Space disease physiotherapy system fusing multi-source heterogeneous biological information

Abstract

The invention provides a space disease physiotherapy system fusing multi-source heterogeneous biological information, and relates to the technical field of intelligent spaceflight physiotherapy. The system comprises a broadcast module, a first acquisition module, a second acquisition module, a category determination module and a physiotherapy scheme module. The broadcasting module is used for generating and broadcasting space disease diagnosis problems according to a preset symptom description list, the first acquisition module and the second acquisition module are used for acquiring answer voice information and physiological index data, the category determination module is used for determining category information of space diseases suffered by astronauts, and the physiotherapy scheme module is used for determining physiotherapy schemes. According to the method and the device, accurate space disease screening prediction can be carried out on the astronaut according to the collected answer voice and the physiological indexes without guidance of professional physicians, dependence on the professional physicians is reduced, the accuracy and the efficiency of screening prediction of the space disease suffered by the astronaut are improved, effective support is provided for further determination of a physical therapy scheme, and the method and the device are suitable for popularization and application. And the physical therapy effect of space diseases is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent aerospace physiotherapy, and in particular to a space pathophysiotherapy system fusing multi-source heterogeneous biological information. BACKGROUND

[0002] There are many risk factors threatening the physical and mental health of astronauts in the long-term on-orbit working environment. Due to the great difference between the space weightless environment and the ground environment, bacteria and fungi are easily bred, causing the internal environment homeostasis of astronauts to be unbalanced, which may cause various space diseases. In addition, the closed working and living environment also threatens the mental health of astronauts, which is easy to cause various space psychological diseases. In order to carry out timely physiotherapy for astronauts suffering from space diseases, it is necessary to accurately screen and predict the space diseases suffered by astronauts and determine the physiotherapy scheme. In the related technology, professional physicians are usually required to manually screen and predict the space diseases suffered by astronauts, but in the closed space environment, it is difficult to obtain real-time guidance and suggestions from professional physicians, and the astronauts themselves also lack medical knowledge, so it is difficult to accurately screen and predict the space diseases suffered by astronauts, which affects the further determination of the physiotherapy scheme.

[0003] The information disclosed in the background section of this application is only intended to deepen the understanding of the general background of the application and should not be considered as an acknowledgment or implicit suggestion that this information constitutes prior art known to those skilled in the art. SUMMARY

[0004] The present application provides a space pathophysiotherapy system fusing multi-source heterogeneous biological information, which can solve the technical problem that the related technology cannot accurately screen and predict the space diseases suffered by astronauts.

[0005] According to a first aspect of the present application, a space pathophysiotherapy system fusing multi-source heterogeneous biological information is provided, comprising:

[0006] The broadcast module is configured to generate space disease diagnosis questions according to a preset symptom description list and broadcast the questions.

[0007] The first acquisition module is configured to acquire answer voice information corresponding to the space disease diagnosis questions.

[0008] The second acquisition module is configured to acquire a plurality of physiological index data of the astronauts.

[0009] The category determination module is configured to determine space disease category information of the astronauts according to the answer voice information and the physiological index data.

[0010] The physiotherapy scheme module is configured to determine a physiotherapy scheme according to the space disease category information.

[0011] According to the present application, the space disease diagnosis questions are generated and broadcasted, which includes:

[0012] According to the preset symptom description list, a plurality of symptom description texts are obtained;

[0013] According to the symptom description text, a space sickness diagnosis question is obtained;

[0014] The space sickness diagnosis question is broadcast through a sound equipment.

[0015] According to the present application, the space sickness category information of the astronaut is determined, including:

[0016] The speech recognition result of the answer voice information is obtained through a speech recognition model;

[0017] According to the speech recognition result, the judgment result of the symptom corresponding to the space sickness diagnosis question is determined;

[0018] According to the physiological index data and the preset physiological index threshold, a physiological index comparison result is obtained;

[0019] According to the symptom type corresponding to the plurality of space sickness categories, the judgment result is grouped to obtain a judgment result set corresponding to the space sickness category;

[0020] The physiological index comparison result and the judgment result set are input into a space sickness recognition model corresponding to the space sickness category to obtain a space sickness recognition result corresponding to the space sickness category;

[0021] According to the space sickness recognition result of the plurality of categories, the space sickness category information of the astronaut is determined.

[0022] According to the present application, the judgment result of the symptom corresponding to the space sickness diagnosis question is determined, including:

[0023] The speech recognition result is processed by word segmentation to obtain a plurality of word segmentation texts;

[0024] Each word segmentation text is processed by word embedding through a pre-trained language model to obtain a word segmentation word vector of each word segmentation text;

[0025] The symptom description text corresponding to the space sickness diagnosis question is processed by word embedding through a pre-trained language model to obtain a symptom word vector of each symptom description text;

[0026] The high-dimensional space distance of each word segmentation word vector and symptom word vector is determined respectively;

[0027] In the case where the high-dimensional space distance between the symptom word vector of the symptom description text and any one word segmentation word vector is less than or equal to a preset distance threshold, the judgment result of the symptom corresponding to the symptom description text in the space sickness diagnosis question is determined to exist.

[0028] According to the present application, the training step of the space disease identification model corresponding to the space disease category comprises:

[0029] Obtaining the sample physiological index comparison result of the training personnel and the sample judgment result of the plurality of symptoms corresponding to the space disease category;

[0030] Assembling the sample physiological index comparison result and the sample judgment result of the plurality of symptoms into a sample input vector;

[0031] Inputting the sample input vector into the space disease identification model corresponding to the space disease category to obtain a predicted identification result;

[0032] According to the predicted identification result and the labeled information of the training personnel, training the space disease identification model corresponding to the space disease category.

