Ankle rehabilitation exercise intelligent management system and method based on multi-modal data fusion

Abstract

The invention provides an ankle rehabilitation exercise intelligent management system and method based on multi-modal data fusion, and relates to the technical field of medical rehabilitation, and the method comprises the steps: collecting sole state data and electromyographic signal data of a patient end through a sensor network; determining a target gait mode and gait parameters according to the sole state data, and summarizing to obtain real-time gait data; determining muscle strength and muscle coordination degree according to the electromyographic signal data, and summarizing to obtain muscle state data; constructing an ankle motion function abnormity prediction model, and determining an ankle fault score according to the real-time gait data and the muscle state data; the ankle fault grade is judged according to the ankle fault score, and a personalized ankle rehabilitation strategy is made based on the ankle fault grade and sent to the visual display terminal, so that the ankle movement data can be accurately collected, a personalized rehabilitation exercise scheme is customized, the rehabilitation effect of a patient is improved in combination with an exercise auxiliary structure, and the rehabilitation effect of the patient is improved. The rapid recovery of the ankle function is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of medical rehabilitation, and particularly to an intelligent management system and method for ankle rehabilitation exercise based on multimodal data fusion. Background Art

[0002] Ankle injuries or diseases are relatively common in daily life. For example, stroke hemiplegia patients often have problems such as foot drop and varus foot, which seriously affect the walking function and quality of life of patients.

[0003] Traditional ankle rehabilitation methods cannot comprehensively collect the ankle movement data of patients, and thus may not be able to provide personalized rehabilitation programs and effective exercise assistance according to the real-time conditions of patients, making it difficult to meet the rehabilitation needs of different patients.

[0004] Therefore, it is necessary to provide an intelligent management system and method for ankle rehabilitation exercise based on multimodal data fusion to solve the above technical problems. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides an intelligent management system and method for ankle rehabilitation exercise based on multimodal data fusion, which is used to solve the problem that traditional ankle rehabilitation methods cannot comprehensively collect the ankle movement data of patients, and thus may not be able to provide personalized rehabilitation programs and effective exercise assistance according to the real-time conditions of patients, making it difficult to meet the rehabilitation needs of different patients.

[0006] The intelligent management system for ankle rehabilitation exercise based on multimodal data fusion provided by the present invention includes:

[0007] A sensor acquisition module, configured to acquire plantar state data and electromyogram signal data of the patient end through a sensor network;

[0008] A gait data acquisition module, configured to determine a target gait pattern and gait parameters according to the plantar state data, and aggregate to obtain real-time gait data;

[0009] A muscle state determination module, configured to determine muscle force intensity and muscle coordination degree according to the electromyogram signal data, and aggregate to obtain muscle state data;

[0010] A fault score prediction module, configured to construct an abnormal prediction model for ankle movement function, and determine an ankle fault score according to the real-time gait data and the muscle state data;

[0011] A rehabilitation strategy formulation module, configured to judge the ankle fault level according to the ankle fault score, and formulate a personalized ankle rehabilitation strategy based on the ankle fault level and send it to a visual display terminal.

[0012] Preferably, the sensor network includes a plantar status sensor and an electromyogram signal sensor;

[0013] Deploy a plurality of the plantar status sensors in different areas of the sole at the patient end, and the plantar status sensors are used to collect the plantar status data;

[0014] Deploy the electromyogram signal sensor on the surface of the muscle group at the patient end, and the electromyogram signal sensor is used to collect the electromyogram signal data;

[0015] Preprocess the plantar status data and the electromyogram signal data, and transmit the preprocessed plantar status data and electromyogram signal data to the microprocessor by wireless transmission.

[0016] Preferably, after receiving the plantar status data, the microprocessor constructs a gait pattern prediction model and predicts the target gait pattern according to the plantar status data, specifically including:

[0017] Input the plantar status data into the input layer of the gait pattern prediction model, and the input value of the u-th node in the hidden layer of the gait pattern prediction model is as follows:

[0018]

[0019] In the formula, H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; R represents the number of plantar status data in the input layer; w ru represents the weight from the r-th plantar status data in the input layer to the u-th node in the hidden layer; z r represents the r-th plantar status data in the input layer; b u represents the bias term of the u-th node in the hidden layer;

[0020] The corresponding output value of the u-th node in the hidden layer is as follows:

[0021]

[0022] In the formula, G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; e represents the natural constant;

[0023] The input value of the output layer of the gait pattern prediction model is as follows:

[0024]

[0025] In the formula, H vrepresents the input value of the v-th node in the output layer of the gait pattern prediction model; U represents the number of nodes in the hidden layer; w uv represents the weight from the u-th node in the hidden layer to the v-th node in the output layer; G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; b v represents the bias term of the v-th node in the output layer;

[0026] The output of the corresponding output layer is as follows:

[0027]

[0028] where G v represents the output value of the v-th node in the output layer of the gait pattern prediction model; H v represents the input value of the v-th node in the output layer of the gait pattern prediction model; e represents the natural constant;

[0029] The output value of the v-th node in the output layer is the predicted probability of the v-th gait pattern, and the gait pattern corresponding to the maximum predicted probability is the target gait pattern.

