the speech signal and establish a model library, wherein the model library comprises a sound model library and a language model library, and the back-end module (22) is used to recognize the pre-processed speech signal and match it with the model library; The back-end module (22) recognizes the pre-processed speech signal, including recognizing the language information coding and using the distinguished discrete symbols to represent the sound signal, and analyzing the grammar, semantics and terminology norms; The back-end module (22) also includes an adaptive feedback module for correcting the sound model library and the language model library; The doctor's infusion plan includes at least one of the following: infusion equipment, patient information, infusion parameters, and consumables brand and specifications.
2. A voice input system for intelligently controlling an infusion device according to claim 1, Features: The front-end module (21) pre-processes the speech signal including sampling, filtering to overcome aliasing, endpoint detection, removing pronunciation differences and reducing environmental noise.
3. A voice input system for intelligently controlling an infusion device according to claim 1, Features: The steps of establishing the model library by the front-end module (21) are as follows: firstly, extracting features from the pre-processed speech signal, then training the extracted features, and finally establishing the model library according to predetermined rules.
4. A voice input system for intelligently controlling an infusion device according to claim 1, Features: Matching with the model library includes comparing the similarity between the extracted features of the preprocessed speech signal and the stored features in the model library to match the recognition result of the speech output signal.
5. A voice input system for intelligently controlling an infusion device according to claim 3, Features: The training of the extracted features is done offline.
6. A voice input system for intelligently controlling an infusion device according to any one of claims 1 to 5, Features: The handheld terminal (1) is a mobile phone, a tablet or a computer.
7. A voice input system for intelligently controlling an infusion device according to any one of claims 1 to 5, Features: The infusion equipment (3) comprises an infusion pump, an injection pump, a nutrition pump, a blood transfusion pump and a target control pump.