[0033] According to the present application, the predicted identification result comprises:

[0034] According to the formula

[0035]

[0036] The predicted identification result is obtained, wherein P(Y h |X j ) is the predicted identification result of the jth training personnel corresponding to the Yth space disease category, X j =(x 1,j ,x 2,j ,…,x m,j ) is the sample input vector of the jth training personnel, x 1,j is the data of the first dimension of the sample input vector, x 2,j is the data of the second dimension of the sample input vector, x i,j is the data of the ith dimension of the sample input vector, x m,j is the data of the mth dimension of the sample input vector, ω0 and ω i are parameters of the space disease identification model corresponding to the Yth space disease category, i≤m, and i, j and m are positive integers.

[0037] According to the present application, the training of the space disease identification model corresponding to the space disease category comprises:

[0038] According to the labeled information of the training personnel, determining the labeled identification result of the Yth space disease of the training personnel;

[0039] According to the labeled identification result and the predicted identification result, determining the log-likelihood function;

[0040] According to the log-likelihood function, parameters of the space disease identification model corresponding to the space disease category are adjusted to train the space disease identification model corresponding to the space disease category.

[0041] According to the present application, the log-likelihood function is determined, comprising:

[0042] According to the formula

[0043]

[0044] The log-likelihood function L is determined, wherein y j is the labeled identification result of the Yth space disease of the jth training personnel, N is the number of training personnel, j≤N, and j and N are positive integers.

[0045] According to the present application, the parameters of the space disease identification model corresponding to the space disease category are adjusted to train the space disease identification model corresponding to the space disease category, comprising:

[0046] According to the formula

[0047]

[0048] The adjusted i th parameter of the space disease identification model corresponding to the Y th space disease category is obtained Wherein, L is the log-likelihood function, alpha is the learning rate, y j is the labeled identification result of the Yth space disease of the jth training personnel, N is the number of training personnel, j≤N, and j and N are positive integers.

[0049] According to the second aspect of the present application, a space disease treatment method fusing multi-source heterogeneous biological information is provided, comprising:

[0050] According to the preset symptom description list, a space disease diagnosis question is generated and broadcasted;

[0051] Collecting answer voice information corresponding to the space disease diagnosis question;

[0052] Collecting a plurality of physiological index data of astronauts;

[0053] According to the answer voice information and the physiological index data, the space disease category information of the astronauts is determined;

[0054] According to the space disease category information, a treatment scheme is determined.

[0055] By adopting the above technical solutions, the present application can achieve the following technical effects:

[0056] According to the application, the space disease diagnosis question can be broadcast according to the preset symptom description list, the answer voice information and the physiological index data of the astronauts are collected, the space disease category information of the astronauts is determined, and the physiotherapy scheme is determined. The dependence on professional physicians is reduced, the accuracy and efficiency of screening and prediction of the space disease of the astronauts are improved, effective support is provided for further determination of the physiotherapy scheme, and the physiotherapy effect of the space disease is improved. When the space disease diagnosis question is determined, the space disease diagnosis question can be generated and broadcast based on the preset symptom description list, and the answer voice information and the physiological index data of the astronauts are collected. Without the guidance of professional physicians, the physical condition of the astronauts can be understood, and basic data for determining the space disease category information of the astronauts is provided. When the space disease recognition model is trained, the probability of each training personnel suffering from various space diseases is determined based on the Logistic regression model, and the physiological index and the symptoms of the astronauts are comprehensively considered, and the prediction accuracy is improved. Moreover, the parameters of the space disease recognition model are updated according to the maximum likelihood function and the gradient ascent method, so that the space disease recognition model is trained, and the accuracy of the space disease recognition model is improved. When the physiotherapy scheme is determined, the physiotherapy scheme corresponding to the space disease of the astronauts can be determined based on the space disease category information. The dependence on professional physicians is reduced, the accuracy and efficiency of screening and prediction of the space disease of the astronauts are improved, effective support is provided for further treatment decision, and the treatment effect of the space disease is improved.

[0057] It should be understood that the above general description and the following detailed description are exemplary and explanatory, but not limiting the present application. Other features and aspects of the present application will be more clearly understood from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other embodiments from these drawings without creative labor;

[0059] Figure 1 An exemplary schematic diagram of a space disease physiotherapy system fusing multi-source heterogeneous biological information according to an embodiment of the present application is shown;

[0060] Figure 2 An exemplary application schematic diagram of a space disease physiotherapy system fusing multi-source heterogeneous biological information according to an embodiment of the present application is shown;

[0061] Figure 3 An exemplary flowchart of a space disease physiotherapy method fusing multi-source heterogeneous biological information according to an embodiment of the present application is shown;

[0062] 101 - sleeping bag, 102 - space disease diagnosis and treatment intelligent controller, 103 - U-shaped groove, 104 - electrode sheet. DETAILED DESCRIPTION

[0063] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0064] The technical solutions of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described in some embodiments.

[0065] Figure 1 An exemplary schematic diagram of a space pathological treatment system fusing multi-source heterogeneous biological information according to an embodiment of the present application is shown, and the system comprises:

[0066] A broadcast module is configured to generate and broadcast space disease diagnosis questions according to a preset symptom description list.