[0030] Preferably, calculating the gait parameters according to the plantar state data specifically includes:

[0031] The gait parameters include a step length parameter and a step frequency parameter;

[0032] Determine the heel-off time and the next heel-strike time of the patient end according to the plantar state data, and calculate the step length parameter in combination with the walking speed of the patient end as follows:

[0033] L = V × (t strike - t lift )

[0034] where L represents the step length parameter of the patient end; V represents the walking speed; t strike represents the heel-strike time; t lift represents the heel-off time;

[0035] Based on the plantar state data, determine the number of heel strikes N of the patient end within a preset time period T, and calculate the step frequency parameter as follows:

[0036]

[0037] where f represents the step frequency parameter of the patient end; N represents the number of heel strikes; T represents the preset time period;

[0038] Summarize the target gait pattern and the gait parameters to obtain the real-time gait data.

[0039] Preferably, based on the root mean square value algorithm, the muscle force intensity is calculated according to the electromyogram signal data as follows:

[0040]

[0041] In the formula, FLQD represents the amplitude of the electromyogram signal data at the patient end, that is, the muscle force intensity; I represents the number of sampling times; t i represents the i-th sampling time; EMG(t i ) represents the electromyogram signal data at the sampling time t i ;

[0042] Based on the Pearson correlation coefficient, the muscle coordination degree is determined according to the electromyogram signal data as follows:

[0043]

[0044] In the formula, XTCD represents the correlation degree between the anterior tibial muscle part and the related muscle part at the patient end, that is, the muscle coordination degree; X and Y respectively represent the electromyogram signal data sequences of the anterior tibial muscle part and the related muscle part; X j represents the electromyogram signal data sequence of the j-th anterior tibial muscle part; Y j represents the electromyogram signal data sequence of the j-th related muscle part; represents the average value of the electromyogram signal data sequence of the j-th anterior tibial muscle part; represents the average value of the electromyogram signal data sequence of the j-th related muscle part; J represents the number of sampling points;

[0045] The muscle force intensity and the muscle coordination degree are summarized to obtain the muscle state data.

[0046] Preferably, the real-time gait data and the muscle state data are normalized;

[0047] Based on the ankle movement function abnormality prediction model, the ankle fault score is calculated according to the normalized real-time gait data and the muscle state data, and the corresponding calculation formula is as follows:

[0048] GZFS = w SSBT × norm SSBT (SSBT) + w JRZT × norm JRZT (JRZT)

[0049] In the formula, GZFS represents the ankle fault score at the patient end; SSBT represents the real-time gait data; w SSBT represents the weight corresponding to the real-time gait data; norm SSBT( ) represents the normalization function corresponding to the real-time gait data; JRZT represents the muscle state data; w JRZT represents the weight corresponding to the muscle state data; norm JRZT ( ) represents the normalization function corresponding to the muscle state data.

[0050] Preferably, the ankle failure level is judged according to the ankle failure score, and the corresponding calculation formula is as follows:

[0051]

[0052] In the formula, GZDJ represents the ankle failure level at the patient end; GZFS represents the ankle failure score at the patient end; GZFS min1 and GZFS max1 represent the lower and upper limits of the ankle failure score corresponding to level 1; GZFS min2 and GZFS max2 represent the lower and upper limits of the ankle failure score corresponding to level 2; GZFS min3 and GZFS max3 represent the lower and upper limits of the ankle failure score corresponding to level 3;

[0053] Among them, the degree of ankle failure corresponding to the ankle failure level increases sequentially in the order of level 1, level 2, and level 3.

[0054] Preferably, the personalized ankle rehabilitation strategy is formulated based on the ankle failure level and sent to the visualization display terminal, specifically including:

[0055] Obtain the preset rehabilitation action library, and select the corresponding rehabilitation exercise actions according to the ankle failure level;

[0056] Calculate the rehabilitation exercise resistance according to the ankle failure level as follows:

[0057] XLZL = XLZL0 × k1

[0058] In the formula, XLZL represents the rehabilitation exercise resistance at the patient end; XLZL0 represents the basic resistance value; k1 represents the resistance adjustment coefficient corresponding to the ankle failure level;

[0059] Determine the rehabilitation exercise intensity at the patient end based on the rehabilitation exercise resistance;

[0060] Calculate the rehabilitation exercise time according to the ankle failure level as follows:

[0061] XLSJ = XLSJ0 × k2

[0062] Wherein, XLSJ represents the rehabilitation exercise time of the patient end; XLSJ0 represents the basic exercise time; k2 represents the time adjustment coefficient corresponding to the ankle failure level;

[0063] Summarize the rehabilitation exercise actions, the rehabilitation exercise intensity, and the rehabilitation exercise time to generate the personalized ankle rehabilitation strategy, and send the personalized ankle rehabilitation strategy to the visual display terminal.

