Gait correction system

Abstract

The application provides a gait correction system, comprising a processing device and an assisting device, the processing device is used for: acquiring gait data of a correction object; performing dimension reduction on the gait data by using isometric mapping to obtain a low-dimensional point set of the gait data; processing the low-dimensional point set of the gait data by using a kernel density estimation method to obtain a walking gait distribution curve; determining an assisting parameter according to the walking gait distribution curve; and the assisting device is used for assisting the correction object to walk according to the assisting parameter. Thus, the walking gait distribution curve of the motion trajectory of the correction object is constructed by using the isometric mapping and the kernel density estimation method, the symmetry correction of the walking gait of the correction object in the multi-dimensional index is realized by the assisting device without losing the high-dimensional information, and thus the gait correction effect is improved.

Description

Technical Field

[0001] This application belongs to the field of medical device technology, and in particular relates to a gait correction system. Background Technology

[0002] Stroke is one of the leading causes of limb disability worldwide. Most stroke patients experience neurogenic hemiplegic gait, characterized by gait asymmetry during walking. This gait asymmetry limits walking speed, increases energy expenditure, impairs balance, and leads to compensatory movements in the unaffected limbs, increasing the load on the affected side, raising the risk of musculoskeletal injury, and reducing quality of life. Therefore, correcting gait asymmetry is of significant clinical value in the rehabilitation of stroke patients.

[0003] In related technologies, to correct gait asymmetry in stroke patients, gait data is collected and optimization indicators are set. These optimization indicators are then used to update the assistive device's parameters, which in turn assists the stroke patient in walking. However, due to the diversity of gait asymmetry assessment indicators, correcting only a single indicator after feature extraction is insufficient. How to improve the effectiveness of gait correction remains a pressing issue. Summary of the Invention

[0004] This application provides a gait correction system that can improve the effectiveness of gait correction.

[0005] A first aspect of this application provides a gait correction system, including a processing device and an assisting device. The processing device is used to: acquire gait data of the correction subject; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation to obtain a walking gait distribution curve; and determine assisting parameters based on the walking gait distribution curve. The assisting device is used to assist the correction subject in walking according to the assisting parameters.

[0006] Optionally, in one possible implementation of the first aspect, the above-mentioned correction target is a stroke patient, and the correction target includes the affected limb and the unaffected limb.

[0007] Optionally, in another possible implementation of the first aspect, the dimensionality reduction of the gait data using isometric mapping to obtain a low-dimensional point set of the gait data includes:

[0008] Noise reduction of gait data is achieved using a low-pass filter;

[0009] For the denoised gait data, multiple gait cycles are identified using an adaptive thresholding method.

[0010] Based on multiple gait cycles, multiple gait cycle samples of the affected side and multiple gait cycle samples of the healthy side of the subject were determined;

[0011] We used isometric mapping to reduce the dimensionality of multiple gait cycle samples from the affected side and multiple gait cycle samples from the healthy side, respectively, to obtain a low-dimensional point set of gait data.

[0012] Optionally, in another possible implementation of the first aspect, the aforementioned low-dimensional point set of gait data includes a low-dimensional point set of gait data on the affected side and a low-dimensional point set of gait data on the healthy side. The low-dimensional point set of gait data on the affected side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the affected side, and the low-dimensional point set of gait data on the healthy side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the healthy side.

[0013] Optionally, in another possible implementation of the first aspect, the aforementioned low-pass filter is a fourth-order Butterworth low-pass filter.

[0014] Optionally, in another possible implementation of the first aspect, the aforementioned gait distribution curve includes the gait distribution curve of the affected side and the gait distribution curve of the unaffected side, and the determination of the assist parameters based on the gait distribution curve includes:

[0015] Determine the Wasserstein distance between the gait distribution curves of the affected side and the healthy side, whereby the Wasserstein distance is used to characterize the gait difference between the affected and healthy limbs of the subject undergoing correction.

[0016] The current boost parameters are updated based on the Wasserstein distance to obtain the updated boost parameters.