[0067] A first collection module is configured to collect answer voice information corresponding to the space disease diagnosis questions.

[0068] A second collection module is configured to collect a plurality of physiological index data of astronauts.

[0069] A category determination module is configured to determine space disease category information of astronauts according to the answer voice information and the physiological index data.

[0070] A treatment scheme module is configured to determine a treatment scheme according to the space disease category information.

[0071] The space pathological treatment system fusing multi-source heterogeneous biological information according to the embodiment of the present application can broadcast space disease diagnosis questions according to a preset symptom description list, collect answer voice information and physiological index data of astronauts, thereby determining space disease category information of astronauts and determining a treatment scheme. The system reduces the dependence on professional physicians, improves the accuracy and efficiency of screening and prediction of space diseases of astronauts, provides effective support for further determination of a treatment scheme, and improves the treatment effect of space diseases.

[0072] Embodiment one:

[0073] According to the embodiment of the present application, the system can include a plurality of functional modules, the functions of which can be realized by the space disease diagnosis and treatment intelligent controller, the space disease diagnosis and treatment intelligent controller can be electrically connected with the sleeping bag of the astronaut, when the astronaut enters the sleeping bag to rest, the physiological index data of the astronaut can be collected and entered into the space disease diagnosis and treatment intelligent controller, the space disease diagnosis and treatment intelligent controller can also play the space disease diagnosis questions and collect the answer voice information of the astronaut's answers, and then analyze the answer voice information and the physiological index data to determine the type of the space disease suffered by the astronaut, thereby providing a basis for formulating a physiotherapy scheme. Moreover, the sleeping bag is a sleeping bag customized according to the astronaut's body shape, and electrode pads can be arranged at positions corresponding to a plurality of acupoints of the astronaut's body in the sleeping bag. The space disease diagnosis and treatment intelligent controller can select corresponding acupoints according to the type of the space disease suffered by the astronaut, and control the electrode pads corresponding to the acupoints to output current pulses, so as to stimulate the acupoints in the form of acupuncture physiotherapy, thereby producing the effect of physiotherapy.

[0074] Embodiment two:

[0075] According to an embodiment of the present application, the broadcast module can obtain the symptom description text from a preset symptom description list, generate the space disease diagnosis question and broadcast, wherein the preset symptom description list can be stored in the memory of the space disease diagnosis and treatment intelligent controller, and the preset symptom description list can be stored in groups according to the symptoms corresponding to different types of space diseases. For example, the symptom description text corresponding to the space psychological disease includes "low mood", "emotional fluctuation", "mental weakness", "anxiety", "fear"; the symptom description text corresponding to bacterial inflammation (conjunctivitis) includes "red eyes", "dryness", "itchy eyes", "eye pain", "foreign body sensation", "fear of light", "excessive secretion", "tears"; the symptom description text corresponding to bacterial inflammation (respiratory disease) includes "cough", "sputum", "hemoptysis", "chest tightness", "chest pain", "dyspnea", "tachypnea", "abnormal respiratory rate", "abnormal blood oxygen saturation"; the symptom description text corresponding to bacterial inflammation (tooth infection) includes "gum swelling", "gum bleeding", "gum pain"; the symptom description text corresponding to space blindness includes "blurred vision"; the symptom description text corresponding to osteoporosis includes "back pain", "leg pain", "other parts of the body hidden pain, acid pain and swelling pain"; the symptom description text corresponding to orthostatic intolerance includes "headache", "fatigue", "sleep disorder", "visual problems", "reduced exercise tolerance", "weakness", "wheeze", "shaking", "sweating", "anxiety", "palpitations", "syncope", "pre-syncope state", "abnormal heart rate", "abnormal blood pressure"; the symptom description text corresponding to muscle atrophy includes "muscle weakness", "limb thinning", "joint stiffness", "joint softness"; the symptom description text corresponding to space adaptation syndrome includes "loss of appetite", "drowsiness", "pale complexion", "cold sweat", "nausea and vomiting", "dizziness", "increased saliva", "belching", "headache".

[0076] According to an embodiment of the present application, the space disease diagnosis question is generated and broadcasted according to the preset symptom description list, including: obtaining a plurality of symptom description texts according to the preset symptom description list; obtaining a space disease diagnosis question according to the symptom description text; and broadcasting the space disease diagnosis question through a sound device.

[0077] According to the embodiment of the present application, the symptom description texts can be grouped according to the type of space disease to be determined, and the symptom description texts are combined to form a piece of text, i.e. the space disease diagnosis question. For example, when determining whether an astronaut suffers from muscle atrophy, the symptom description texts of muscle weakness, thin limbs, joint stiffness and joint soreness are combined and expanded to form a piece of text, i.e. the space disease diagnosis question of muscle atrophy, such as "Have you had the following symptoms recently: muscle weakness, thin limbs, joint stiffness, and joint soreness? If yes, please answer the specific symptoms; if no, answer 'no'". Further, the generated various space disease diagnosis questions can be broadcast through the loudspeaker of the space disease diagnosis and treatment intelligent controller.