[0064] Preferably, an adjustable elastic support structure and a detachable auxiliary rehabilitation accessory are provided, and according to the personalized ankle rehabilitation strategy, the elastic support strength and angle of the adjustable elastic support structure are adjusted.

[0065] An intelligent management method for ankle rehabilitation exercise based on multi-modal data fusion, the method includes:

[0066] Collect the plantar state data and electromyogram signal data of the patient end through a sensor network;

[0067] Determine the target gait pattern and gait parameters according to the plantar state data, and summarize to obtain real-time gait data;

[0068] Determine the muscle force intensity and muscle coordination degree according to the electromyogram signal data, and summarize to obtain muscle state data;

[0069] Construct an ankle movement function abnormality prediction model, and determine the ankle failure score according to the real-time gait data and the muscle state data;

[0070] Judge the ankle failure level according to the ankle failure score, and formulate a personalized ankle rehabilitation strategy based on the ankle failure level and send it to the visual display terminal.

[0071] Compared with the related technology, the intelligent management system and method for ankle rehabilitation exercise based on multi-modal data fusion provided by the present invention have the following beneficial effects:

[0072] The present invention collects the plantar state data and electromyogram signal data of the patient end through a sensor network; determines the target gait pattern and gait parameters according to the plantar state data, and summarizes to obtain real-time gait data; determines the muscle force intensity and muscle coordination degree according to the electromyogram signal data, and summarizes to obtain muscle state data; constructs an ankle movement function abnormality prediction model, and determines the ankle failure score according to the real-time gait data and the muscle state data; judges the ankle failure level according to the ankle failure score, and formulates a personalized ankle rehabilitation strategy based on the ankle failure level and sends it to the visual display terminal, so as to accurately collect ankle movement data, customize a personalized rehabilitation exercise plan, and combine with an exercise assistance structure to improve the patient's rehabilitation effect and promote the rapid recovery of ankle function.

[0073] The present invention can accurately collect and analyze the ankle movement data of patients, customize personalized rehabilitation exercise programs for patients, and improve the pertinence and effectiveness of rehabilitation; the system of the present invention can increase the flexibility and initiative of rehabilitation exercises, reduce the occurrence of incorrect exercise movements, and improve the rehabilitation effect of patients through an adjustable elastic support structure and a detachable auxiliary rehabilitation attachment; at the same time, the present invention can display personalized ankle rehabilitation strategies in real time through a visual display terminal, improve the efficiency and quality of patients' rehabilitation treatment, and promote the rapid recovery of ankle function. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 is a system block diagram of the intelligent management system for ankle rehabilitation exercises based on multi-modal data fusion of the present invention;

[0075] Figure 2 is a schematic diagram of the distribution of the plantar state sensors of the present invention;

[0076] Figure 3 is a flowchart of the intelligent management method for ankle rehabilitation exercises based on multi-modal data fusion of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0077] The present invention will be further described below with reference to the drawings and embodiments.

[0078] Embodiment 1

[0079] As Figure 1 shown, an intelligent management system for ankle rehabilitation exercises based on multi-modal data fusion, the system includes:

[0080] A sensor acquisition module, configured to acquire plantar state data and electromyogram signal data of a patient end through a sensor network;

[0081] Among them, the sensor network refers to a network composed of multiple sensors. In the intelligent management system for ankle rehabilitation exercises, these sensors cooperate with each other to collect the physiological data of patients. The plantar state data refers to the real-time state data of the patient's plantar. The electromyogram signal data refers to the bioelectric signals generated by the patient's muscles during contraction or relaxation, which can intuitively reflect the activity state of the muscles. The patient end refers to the patient who needs to perform ankle rehabilitation exercises.

[0082] It can be understood that the sensor acquisition module can, based on the sensor network, acquire the plantar state data and electromyogram signal data of the patient end in real time.

[0083] A gait data acquisition module, configured to determine a target gait pattern and gait parameters according to the plantar state data, and summarize to obtain real-time gait data;

[0084] Among them, the target gait pattern refers to the specific pattern of the contact between the patient's feet and the ground and the changes in body posture during walking. Different ankle health conditions will lead to different gait patterns. For example, the gait pattern of a patient with foot drop may be manifested as toe dragging on the ground and a reduced stride length when walking. Gait parameters refer to a series of indicators for measuring gait characteristics, including step length, step frequency, etc. These parameters can reflect the patient's movement state and the health condition of the ankles.

[0085] Through the target gait pattern and gait parameters, the gait situation of the patient can be comprehensively understood, real-time gait data can be generated, providing key support for subsequent evaluation and diagnosis.