[0017] Optionally, in another possible implementation of the first aspect, the updated boosting parameters, obtained by updating the current boosting parameters based on the Wasserstein distance, include:

[0018] Based on the Wasserstein distance, the current assist parameters are continuously updated using a Bayesian optimization algorithm to obtain the updated assist parameters.

[0019] Optionally, in another possible implementation of the first aspect, an inertial measurement unit is set on both the affected limb and the unaffected limb of the subject being corrected to acquire gait data of the subject being corrected, including:

[0020] Gait data is acquired by using inertial measurement units set on the affected and healthy limbs of the patient undergoing correction.

[0021] Optionally, in another possible implementation of the first aspect, the aforementioned assistive device is an exoskeleton.

[0022] A second aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the following: acquiring gait data of a correction subject; reducing the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; processing the low-dimensional point set of gait data using a kernel density estimation method to obtain a walking gait distribution curve; determining assistance parameters based on the walking gait distribution curve; and using an assistance device to assist the correction subject in walking according to the assistance parameters.

[0023] A third aspect of this application provides a computer program product that, when run on a gait correction system, causes the gait correction system to perform: acquiring gait data of the correction object;

[0024] Dimensionality reduction of gait data is achieved by using isometric mapping to obtain a low-dimensional point set of gait data;

[0025] The kernel density estimation method is used to process the low-dimensional point set of gait data to obtain the walking gait distribution curve;

[0026] The assist parameters are determined based on the walking gait distribution curve.

[0027] The beneficial effects of this application embodiment compared with the prior art are as follows: This application embodiment provides a gait correction system, including a processing device and an assisting device. The processing device is used to: acquire gait data of the correction object; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation method to obtain a walking gait distribution curve; determine assisting parameters based on the walking gait distribution curve; and the assisting device is used to assist the correction object in walking according to the assisting parameters. Therefore, by constructing the walking gait distribution curve of the correction object's movement trajectory using isometric mapping and kernel density estimation method, the assisting device achieves symmetrical correction of the correction object's walking gait under multi-dimensional indicators without losing high-dimensional information, thereby improving the gait correction effect. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram of the structure of a gait correction system provided in an embodiment of this application;

[0030] Figure 2This is a flowchart illustrating the execution steps of a processing device in a gait correction system provided in an embodiment of this application;

[0031] Figure 3 This is a schematic diagram of a scenario executed by a processing device in a gait correction system provided in an embodiment of this application. Detailed Implementation

[0032] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0033] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0034] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0035] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0036] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0037] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0038] It should be understood that the sequence number of each step in this embodiment does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application embodiment.

[0039] In related technologies, to correct gait asymmetry in stroke patients, gait data is collected and optimization indicators are set. These optimization indicators are then used to update the assistive device's parameters, which in turn assists the stroke patient in walking. However, due to the diversity of gait asymmetry assessment indicators, correcting only a single indicator after feature extraction is insufficient. How to improve the effectiveness of gait correction remains a pressing issue.

[0040] In view of this, embodiments of this application provide a gait correction system, including a processing device and an assisting device. The processing device is used to: acquire gait data of the correction subject; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation to obtain a walking gait distribution curve; determine assisting parameters based on the walking gait distribution curve; and the assisting device is used to assist the correction subject in walking according to the assisting parameters. Thus, by constructing the walking gait distribution curve of the correction subject's movement trajectory using isometric mapping and kernel density estimation, the assisting device achieves symmetrical correction of the correction subject's walking gait under multi-dimensional indicators without losing high-dimensional information, thereby improving the gait correction effect.

[0041] To illustrate the technical solution of this application, specific embodiments are described below.

[0042] See Figure 1 The diagram shows a structural schematic of a gait correction system provided in an embodiment of this application. For ease of explanation, only the parts related to the embodiment of this application are shown.

[0043] The gait correction system includes a processing device 10 and an assist device 11. The processing device 10 may specifically include a data acquisition module 100, an evaluation index calculation module 101, and an assist parameter optimization module 102.