[0078] According to the embodiment of the present application, the function of the first acquisition module can be realized by the microphone of the space disease diagnosis and treatment intelligent controller, and the answer voice information corresponding to the space disease diagnosis question can be acquired. For example, after broadcasting the space disease diagnosis question of muscle atrophy "Have you had the following symptoms recently: muscle weakness, thin limbs, joint stiffness, and joint soreness? If yes, please answer the specific symptoms; if no, answer 'no'", the answer voice information of the astronaut collected through the microphone can be "I have been experiencing muscle weakness and joint stiffness and soreness in my limbs recently", which is the answer voice information corresponding to the space disease diagnosis question of muscle atrophy.

[0079] According to the embodiment of the present application, the second acquisition module can be used to acquire various physiological index data of the astronaut. The astronaut wears various physiological index monitors to monitor various physiological index data of the astronaut, which can be respiratory rate, heart rate, blood pressure, and blood oxygen saturation, etc. The function of the second acquisition module can be realized by the touch screen of the space disease diagnosis and treatment intelligent controller. For example, the astronaut can manually input the physiological index data through the touch screen after learning the physiological index data of the astronaut through the physiological index monitor. The function of the second acquisition module can also be realized by the communication component of the space disease diagnosis and treatment intelligent controller. For example, the space disease diagnosis and treatment intelligent controller can be communicatively connected with the physiological index monitor, so as to acquire various physiological index data.

[0080] In this way, the space disease diagnosis questions can be generated and broadcast based on the preset symptom description list, and the answer voice information and various physiological index data of the astronaut can be collected, so as to understand the physical condition of the astronaut without the guidance of a professional doctor, and to provide basic data for determining the space disease category information of the astronaut.

[0081] According to an embodiment of the present application, the function of the category determining module can be realized by the intelligent controller for space disease diagnosis and treatment. According to the answer voice information and the physiological index data, the category information of the space disease suffered by the astronaut is determined, including: obtaining the voice recognition result of the answer voice information through a voice recognition model; determining the judgment result of the symptoms corresponding to the space disease diagnosis question according to the voice recognition result; obtaining the physiological index comparison result according to the physiological index data and the preset physiological index threshold; grouping the judgment result according to the symptom types corresponding to the multiple space disease categories, to obtain the judgment result set corresponding to the space disease category; inputting the physiological index comparison result and the judgment result set into the space disease recognition model corresponding to the space disease category, to obtain the space disease recognition result corresponding to the space disease category; and determining the category information of the space disease suffered by the astronaut according to the space disease recognition results of the multiple categories.

[0082] According to an embodiment of the present application, the voice recognition model is, for example, a GMM-HMM-based voice recognition system, an end-to-end deep learning model, etc., which can convert the answer voice information of the astronaut into text, i.e., obtain the voice recognition result.

[0083] According to an embodiment of the present application, the judgment result of the symptoms corresponding to the space disease diagnosis question is determined according to the voice recognition result, including: performing word segmentation processing on the voice recognition result to obtain multiple segmented texts; performing word embedding processing on each segmented text through a pre-trained language model to obtain the segmented word vector of each segmented text; performing word embedding processing on the symptom description text corresponding to the space disease diagnosis question through the pre-trained language model to obtain the symptom word vector of each symptom description text; respectively determining the high-dimensional space distance of each segmented word vector and symptom word vector; and in the case that the high-dimensional space distance between the symptom word vector of the symptom description text and any one segmented word vector is less than or equal to a preset distance threshold, determining the judgment result of the symptom corresponding to the symptom description text in the space disease diagnosis question as existing.

[0084] According to embodiments of the present invention, the speech recognition result can be segmented (e.g., using Jieba segmentation) to obtain multiple segmented texts. For example, if the speech recognition result is "Recently feeling chest tightness, chest pain, and difficulty breathing", the result obtained through Jieba segmentation is "Recently / feeling / chest tightness / chest pain / difficulty breathing / ". Each segmented text is then embedded using a pre-trained language model (e.g., BERT, XLNet, etc.) to obtain a vector for each segmented text, which is the segmented word vector. Similarly, a pre-trained language model can be used to embed the symptom description text in the space disease diagnosis problem to obtain the symptom word vector for each symptom description text. The dimensions of the segmented word vectors and symptom word vectors are determined by the pre-trained language model and are the same. For example, the segmented word vectors and symptom word vectors obtained through the BERT model are both 768-dimensional. By determining the absolute value of the difference between the distances in each dimension of each segmented word vector and the symptom word vector, and summing them, the high-dimensional spatial distance between each segmented word vector and the symptom word vector can be obtained. Of course, the high-dimensional spatial distance can also be the Euclidean distance between the segmented word vector and the symptom word vector, and this invention does not limit this. If the high-dimensional spatial distance between the symptom word vector of the symptom description text and any segmented word vector is less than or equal to a preset distance threshold, the judgment result of the symptom corresponding to the symptom description text is determined to exist. For example, the high-dimensional spatial distance between the segmented word vector corresponding to lower back and knee pain and the symptom word vector corresponding to the symptom description text "joint weakness" of muscular dystrophy is 0.1, which is less than the preset distance threshold of 0.5. Therefore, the judgment result of the symptom "joint weakness" in the space disease diagnosis question corresponding to muscular dystrophy can be determined to exist, thereby determining whether the symptoms answered by the astronauts exist as the symptoms corresponding to the various space disease diagnosis questions.