[0086] The muscle state determination module is used to determine the muscle force intensity and muscle coordination degree according to the electromyogram signal data, and summarize to obtain muscle state data;

[0087] Among them, the muscle force intensity refers to the magnitude of the force generated by the patient's muscles during contraction, which is used to evaluate the functional state of the patient's muscles. The muscle coordination degree refers to the ability of different muscles of the patient to work together. The muscle state determination module can summarize the muscle force intensity and muscle coordination degree to obtain muscle state data.

[0088] In normal ankle movement, multiple muscles need to cooperate with each other to complete complex actions. For example, when walking, the anterior tibial muscle group and the posterior calf muscle group of the patient need to contract and relax synergistically to ensure the normal movement of the ankles.

[0089] The fault score prediction module is used to construct a prediction model for ankle movement function abnormality, and determine the ankle fault score according to the real-time gait data and the muscle state data;

[0090] Among them, the prediction model for ankle movement function abnormality is a model constructed based on a large amount of clinical data and machine learning algorithms, which is used to predict whether the ankle movement function is abnormal. The ankle fault score is an index used to quantify the degree of abnormality of the ankle movement function.

[0091] By constructing a prediction model for ankle movement function abnormality, the fault score prediction module can quantitatively evaluate the ankle health condition of the patient. Based on a large amount of clinical data and machine learning algorithms, this model can accurately analyze the correlation between real-time gait data and muscle state data, and determine the ankle fault score. And the higher the ankle fault score, the greater the possibility that the patient's ankle movement function is abnormal.

[0092] The rehabilitation strategy formulation module is used to judge the ankle fault level according to the ankle fault score, and formulate a personalized ankle rehabilitation strategy based on the ankle fault level and send it to the visualization display terminal.

[0093] In practical applications, the ankle fault levels are usually divided into different levels such as mild, moderate, and severe, and each level corresponds to different rehabilitation needs. The rehabilitation strategy formulation module can formulate personalized ankle rehabilitation strategies based on the ankle fault levels, combined with individual differences of the patient, such as age, physical condition, rehabilitation goals, etc. The strategy covers exercise movements, exercise intensity, exercise time, etc., and is presented to the patient and the physician through a visual display terminal. The visual display terminal can be a display screen, a computer software interface, etc., which is used to display the personalized ankle rehabilitation strategy.

[0094] In the specific implementation process, the sensor network includes a plantar status sensor and an electromyogram signal sensor;

[0095] As Figure 2 shown, a plurality of the plantar status sensors are deployed in different areas of the plantar of the patient end, and the plantar status sensors are used to collect the plantar status data;

[0096] The electromyogram signal sensor is deployed on the surface of the muscle group of the patient end, and the electromyogram signal sensor is used to collect the electromyogram signal data;

[0097] Preprocess the plantar status data and the electromyogram signal data, and transmit the preprocessed plantar status data and electromyogram signal data to the microprocessor by wireless transmission.

[0098] In practical applications, multiple plantar status sensors can be reasonably deployed in different areas such as the heel, arch, and forefoot of the patient's plantar to collect plantar status data. Whether the patient is in a static standing state or in a moving state such as walking or running, the force conditions such as the pressure distribution and pressure change on the ankle will be accurately captured by these sensors and converted into corresponding data information, providing a key basis for subsequent in-depth analysis of the mechanical state of the patient's ankle.

[0099] On the surface of the muscle group of the patient, such as the tibialis anterior muscle group on the front side of the calf and the gastrocnemius muscle on the back side, etc., the electromyogram signal sensor can be accurately deployed to efficiently collect the electromyogram signal data.

[0100] Since the collected plantar status data and electromyogram signal data may be affected by factors such as environmental interference, it is necessary to preprocess them. The preprocessing process covers professional operations such as denoising and filtering, and thus can improve the accuracy and reliability of the data.

[0101] Then, wireless transmission technologies such as Bluetooth and Wi-Fi can be used to quickly and stably transmit the preprocessed plantar status data and electromyogram signal data to the microprocessor.

[0102] After receiving the plantar state data, the microprocessor constructs a gait pattern prediction model and predicts the target gait pattern according to the plantar state data, specifically including:

[0103] Input the plantar state data into the input layer of the gait pattern prediction model. The input value of the u-th node in the hidden layer of the gait pattern prediction model is as follows:

[0104]

[0105] In the formula, H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; R represents the number of plantar state data in the input layer; w ru represents the weight from the r-th plantar state data in the input layer to the u-th node in the hidden layer; z r represents the r-th plantar state data in the input layer; b u represents the bias term of the u-th node in the hidden layer;

[0106] The output value of the corresponding u-th node in the hidden layer is as follows:

[0107]

[0108] In the formula, G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; e represents the natural constant;

[0109] The input value of the output layer of the gait pattern prediction model is as follows:

[0110]

[0111] In the formula, H v represents the input value of the v-th node in the output layer of the gait pattern prediction model; U represents the number of nodes in the hidden layer; w uv represents the weight from the u-th node in the hidden layer to the v-th node in the output layer; G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; b v represents the bias term of the v-th node in the output layer;

[0112] The corresponding output of the output layer is as follows:

[0113]

[0114] In the formula, G v represents the output value of the v-th node in the output layer of the gait pattern prediction model; H vDenote the input value of the v-th node in the output layer of the gait pattern prediction model; e denotes the natural constant;

[0115] The output value of the v-th node in the output layer is the predicted probability of the v-th gait pattern, and the gait pattern corresponding to the maximum predicted probability is determined as the target gait pattern.