[0044] The following sections will introduce the various modules in the processing device 10 and the assist device 11.

[0045] The data acquisition module 100 is used to acquire the gait data of the object being corrected.

[0046] In this embodiment, the correction target can be a stroke patient. The correction target includes the affected limb and the unaffected limb. Gait data can include walking motion data and hip joint angle data of the correction target over a period of time.

[0047] In one embodiment, an inertial measurement unit (IMU) can be placed on the affected limb and the unaffected limb of the subject for correction, respectively, to acquire gait data. Specifically, the data acquisition module 101 can acquire the gait data of the subject for correction through two IMUs placed on the thighs of the subject.

[0048] In one possible implementation, the processing device 10 may further include a data preprocessing module 104, which is used to reduce noise in gait data using a low-pass filter; identify multiple gait cycles using an adaptive threshold method for the denoised gait data; and determine multiple affected-side gait cycle samples and multiple healthy-side gait cycle samples of the correction object based on the multiple gait cycles.

[0049] It should be noted that the data preprocessing module 104 can use low-pass filtering to reduce noise in the gait data of the subject being corrected, then use an adaptive thresholding method to identify the start and end points of each gait cycle, and divide the gait cycles accordingly. Further, through data integration, gait cycle samples of both limbs of the subject being corrected are obtained. The low-pass filter can be a fourth-order Butterworth low-pass filter, which is a maximally flat amplitude response filter, possessing the smoothest frequency response (ripple-free) in the passband, but with slower attenuation in the stopband (compared to Chebyshev or elliptic filters). The fourth-order Butterworth low-pass filter is composed of two cascaded second-order Butterworth filters, providing a steeper roll-off characteristic.

[0050] Specifically, you can refer to, for example Figure 2 As shown in the flowchart, the data preprocessing module 104 first performs data filtering, applying a fourth-order Butterworth low-pass filter to the original hip joint angle data of the object being corrected. Its transfer function can be expressed by the following formula:

[0051]

[0052] Where s is a complex variable in the Laplace domain, f c is the cutoff frequency, and n is the filter order.

[0053] Gait is further segmented using heel-off, and an adaptive thresholding method is used to identify the peak hip extension angle. The threshold update formula is shown below:

[0054] T k =μ k-1 +α·σ k-1 ;

[0055] Where, μ k-1 and σ k-1 These are the mean and standard deviation of the angles for the first k-1 gait cycles, respectively, and α is an empirical coefficient with a value of 2.5. The start and end points of each gait cycle are identified, and gait cycles are divided accordingly. Abnormal gait cycles with a length less than 0.5s or greater than 2s are removed.

[0056] After the above processing, the data is integrated to obtain N gait cycle samples for both the affected and healthy limbs. Each sample contains 1-dimensional angle data (hip flexion-extension sagittal plane joint angle), which can be expressed by the following formula:

[0057] X = [x1, x2, ..., x M ] T ∈R M×D = [x1,x2,…,x M ] T ∈R M ;

[0058] Where M = 2N is the total number of samples.

[0059] The evaluation index calculation module 101 is used to reduce the dimensionality of gait data using isometric mapping (Isomap) to obtain a low-dimensional point set of gait data.

[0060] In one embodiment, the evaluation index calculation module 101 uses isometric mapping to reduce the dimensionality of multiple affected-side gait cycle samples and multiple healthy-side gait cycle samples respectively, to obtain a low-dimensional point set of gait data.

[0061] Specifically, the evaluation index calculation module 101 uses isometric mapping to perform nonlinear dimensionality reduction on the hip joint angle trajectory between every two heel-off moments of the affected and unaffected limbs of the corrected subject. (See example...) Figure 2 The flowchart shown first constructs a neighborhood graph and uses k-nearest neighbors (k-NN) to determine each sample point x.j neighborhood N i Then, the adjacency weight matrix W is calculated, as shown in the following formula:

[0062]

[0063] Then, the geodesic distance matrix is ​​calculated, and the shortest path distance d between all sample pairs is calculated using Dijkstra's algorithm. G (x i ,x j This forms the geodesic distance matrix D. G =ShortestPath(W)∈R M×M .