[0085] According to embodiments of the present invention, physiological indicator comparison results can also be obtained based on the physiological indicator data and preset physiological indicator thresholds. The physiological indicator thresholds can be normal ranges for the physiological indicators, for example, a respiratory rate threshold of 12-20 breaths / minute, a heart rate threshold of 60-100 beats / minute, a blood pressure threshold of 90-139 mmHg systolic blood pressure, 60-89 mmHg diastolic blood pressure, and a blood oxygen saturation threshold of 95%-99%. If the physiological indicator data is not within the preset physiological indicator threshold range, the physiological indicator comparison result is determined to be 1; if the physiological indicator data is within the preset physiological indicator threshold range, the physiological indicator comparison result is determined to be 0. Combining the comparison results of each physiological indicator yields a physiological indicator comparison result vector for the astronaut. For example, if the comparison results of the astronaut's five physiological indicators are 1, 1, 1, 0, and 1, the resulting physiological indicator comparison result vector is (1, 1, 1, 0, 1).

[0086] According to an embodiment of the present application, the judgment results are grouped according to the symptom types corresponding to the multiple space disease categories, to obtain a judgment result set corresponding to a space disease category. For example, the judgment results of multiple symptoms corresponding to each space disease are obtained above for the types of multiple space diseases, and the judgment results can still be grouped according to the symptoms corresponding to the space disease, to obtain a judgment result set corresponding to a space disease category. For example, the judgment results of four symptoms of muscle weakness, thin limbs, joint stiffness, and joint softness corresponding to muscular dystrophy are existence, existence, existence, and nonexistence, respectively, and 1 is used to represent existence of a symptom and 0 is used to represent nonexistence of a symptom. Therefore, the judgment result set corresponding to muscular dystrophy is {1, 1, 1, 0}, which can be expressed in the form of a vector, i.e., (1, 1, 1, 0).

[0087] According to an embodiment of the present application, the physiological index comparison result and the judgment result set are input into a space disease recognition model corresponding to a space disease category, to obtain a space disease recognition result corresponding to the space disease category. The space disease recognition model can be any type of classification model, such as a Logistic regression model, a support vector machine model, a neural network model, etc., and the present application does not limit this. In addition, each space disease can correspond to a trained space disease recognition model. The physiological index comparison result vector and the vector corresponding to the judgment result set of the space disease are spliced into a long vector. For example, the physiological index comparison result vector is (1, 1, 1, 0, 1), the vector corresponding to the judgment result set of muscular dystrophy is (1, 1, 1, 0), the spliced long vector is (1, 1, 1, 0, 1, 1, 1, 1, 0), which is input into the space disease recognition model corresponding to muscular dystrophy, to obtain the space disease recognition result of muscular dystrophy. Similarly, the vector corresponding to the judgment result set of each space disease and the physiological index comparison result vector are spliced into a long vector, which is input into the space disease recognition model corresponding to the space disease, to obtain the space disease recognition result of the space disease. The space disease recognition result corresponding to the space disease category is a value in the range of [0, 1], which can describe the probability of the astronaut suffering from the corresponding space disease.

[0088] According to the embodiment of the present application, the space disease category information of the astronauts can be determined according to the identification results of various categories of space diseases. When the space disease identification result of a kind of space disease is greater than 0.5, it can be considered that the astronaut suffers from the space disease. For example, when the space disease identification result of muscle atrophy is greater than 0.5, it can be considered that the astronaut suffers from muscle atrophy. Thus, all the space diseases suffered by the astronaut can be determined according to the identification results of various categories of space diseases, that is, the space disease category information of the astronaut. For example, when the space disease identification results of muscle atrophy and orthostatic intolerance of the astronaut are 0.8 and 0.7, respectively, and the space disease identification results of the remaining categories are all less than 0.5, the space disease category information of the astronaut can be determined as suffering from muscle atrophy and orthostatic intolerance.

[0089] According to the embodiment of the present application, the space disease identification model corresponding to each category of space disease can be trained respectively. The training steps of the space disease identification model corresponding to the space disease category include: obtaining the sample physiological index comparison result of the training personnel and the sample judgment result of various symptoms corresponding to the space disease category; grouping the sample physiological index comparison result and the sample judgment result of various symptoms to form a sample input vector; inputting the sample input vector into the space disease identification model corresponding to the space disease category to obtain a predicted identification result; and training the space disease identification model corresponding to the space disease category according to the predicted identification result and the labeled information of the training personnel.

[0090] According to the embodiment of the present application, similar to the determination of the space disease identification result of the astronaut, a plurality of astronauts can be selected as training personnel, and the physiological index data of the training personnel and the judgment results of various symptoms corresponding to various space diseases can be obtained. These data can be stored in an external memory, without occupying the storage space of the memory of the space disease diagnosis and treatment intelligent controller, and are only used in training. The sample physiological index comparison result of the training personnel (similar to the physiological index comparison result, which is a value of 0 or 1) and the sample judgment result of various symptoms corresponding to various categories of space diseases (similar to the judgment result corresponding to the space disease category, which is a value of 0 or 1) are obtained, and the sample judgment results of various symptoms are grouped according to the space disease type. After that, the sample judgment results of the symptoms of each type of space disease and the sample physiological index comparison result are grouped to form a long vector corresponding to the space disease, that is, a sample input vector corresponding to the space disease. Inputting the long vector corresponding to the space disease into the space disease identification model corresponding to the space disease category can obtain a predicted identification result. The predicted identification result is a value in the range of [0, 1], which can describe the probability of the training personnel suffering from the corresponding space disease.