[0116] Calculate the gait parameters according to the plantar state data, specifically including:

[0117] The gait parameters include a step length parameter and a step frequency parameter;

[0118] Determine the heel-off time and the next heel-strike time of the patient side according to the plantar state data, and combine the walking speed of the patient side to calculate the step length parameter as follows:

[0119] L = V × (t strike - t lift )

[0120] In the formula, L represents the step length parameter of the patient side; V represents the walking speed; t strike represents the heel-strike time; t lift represents the heel-off time;

[0121] Based on the plantar state data, determine the number of heel strikes N of the patient side within a preset time period T, and calculate the step frequency parameter as follows:

[0122]

[0123] In the formula, f represents the step frequency parameter of the patient side; N represents the number of heel strikes; T represents the preset time period;

[0124] Summarize the target gait pattern and the gait parameters to obtain the real-time gait data.

[0125] It should be noted that first, based on the plantar state data, the moment when the patient's heel leaves the ground and the moment of the next heel strike can be accurately determined, and then combined with the real-time walking speed of the patient side, the step length parameter can be obtained. Then, by counting the number of heel strikes N of the patient side within the preset duration T, the step frequency parameter can be calculated. Finally, the target gait pattern and the gait parameters can be summarized to obtain the real-time gait data.

[0126] Based on the root mean square value algorithm, calculate the muscle force intensity according to the electromyogram signal data as follows:

[0127]

[0128] Wherein, FLQD represents the amplitude of the electromyogram signal data at the patient end, that is, the muscle force intensity; I represents the number of sampling moments; t i represents the i-th sampling moment; EMG(t i ) represents the electromyogram signal data at the sampling moment t i ;

[0129] Based on the Pearson correlation coefficient, the muscle coordination degree is determined according to the electromyogram signal data as follows:

[0130]

[0131] Wherein, XTCD represents the correlation degree between the tibialis anterior muscle part and the related muscle part at the patient end, that is, the muscle coordination degree; X and Y respectively represent the electromyogram signal data sequences of the tibialis anterior muscle part and the related muscle part; X j represents the electromyogram signal data sequence of the j-th tibialis anterior muscle part; Y j represents the electromyogram signal data sequence of the j-th related muscle part; represents the average value of the electromyogram signal data sequence of the j-th tibialis anterior muscle part; represents the average value of the electromyogram signal data sequence of the j-th related muscle part; J represents the number of sampling points;

[0132] The muscle force intensity and the muscle coordination degree are summarized to obtain the muscle state data.

[0133] Among them, when performing in-depth analysis on the electromyogram signal data, the root mean square value algorithm can be used to accurately calculate the amplitude of the electromyogram signal data, that is, the muscle force intensity.

[0134] At the same time, based on the Pearson correlation coefficient, the correlation between the electromyogram signal data sequences of the patient's tibialis anterior muscle part and the related muscle part can be deeply analyzed, and then the muscle coordination degree can be determined.

[0135] Finally, the calculated muscle force intensity and muscle coordination degree can be summarized and integrated to obtain muscle state data, which is used to comprehensively reflect the real-time state of the patient's muscles.

[0136] Normalize the real-time gait data and the muscle state data;

[0137] Based on the ankle movement function abnormality prediction model, calculate the ankle fault score according to the normalized real-time gait data and muscle state data, and the corresponding calculation formula is as follows:

[0138] GZFS = w SSBT ×norm SSBT (SSBT)+w JRZT ×normJRZT (JRZT)

[0139] Wherein, GZFS represents the ankle fault score at the patient end; SSBT represents the real-time gait data; w SSBT represents the weight corresponding to the real-time gait data; norm SSBT () represents the normalization function corresponding to the real-time gait data; JRZT represents the muscle state data; w JRZT represents the weight corresponding to the muscle state data; norm JRZT () represents the normalization function corresponding to the muscle state data.

[0140] Judge the ankle fault level according to the ankle fault score, and the corresponding calculation formula is as follows:

[0141]

[0142] Wherein, GZDJ represents the ankle fault level at the patient end; GZFS represents the ankle fault score at the patient end; GZFS min1 and GZFS max1 represent the lower limit and upper limit of the ankle fault score corresponding to level 1; GZFS min2 and GZFS max2 represent the lower limit and upper limit of the ankle fault score corresponding to level 2; GZFS min3 and GZFS max3 represent the lower limit and upper limit of the ankle fault score corresponding to level 3;

[0143] Among them, the degree of ankle fault corresponding to the ankle fault level increases in sequence according to level 1, level 2, and level 3.