[0064] Then, multidimensional scaling (MDS) is performed to reduce the dimensionality and construct the central matrix. Where 1 represents an all-1 vector; then calculate the inner product matrix. in This is the geodesic distance square matrix; further, eigenvalue decomposition K is performed on K = UΛU T Retain the first d largest eigenvalues: λ1≥λ2≥…≥λ d and their corresponding eigenvectors u1, u2, ..., u d Finally, the inner product matrix K is embedded in low dimension, i.e. Among them Λ d =diag(λ1,…,λ) d The low-dimensional point set of bilateral walking gait data is obtained as shown in the following formula:

[0065] Y = [y1, y2, ..., y M ] T ∈R M×d ;

[0066] Where each point y i d represents a one-dimensional reduced point of the hip joint trajectory corresponding to the gait cycle, where d is the data dimension.

[0067] Therefore, as a possible implementation of this application, the aforementioned low-dimensional point set of gait data includes a low-dimensional point set of gait data on the affected side and a low-dimensional point set of gait data on the healthy side. The low-dimensional point set of gait data on the affected side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the affected side, and the low-dimensional point set of gait data on the healthy side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the healthy side.

[0068] The aforementioned evaluation index calculation module 101 is also used to process the low-dimensional point set of gait data using the kernel density estimation (KDE) method to obtain the walking gait distribution curve.

[0069] In one embodiment, a Gaussian kernel is first selected to perform weighted smoothing on each low-dimensional point. The Gaussian kernel is shown in the following formula:

[0070]

[0071] Among them, u=yy i Represent the midpoint y and the embedding point y in the low-dimensional space i .

[0072] The bandwidth parameter h is selected using the Silverman rule to control the smoothness of the Gaussian kernel, as shown in the following formula:

[0073]

[0074] Where σ is the sample standard deviation.

[0075] The walking gait distribution curve is calculated using the kernel density estimation method, as shown in the following formula:

[0076]

[0077] Where f(y) is the gait distribution curve. The final gait distribution curves for the affected and healthy sides are obtained and denoted as the affected side gait distribution curve f(y). impaired Gait distribution curves of the healthy side and the unhealthy side healthy .

[0078] The assist parameter optimization module 102 is used to determine the assist parameters based on the walking gait distribution curve.

[0079] In this embodiment of the application, the assist parameter optimization module 102 can continuously optimize and adjust the assist parameters according to the walking gait distribution curve.

[0080] In one embodiment, reference can be made to, for example Figure 2 As shown in the flowchart, after calculating the bilateral walking gait distribution curves, the assist parameter optimization module 102 is also used to calculate the Wasserstein distance between the affected side walking gait distribution curve and the healthy side walking gait distribution curve, wherein the Wasserstein distance is used to characterize the walking difference between the affected side limb and the healthy side limb of the correction subject; the current assist parameters are updated according to the Wasserstein distance to obtain the updated assist parameter set.

[0081] Specifically, Wasserstein is far from W p (f impaired ,f healthy The calculation formula for ) is as follows:

[0082]

[0083] Among them, Π(f impaired ,f healthy Let f be the joint distribution, and let f be the marginal distribution. impaired and f healthy , where p is the power of the distance.

[0084] In this embodiment of the application, the optimization objective is calculated with d=1 and p=1, then the Wasserstein distance W1(f) is... impaired ,f healthy The calculation formula for ) is as follows:

[0085]

[0086] Among them, F impaired (t) and F healthy (t) represent the cumulative distribution functions (CDF) of the gait distribution on the affected and healthy sides, respectively, where t is the integral variable on the real number axis. W1(f impaired ,f healthy This will be used as an optimization target to assist in parameter optimization.