[0091] According to the embodiment of the present application, inputting the sample input vector into the space disease identification model corresponding to the space disease category to obtain a predicted identification result includes: obtaining the predicted identification result according to formula (1),

[0092]

[0093] wherein P(Y j | X j ) is the prediction recognition result of the jth training personnel corresponding to the Yth space disease category, X j = (x 1,j , x 2,j , …, x m,j ) is the sample input vector of the jth training personnel, x 1,j is the data of the first dimension of the sample input vector, x 2,j is the data of the second dimension of the sample input vector, x i,j is the data of the ith dimension of the sample input vector, x m,j is the data of the mth dimension of the sample input vector, ω0and ω i are parameters of the space disease recognition model corresponding to the Yth space disease category, i≤m, and i, j and m are positive integers.

[0094] According to an embodiment of the present application, in formula (1), ω0and ω i are parameters of the space disease recognition model corresponding to the Yth space disease category, and can obtain specific values after training by using the sample input vectors of multiple training personnel. x 1,j is the data of the first dimension of the sample input vector of the jth training personnel, and is 0 or 1, and similarly, x i,j is the data of the ith dimension of the sample input vector of the jth training personnel, and by inputting the data of each dimension of the sample input vector of the jth training personnel into the Logistic regression model (the space disease recognition model corresponding to the Yth space disease category), the probability of the jth training personnel suffering from the Yth space disease category can be obtained, that is, the prediction recognition result of the jth training personnel corresponding to the Yth space disease category.

[0095] According to an embodiment of the present application, the space disease recognition model corresponding to the space disease category is trained according to the prediction recognition result and the annotation information of the training personnel, comprising: determining the annotation recognition result of the Yth space disease of the training personnel according to the annotation information of the training personnel; determining the log-likelihood function according to the annotation recognition result and the prediction recognition result; and adjusting the parameters of the space disease recognition model corresponding to the space disease category according to the log-likelihood function, so as to train the space disease recognition model corresponding to the space disease category.

[0096] According to an embodiment of the present application, according to the annotation information of the training personnel, the annotation recognition result of the Yth space disease of the training personnel is determined. For example, the annotation recognition result of each space disease of the training personnel is 0 or 1, and the annotation recognition result is 0, indicating that the training personnel does not have the space disease, and the annotation recognition result is 1, indicating that the training personnel has the space disease.

[0097] According to an embodiment of the present application, according to the annotation recognition result and the prediction recognition result, the log-likelihood function is determined, including: determining the log-likelihood function L according to formula (2),

[0098]

[0099] wherein y j is the annotation recognition result of the Yth space disease of the jth training personnel, N is the number of training personnel, j≤N, and j and N are positive integers.

[0100] According to an embodiment of the present application, in formula (2), y j is the annotation recognition result of the Yth space disease of the jth training personnel, which is 0 or 1, and in the case of having the Yth space disease, the annotation recognition result of the Yth space disease is 1, and in the case of not having the Yth space disease, the annotation recognition result of the Yth space disease is 0, and the prediction recognition result is (the probability of the jth training personnel having the Yth space disease output by the space disease recognition model) or (the probability of the jth training personnel not having the Yth space disease output by the space disease recognition model), and then the log-likelihood function L based on the annotation recognition result and the prediction recognition result of the Yth space disease of the jth training personnel can be obtained After simplification, we can get Therefore, formula (2) can represent the sum of the log-likelihood functions of the Yth space disease of each training personnel, which can be used as the log-likelihood function L of the Yth space disease. During training, the log-likelihood function can be back-propagated to maximize the log-likelihood function, thereby improving the similarity between the annotation recognition result and the prediction recognition result of the Yth space disease.

[0101] According to an embodiment of the present application, according to the log-likelihood function, the parameters of the space disease recognition model corresponding to the space disease category are adjusted to train the space disease recognition model corresponding to the space disease category, including: obtaining the adjusted i th parameter of the space disease recognition model corresponding to the Yth space disease category according to formula (3)

[0102]

[0103] wherein L is the log-likelihood function, and a is the learning rate.j Let N be the labeling and identification result of the Yth type of space disease for the jth trainee, where N is the number of trainees, j≤N, and both j and N are positive integers.

[0104] According to an embodiment of the present invention, in formula (3), Indicates the relationship with ω i Find the partial derivative of the maximum likelihood function L for the Y-th type of space-related illness. Using this partial derivative, the gradient ascent method can be used to calculate ω. i Perform an update to obtain the adjusted version. α is the learning rate, which can control ω i The magnitude of each update. The initial value of the learning rate can be 0.1, and it gradually decreases as the number of training updates increases. Similarly, the adjusted parameters of the space disease identification model can be obtained. The sample input vectors of multiple trainees corresponding to the Yth type of space disease are divided into a training set, a validation set, and a test set according to a 6:2:2 ratio. The sample input vectors in the training set can be used to train the space disease identification model corresponding to the Yth type of space disease as described above. The validation set is used for further parameter tuning and model hyperparameter optimization. The prediction accuracy is tested in the test set, that is, the accuracy of determining whether each trainee in the test set has the Yth type of space disease is determined. If the accuracy is not up to standard, training is performed again until the accuracy meets the requirements, thus completing the training of the space disease identification model corresponding to the Yth type of space disease. Similarly, the space disease identification models corresponding to various types of space diseases can be trained using a similar method.

[0105] In this way, based on the Logistic regression model, the probability of each trainee suffering from various space-related illnesses can be determined, and the prediction accuracy is improved by comprehensively considering physiological indicators and astronaut symptoms. Furthermore, the parameters of the space-related illness identification model are updated according to the maximum likelihood function and gradient ascent method, thereby training the space-related illness identification model and improving its accuracy.