[0144] Formulate the personalized ankle rehabilitation strategy based on the ankle fault level and send it to the visual display terminal, specifically including:

[0145] Obtain the preset rehabilitation action library, and select the corresponding rehabilitation exercise actions according to the ankle fault level;

[0146] Calculate the rehabilitation exercise resistance according to the ankle fault level as follows:

[0147] XLZL = XLZL0 × k1

[0148] Wherein, XLZL represents the rehabilitation exercise resistance at the patient end; XLZL0 represents the basic resistance value; k1 represents the resistance adjustment coefficient corresponding to the ankle fault level;

[0149] Determine the rehabilitation exercise intensity at the patient end based on the rehabilitation exercise resistance;

[0150] Calculate the rehabilitation exercise time according to the ankle fault level as follows:

[0151] XLSJ = XLSJ0 × k2

[0152] In the formula, XLSJ represents the rehabilitation exercise time at the patient end; XLSJ0 represents the basic exercise time; k2 represents the time adjustment coefficient corresponding to the ankle failure level.

[0153] Summarize the rehabilitation exercise actions, the rehabilitation exercise intensity, and the rehabilitation exercise time to generate the personalized ankle rehabilitation strategy, and send the personalized ankle rehabilitation strategy to the visual display terminal.

[0154] It can be understood that in the intelligent management system for ankle rehabilitation exercise, it is first necessary to normalize the real-time gait data and muscle state data, so that data with different magnitudes and different characteristics can be unified into the same scale range, eliminate the dimensional differences between the data, and improve the accuracy and stability of subsequent analysis and calculation.

[0155] Furthermore, based on the ankle movement function abnormality prediction model, according to the normalized real-time gait data and muscle state data, the ankle failure score can be accurately calculated. Then, the ankle failure score can be compared with the score intervals corresponding to different levels, and further determine the ankle failure level of the patient. And the ankle failure level ranges from level 1 to level 3, and the corresponding ankle failure degree increases in turn.

[0156] Immediately afterwards, a personalized ankle rehabilitation strategy can be formulated based on the ankle failure level. Specifically, the rehabilitation exercise actions suitable for the current ankle failure level can be accurately selected from the preset rehabilitation action library. Then, according to the ankle failure level, combined with the basic resistance value and the resistance adjustment coefficient corresponding to the level, the rehabilitation exercise resistance can be calculated, and then the rehabilitation exercise intensity can be determined. At the same time, according to the basic exercise time and the corresponding time adjustment coefficient, the rehabilitation exercise time can be calculated.

[0157] Finally, the rehabilitation exercise actions, the rehabilitation exercise intensity, and the rehabilitation exercise time can be summarized to generate a complete personalized ankle rehabilitation strategy and sent to the visual display terminal in a timely manner.

[0158] Set an adjustable elastic support structure and a detachable auxiliary rehabilitation attachment, and adjust the elastic support strength and angle of the adjustable elastic support structure according to the personalized ankle rehabilitation strategy.

[0159] In addition, the intelligent management system for ankle rehabilitation exercise is also provided with an adjustable elastic support structure and a detachable auxiliary rehabilitation attachment. The adjustable elastic support structure adopts advanced materials and mechanical designs, and has the function of accurately adjusting the elastic support strength and angle. The detachable auxiliary rehabilitation attachment can enrich the forms and dimensions of rehabilitation exercises.

[0160] In practical applications, based on personalized ankle rehabilitation strategies and combined with the specific ankle fault levels and rehabilitation needs of patients, the elastic support strength and angle of the adjustable elastic support structure can be precisely adjusted to provide the most suitable rehabilitation assistance.

[0161] Embodiment 2

[0162] As Figure 3 shown, an intelligent management method for ankle rehabilitation exercises based on multi-modal data fusion, the method includes:

[0163] S1, collecting plantar state data and electromyogram signal data at the patient end through a sensor network;

[0164] S2, determining a target gait pattern and gait parameters according to the plantar state data, and summarizing to obtain real-time gait data;

[0165] S3, determining muscle force intensity and muscle coordination degree according to the electromyogram signal data, and summarizing to obtain muscle state data;

[0166] S4, constructing an ankle movement function abnormality prediction model, and determining an ankle fault score according to the real-time gait data and the muscle state data;

[0167] S5, judging the ankle fault level according to the ankle fault score, and formulating a personalized ankle rehabilitation strategy based on the ankle fault level and sending it to a visual display terminal.

[0168] Through the introduction of the above embodiments, the present invention can collect plantar state data and electromyogram signal data at the patient end through a sensor network through an intelligent management system and method for ankle rehabilitation exercises based on multi-modal data fusion; determine a target gait pattern and gait parameters according to the plantar state data, and summarize to obtain real-time gait data; determine muscle force intensity and muscle coordination degree according to the electromyogram signal data, and summarize to obtain muscle state data; construct an ankle movement function abnormality prediction model, and determine an ankle fault score according to the real-time gait data and the muscle state data; judge the ankle fault level according to the ankle fault score, and formulate a personalized ankle rehabilitation strategy based on the ankle fault level and send it to a visual display terminal, so that the ankle movement data can be accurately collected, a personalized rehabilitation exercise plan can be customized, and combined with the exercise assistance structure, the rehabilitation effect of patients can be improved, and the rapid recovery of ankle function can be promoted.