[0087] In one embodiment, the assist parameter optimization module 102 is specifically used to continuously update the current assist parameters based on the Wasserstein distance using a Bayesian optimization algorithm to obtain the updated assist parameters. Specifically, the assist parameter optimization module 102 uses minimizing the walking difference between the two limbs, i.e., the Wasserstein distance, as the optimization objective function, and continuously optimizes and adjusts the assist parameters using a Bayesian optimization algorithm.

[0088] It should be noted that you can refer to, for example Figure 2 The flowchart shown illustrates the auxiliary parameter optimization based on the Bayesian optimization method. First, the search space X and the maximum number of iterations T are set. Then, Latin hypercube sampling (LHS) is used to select a set of initial parameters within the search space X. And calculate the walking distribution distance of the subject's bilateral lower limbs after the corresponding initialization parameters are used for assistance. And construct the initialization parameter set, as shown in the following formula:

[0089] D0 = {x, y};

[0090] Where, n init It is the initial parameter x i The number of groups.

[0091] Based on the current dataset Dt-1 The Gaussian Process (GP) is trained as a surrogate model to model the probabilistic mapping relationship between the parameters x and the objective function W1, as shown in the following formula:

[0092] p(W1|D t-1 )~GP(μ(x),k(x,x′));

[0093] Wherein, p(W|D t-1 ) refers to the current dataset D t-1 Given the condition, the posterior probability distribution of the objective function W1 is given, where μ(x) is the mean function and k(x,x′) is the covariance function (kernel function). This patent uses the common squared exponential kernel, which has the following form:

[0094]

[0095] in, is the signal variance, which controls the overall amplitude of the function output; l is the length scale, which determines the smoothness of the function.

[0096] Optimize the acquisition function and calculate the assist parameters for the next iteration. Select the parameter selection mechanism that prioritizes the expected improvement (EI) acquisition function in Bayesian optimization.

[0097] α EI (x)=E[max(0,W1-W 1,min -ξ)];

[0098] Among them, W 1,min Let ξ be the current minimum distance value, and let ξ be the exploration term, which helps prevent the algorithm from prematurely converging to a local optimum. The acquisition function α is improved by maximizing the expected value. EI (x), solve for the next set of assist parameters x t :

[0099]

[0100] Here, arg min represents finding the parameter that minimizes the objective function value.

[0101] Evaluate assist parameters based on the selected parameter x t Adjust the assist curve and assist the subject in walking, repeating all the aforementioned processes for calculating and optimizing the objective function to obtain the evaluation result y of the objective function. t The calculation formula is as follows:

[0102] y t =W1+∈;

[0103] Wherein, ∈ is the noise term, used to simulate measurement errors such as sensor noise and fluctuations in gait data acquisition.

[0104] Update the historical assist parameter set, and add new data points (x) t ,y t Add to historical dataset D t =D t-1 ∪{(x t ,y t If y t <W 1,min Then update the current optimal value to y. t Otherwise, it remains unchanged.

[0105] The process involves continuously optimizing and finding the optimal assist parameters, iteratively repeating the steps of training the surrogate model and updating the historical assist parameter set until the maximum number of iterations T or W1 convergence is achieved. After completing all iterations, the final surrogate model is used to expand the parameter space X. T Selecting the optimal solution x * :

[0106]

[0107] Where, μ T (x) is the mean of the objective function prediction for parameter x by the Gaussian process at the Tth iteration.

[0108] The assist device 11 is used to assist the corrected object in walking according to the assist parameters.

[0109] In this embodiment, the assistive device 11 can specifically be an exoskeleton, a wearable robotic device designed to help patients regain motor function, improve gait, enhance muscle strength, and promote neurological rehabilitation. The exoskeleton compensates for motor impairments (such as hemiplegia, muscle weakness, and balance problems) caused by stroke through mechanical support, power assistance, and intelligent feedback.

[0110] In one embodiment, the assist device 11 can generate an assist curve based on the assist parameters output by the assist parameter optimization module 102 and apply it to the exoskeleton, which can then be applied to stroke patients through the exoskeleton control system.