[0106] According to an embodiment of the present invention, the physiotherapy plan module can be used to determine a physiotherapy plan based on the space disease category information. Based on the astronaut's space disease category information, a physiotherapy plan corresponding to the astronaut's space disease is determined. For example, if an astronaut's space disease category information is muscular dystrophy and orthostatic hypotension, a physiotherapy plan corresponding to muscular dystrophy and orthostatic hypotension can be obtained. As described above, after determining the type of space disease suffered by the astronaut, corresponding acupoints can be selected, and the electrode pads corresponding to these acupoints can be controlled to output current pulses to stimulate the acupoints in a manner simulating acupuncture physiotherapy, thereby producing a physiotherapy effect.

[0107] In this way, the corresponding physiotherapy scheme of the space disease of the astronaut can be determined based on the space disease category information. The dependence on professional physicians is reduced, the accuracy and efficiency of screening and prediction of the space disease of the astronaut are improved, effective support is provided for further treatment decision, and the treatment effect of the space disease is improved.

[0108] Embodiment three:

[0109] Figure 2 An application schematic diagram of the space disease physiotherapy system for fusing multi-source heterogeneous biological information according to the embodiment of the application is exemplarily shown.

[0110] As shown in Figure 2 , the space disease diagnosis and treatment intelligent controller 102 can be electrically connected with the sleeping bag 101 of the astronaut through the U-shaped groove 103. The electrode pieces 104 corresponding to the positions of the multiple acupoints of the astronaut's body can be arranged in the sleeping bag. When the space disease diagnosis and treatment intelligent controller 102 is inserted into the U-shaped groove 103, the electrode pieces 104 can be controlled to output the current pulse, including selecting the electrode pieces that need to output the current pulse, and controlling the intensity and duration of the current pulse output by the electrode pieces. The space disease diagnosis and treatment intelligent controller 102 can determine the type of the space disease of the astronaut by realizing the functions of the above-mentioned broadcast module, the first acquisition module, the second acquisition module, the category determination module and the physiotherapy scheme module, and then retrieve the acupoint selection scheme retrieval table stored in the memory of the space disease diagnosis and treatment intelligent controller 102 to determine the acupoint selection scheme corresponding to the type of the space disease of the astronaut, that is, to determine which acupoints need to be acupunctured and physiotherapied, so as to control the electrode pieces 104 corresponding to the acupoints to output the current pulse through the space disease diagnosis and treatment intelligent controller 102, so as to stimulate the acupoints in the manner of simulating acupuncture and physiotherapy, thereby producing the effect of physiotherapy.

[0111] The space pathy therapy system fusing multi-source heterogeneous biological information according to the embodiment of the present application can broadcast space pathy diagnosis questions according to a preset symptom description list, collect astronaut's answer voice information and physiological index data, so as to determine the space pathy category information of the astronaut and determine a therapy scheme. The dependence on professional physicians is reduced, the accuracy and efficiency of screening and prediction of the space pathy of the astronaut are improved, effective support is provided for further determination of the therapy scheme, and the therapy effect of the space pathy is improved. When the space pathy diagnosis questions are determined, the space pathy diagnosis questions can be generated and broadcast based on the preset symptom description list, and astronaut's answer voice information and various physiological index data are collected. The astronaut's physical condition can be understood without the guidance of professional physicians, and basic data is provided for determination of the space pathy category information of the astronaut. When the space pathy recognition model is trained, the probability of each training personnel suffering from various space pathys is determined based on the Logistic regression model, and the physiological index and the symptom of the astronaut are comprehensively considered, so that the prediction accuracy is improved. Moreover, the parameters of the space pathy recognition model are updated according to the maximum likelihood function and the gradient ascent method, so that the space pathy recognition model is trained, and the accuracy of the space pathy recognition model is improved. When the therapy scheme is determined, the therapy scheme corresponding to the space pathy of the astronaut is determined based on the space pathy category information. The dependence on professional physicians is reduced, the accuracy and efficiency of screening and prediction of the space pathy of the astronaut are improved, effective support is provided for further therapy decision, and the therapy effect of the space pathy is improved.

[0112] Embodiment four:

[0113] Figure 3 An exemplary flowchart of a space pathy therapy method fusing multi-source heterogeneous biological information according to an embodiment of the present application is shown. The method comprises:

[0114] In step S1, space pathy diagnosis questions are generated and broadcast according to a preset symptom description list.

[0115] In step S2, answer voice information corresponding to the space pathy diagnosis questions is collected.

[0116] In step S3, various physiological index data of the astronaut is collected.

[0117] In step S4, space pathy category information of the astronaut is determined according to the answer voice information and the physiological index data.

[0118] In step S5, a therapy scheme is determined according to the space pathy category information.

[0119] The present application can be a method, device, system and / or computer program product. The computer program product can include a computer readable storage medium on which is loaded a computer readable program instruction for executing various aspects of the present application.