[0169] The present invention can accurately collect and analyze the ankle movement data of patients, customize personalized rehabilitation exercise programs for patients, and improve the pertinence and effectiveness of rehabilitation; the system of the present invention can increase the flexibility and initiative of rehabilitation exercises, reduce the occurrence of incorrect exercise movements, and improve the rehabilitation effect of patients through an adjustable elastic support structure and a detachable auxiliary rehabilitation attachment; at the same time, the present invention can display personalized ankle rehabilitation strategies in real time through a visual display terminal, improve the efficiency and quality of patients' rehabilitation treatment, and promote the rapid recovery of ankle function.

[0170] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0171] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program, and this program can be stored in a computer-readable storage medium. The storage medium includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM), or other optical disc memories, magnetic disk memories, tape memories, or any other computer-readable medium capable of carrying or storing data.

[0172] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent in such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, commodity or device comprising the element.

Claims

1. An intelligent management system for foot and ankle rehabilitation based on multimodal data fusion, characterized in that: The system comprises: A sensor acquisition module is used to collect the plantar status data and electromyographic signal data of the patient through a sensor network; A gait data acquisition module, used to determine a target gait pattern and gait parameters according to the plantar state data, and summarize to obtain real-time gait data; A muscle state determination module, used to determine the muscle force intensity and muscle coordination degree according to the electromyographic signal data, and summarize to obtain muscle state data; A fault score prediction module, used to construct a foot and ankle motor function abnormality prediction model and determine the foot and ankle fault score according to the real-time gait data and the muscle state data; A rehabilitation strategy formulation module is used to determine the level of the ankle fault according to the ankle fault score, and formulate a personalized ankle rehabilitation strategy based on the ankle fault level and send it to the visual display terminal.

2. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 1, characterized in that: The sensor network includes a foot state sensor and an electromyographic signal sensor; Deploy a plurality of the plantar state sensors at different areas of the plantar of the patient, wherein the plantar state sensors are used to collect the plantar state data; The electromyographic signal sensor is deployed on the surface of the muscle group at the patient end, and the electromyographic signal sensor is used to collect the electromyographic signal data; The plantar state data and the electromyographic signal data are preprocessed, and the preprocessed plantar state data and the electromyographic signal data are transmitted to a microprocessor via wireless transmission.

3. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 2 is characterized in that: After receiving the plantar state data, the microprocessor constructs a gait pattern prediction model and predicts the target gait pattern according to the plantar state data, specifically including: The plantar state data is input into the input layer of the gait pattern prediction model, and the input value of the u-th node in the hidden layer of the gait pattern prediction model is as follows: In the formula, H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; R represents the number of plantar state data in the input layer; w ru represents the weight from the rth plantar state data in the input layer to the uth node in the hidden layer; z r represents the rth plantar state data in the input layer; b u represents the bias term of the u-th node in the hidden layer; The corresponding output value of the u-th node in the hidden layer is as follows: In the formula, G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; H u represents the input value of the u-th node in the hidden layer of the gait pattern prediction model; e represents a natural constant; The input values ​​of the output layer of the gait pattern prediction model are as follows: In the formula, H v represents the input value of the vth node in the output layer of the gait pattern prediction model; U represents the number of nodes in the hidden layer; w uv represents the weight from the uth node in the hidden layer to the vth node in the output layer; G u represents the output value of the u-th node in the hidden layer of the gait pattern prediction model; b v represents the bias term of the vth node in the output layer; The corresponding output of the output layer is as follows: In the formula, G v represents the output value of the vth node in the output layer of the gait pattern prediction model; H v represents the input value of the vth node in the output layer of the gait pattern prediction model; e represents a natural constant; The output value of the vth node in the output layer is the predicted probability of the vth gait pattern, and the gait pattern corresponding to the maximum predicted probability is determined to be the target gait pattern.

4. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 3 is characterized in that: Calculating the gait parameter according to the plantar state data specifically includes: The gait parameters include a step length parameter and a step frequency parameter; The heel-off time and the next heel-on time of the patient are determined according to the plantar state data, and the step length parameter is calculated in combination with the walking speed of the patient as follows: L=V×(t strike -t lift ) Where L represents the step length parameter of the patient; V represents the walking speed; t strike Indicates the heel strike time; t lift Indicates the heel-off time; Based on the plantar state data, the number of times N when the patient's heel strikes the ground within a preset time period T is determined, and the step frequency parameter is calculated as follows: Where, f represents the patient's step frequency parameter; N represents the number of heel strikes; T represents the preset time period; The target gait pattern and the gait parameters are aggregated to obtain the real-time gait data.

5. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 1, characterized in that: Based on the root mean square value algorithm, the muscle force intensity is calculated according to the electromyographic signal data as follows: Where FLQD represents the amplitude of the electromyographic signal data at the patient end, that is, the muscle strength; I represents the number of sampling moments; t i represents the i-th sampling moment; EMG(t i ) represents the sampling time t i EMG signal data; Based on the Pearson correlation coefficient, the muscle coordination degree is determined according to the electromyographic signal data as follows: Where, XTCD represents the correlation between the tibialis anterior muscle and the related muscle parts at the patient end, that is, the muscle coordination degree; X and Y represent the electromyographic signal data sequences of the tibialis anterior muscle and the related muscle parts respectively; X j Y represents the electromyographic signal data sequence of the jth tibialis anterior muscle; j Represents the electromyographic signal data sequence of the jth related muscle part; represents the average value of the electromyographic signal data sequence of the jth tibialis anterior muscle; represents the average value of the electromyographic signal data sequence of the jth related muscle part; J represents the number of sampling points; The muscle strength and the muscle coordination degree are summarized to obtain the muscle status data.

6. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 1, characterized in that: Normalizing the real-time gait data and the muscle state data; Based on the ankle motor function abnormality prediction model, the ankle fault score is calculated according to the normalized real-time gait data and the muscle state data. The corresponding calculation formula is as follows: GZFS=w SSBT ×norm SSBT SSBT+in JRZT ×norm JRZT JRZT Where GZFS represents the foot and ankle failure score at the patient end; SSBT means real-time gait data; w SSBT Indicates the weight corresponding to the real-time gait data; norm SSBT () represents the normalization function corresponding to the real-time gait data; JRZT represents the muscle state data; w JRZT Indicates the weight corresponding to the muscle state data; norm JRZT Represents the normalization function corresponding to the muscle state data.

7. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 6, characterized in that: The ankle fault level is determined according to the ankle fault score, and the corresponding calculation formula is as follows: Where GZDJ represents the ankle fault grade at the patient end; GZFS represents the ankle fault score at the patient end; GZFS min1 and GZFS max1 Indicates the lower and upper limits of the ankle failure score corresponding to level 1; GZFS min2 and GZFS max2 Indicates the lower and upper limits of the ankle failure score corresponding to level 2; GZFS min3 and GZFS max3 represents the lower and upper limits of the ankle failure score corresponding to level 3; Among them, the degree of ankle fault corresponding to the ankle fault level increases in the order of level 1, level 2, and level 3.

8. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 7, characterized in that: Formulating the personalized ankle rehabilitation strategy based on the ankle fault level and sending it to the visual display terminal specifically includes: Obtaining a preset rehabilitation action library, and selecting corresponding rehabilitation exercise actions according to the ankle fault level; The rehabilitation exercise resistance is calculated according to the ankle fault level as follows: XLZL=XLZL0×k1 Where XLZL represents the rehabilitation exercise resistance at the patient end; XLZL0 represents the basic resistance value; k1 represents the resistance adjustment coefficient corresponding to the ankle fault level; Determining the rehabilitation exercise intensity of the patient based on the rehabilitation exercise resistance; The rehabilitation training time is calculated according to the ankle fault level as follows: XLSJ=XLSJ0×k2 In the formula, XLSJ represents the rehabilitation exercise time on the patient side; XLSJ0 represents the basic exercise time; k2 represents the time adjustment coefficient corresponding to the ankle fault level; The rehabilitation exercise movements, the rehabilitation exercise intensity and the rehabilitation exercise time are summarized to generate the personalized foot and ankle rehabilitation strategy, and the personalized foot and ankle rehabilitation strategy is sent to the visual display terminal.

9. The intelligent management system for foot and ankle rehabilitation based on multimodal data fusion according to claim 1, characterized in that: An adjustable elastic support structure and a detachable auxiliary rehabilitation accessory are provided, and the elastic support strength and angle of the adjustable elastic support structure are adjusted according to the personalized foot and ankle rehabilitation strategy.

10. An intelligent management method for foot and ankle rehabilitation based on multimodal data fusion, applied to an intelligent management system for foot and ankle rehabilitation based on multimodal data fusion as claimed in any one of claims 1 to 9, the method comprising: Collect the plantar status data and electromyographic signal data of the patient through the sensor network; Determine a target gait pattern and gait parameters according to the plantar state data, and summarize to obtain real-time gait data; Determine muscle strength and muscle coordination according to the electromyographic signal data, and summarize to obtain muscle status data; Constructing a prediction model for abnormal ankle motor function, and determining an ankle fault score based on the real-time gait data and the muscle status data; The ankle fault level is determined according to the ankle fault score, and a personalized ankle rehabilitation strategy is formulated based on the ankle fault level and sent to a visualization display terminal.

Images