[0120] Those skilled in the art will understand that the above description and the embodiments of the present application shown in the drawings are only examples and do not limit the present application. The purpose of the present application has been fully and effectively achieved. The functional and structural principles of the present application have been demonstrated and described in the embodiments, and the embodiments of the present application can be modified or changed in any way without departing from the principles.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A space pathy therapy system that fuses multi-source heterogeneous biological information, characterized by, The method comprises the following steps: The announcement module is used to generate space disease diagnosis questions according to a preset symptom description list and to announce the space disease diagnosis questions. The first acquisition module is used to acquire answer voice information corresponding to the space disease diagnosis questions. The second acquisition module is used to acquire multiple physiological index data of astronauts. The category determination module is used to determine space disease category information of astronauts according to the answer voice information and the physiological index data. The physiotherapy scheme module is used to determine a physiotherapy scheme according to the space disease category information.

2. The system of claim 1, wherein, According to the preset symptom description list, space disease diagnosis questions are generated and announced, which comprises the following steps: According to the preset symptom description list, multiple symptom description texts are obtained. According to the symptom description texts, space disease diagnosis questions are obtained. The space disease diagnosis questions are announced through a sound equipment.

3. The system of claim 1, wherein, According to the answer voice information and the physiological index data, the space disease category information of astronauts is determined, which comprises the following steps: Through a voice recognition model, a voice recognition result of the answer voice information is obtained. According to the voice recognition result, a judgment result of a symptom corresponding to the space disease diagnosis question is determined. According to the physiological index data and a preset physiological index threshold value, a physiological index comparison result is obtained. According to the multiple space disease categories corresponding to the symptom types, the judgment result is grouped to obtain a judgment result set corresponding to the space disease categories. The physiological index comparison result and the judgment result set are input into a space disease recognition model corresponding to the space disease categories to obtain a space disease recognition result corresponding to the space disease categories. According to the space disease recognition results of the multiple categories, the space disease category information of astronauts is determined.

4. The system of claim 3, wherein, According to the voice recognition result, the judgment result of the symptom corresponding to the space disease diagnosis question is determined, which comprises the following steps: The voice recognition result is subjected to word segmentation processing to obtain multiple word segmentation texts. Through a pre-trained language model, word embedding processing is performed on each word segmentation text to obtain a word segmentation word vector of each word segmentation text. Through the pre-trained language model, word embedding processing is performed on the symptom description text corresponding to the space disease diagnosis question to obtain a symptom word vector of each symptom description text. The high-dimensional space distance of each word segmentation word vector and symptom word vector is determined. In the case that the high-dimensional space distance between the symptom word vector of the symptom description text and any one word segmentation word vector is less than or equal to a preset distance threshold value, the judgment result of the symptom corresponding to the symptom description text in the space disease diagnosis question is determined as existing.

5. The system of claim 3, wherein, The training steps of the space disease recognition model corresponding to the space disease categories comprise the following steps: Sample physiological index comparison results of training personnel and sample judgment results of multiple symptoms corresponding to the space disease categories are obtained. The sample physiological index comparison results and the sample judgment results of the multiple symptoms are combined to form a sample input vector. The sample input vector is input into the space disease recognition model corresponding to the space disease categories to obtain a predicted recognition result. According to the predicted recognition result and the labeling information of the training personnel, the space disease recognition model corresponding to the space disease categories is trained.

6. The system of claim 5, wherein, The sample input vector is input into the space disease recognition model corresponding to the space disease categories to obtain a predicted recognition result, which comprises the following steps: According to the formula obtaining a predicted recognition result, wherein P(Y j =1|X j ) is a predicted recognition result corresponding to the Yth space disease category of the jth training personnel, X j =(x 1,j ,x 2,j ,…,x m,j ) is a sample input vector of the jth training personnel, x 1,j is data of the first dimension of the sample input vector, x 2,j is data of the second dimension of the sample input vector, x i,j is data of the ith dimension of the sample input vector, x m,j is data of the mth dimension of the sample input vector, ω0 and ω i are parameters of a space disease recognition model corresponding to the Yth space disease category, i≤m, and i, j and m are positive integers.

7. The system of claim 6, wherein, According to the prediction recognition result and the annotation information of the training personnel, a space disease recognition model corresponding to the space disease category is trained, including: According to the annotation information of the training personnel, the annotation recognition result of the Yth space disease of the training personnel is determined; According to the annotation recognition result and the prediction recognition result, the log-likelihood function is determined; According to the log-likelihood function, the parameters of the space disease recognition model corresponding to the space disease category are adjusted to train the space disease recognition model corresponding to the space disease category.

8. The system of claim 7, wherein, According to the annotation recognition result and the prediction recognition result, the log-likelihood function is determined, including: According to the formula determining a log-likelihood function L, wherein y j is the labeled identification result of the jth training personnel for the Yth space disease, N is the number of training personnel, j≤N, and j and N are both positive integers.

9. The system of claim 7, wherein, According to the log-likelihood function, the parameters of the space disease recognition model corresponding to the space disease category are adjusted to train the space disease recognition model corresponding to the space disease category, including: According to the formula obtaining an adjusted i-th parameter of a space disease recognition model corresponding to the Y-th space disease category wherein, L is a log-likelihood function, a is a learning rate, y j is the labeled recognition result of the Y-th space disease of the j-th training personnel, N is the number of training personnel, j≤N, and j and N are both positive integers.

10. A space pathogenic therapy method of fusing multi-source heterogeneous biological information, characterized in that, Including: According to the preset symptom description list, a space disease diagnosis question is generated and broadcasted; Collecting answer voice information corresponding to the space disease diagnosis question; Collecting multiple physiological index data of astronauts; According to the answer voice information and the physiological index data, the space disease category information of the astronauts is determined; According to the space disease category information, a physiotherapy scheme is determined.

Images