[0111] The gait correction system disclosed in the above embodiments of this application includes a processing device and an assistive device. The processing device is used to: acquire gait data of the correction subject; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation to obtain a walking gait distribution curve; determine assistive parameters based on the walking gait distribution curve; and the assistive device is used to assist the correction subject in walking according to the assistive parameters. Thus, by constructing the probability distribution of the correction subject's motion trajectory through isometric mapping and kernel density estimation, a probabilistic representation of the joint distribution of spatiotemporal and dynamic parameters of walking gait is achieved without losing high-dimensional information. Using a Bayesian optimization method based on Wasserstein distance, adaptive adjustment of assistive parameters and synergistic optimization of multi-dimensional gait symmetry are achieved. Optimizing the assistive control strategy by dynamically adjusting assistive parameters based on the hip joint angle of stroke patients helps to achieve personalized adaptive control of the lower limb assistive exoskeleton, thereby improving the effectiveness of gait correction.

[0112] To facilitate understanding of the above embodiments, the following is in conjunction with... Figure 3 Taking gait correction for stroke patients as an example, this paper presents a schematic diagram of a gait correction system's processing device execution scenario. Figure 3 As shown, firstly, the hip joint walking trajectory set of stroke patients during a certain period of walking is collected using an IMU. After preprocessing the hip joint walking trajectory set, the gait cycle sample X corresponding to the impaired hip angle data of the corrected subject is obtained. impaired And the healthy hip angle data corresponding to the healthy side gait cycle sample X healthy Gait cycle sample X on the affected side impaired and healthy side gait cycle sample X healthy The horizontal axis represents the gait phase, the vertical axis represents the hip angle, and the upward arrow represents flexion.

[0113] The gait cycle samples X on the affected side were analyzed separately. impaired and healthy side gait cycle sample X healthy Perform isometric mapping to obtain the low-dimensional point set Y of the affected side gait data. impaired And the low-dimensional point set Y of the healthy side gait data healthy The gait distribution curve f of the affected side was obtained using kernel density estimation. impaired Gait distribution curves of the healthy side and the walking side healthyNext, the computed wasserstein distance is calculated based on the walking gait distribution curve, using W1(f impaired ,f healthy The lines in the figure represent the probability density function of the affected leg (impaired leg PDF), the probability density function of the healthy leg (healthy leg PDF), the cumulative distribution function of the affected leg (impaired leg CDF), the cumulative distribution function of the healthy leg (healthy leg CDF), and the Wasserstein distance, respectively.

[0114] Using Wasserstein distance as the objective function, the assistive torque parameter is continuously optimized and adjusted using the Bayesian optimization algorithm. Based on the optimized assistive torque parameter, the assistive curve is continuously adjusted and applied to the exoskeleton.

[0115] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0116] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0117] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0118] In the embodiments provided in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0119] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0120] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-described embodiments of this application can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can: acquire the gait data of the correction object; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation to obtain a walking gait distribution curve; determine the assistance parameters based on the walking gait distribution curve; and use an assistance device to assist the correction object in walking according to the assistance parameters.

[0122] Furthermore, in one possible implementation of this application embodiment, the above-mentioned correction object is a stroke patient, and the correction object includes the affected limb and the unaffected limb.

[0123] Furthermore, in another possible implementation of this application embodiment, the above-mentioned dimensionality reduction of gait data using isometric mapping to obtain a low-dimensional point set of gait data includes:

[0124] Noise reduction of gait data is achieved using a low-pass filter;

[0125] For the denoised gait data, multiple gait cycles are identified using an adaptive thresholding method.

[0126] Based on multiple gait cycles, multiple gait cycle samples of the affected side and multiple gait cycle samples of the healthy side of the subject were determined;

[0127] We used isometric mapping to reduce the dimensionality of multiple gait cycle samples from the affected side and multiple gait cycle samples from the healthy side, respectively, to obtain a low-dimensional point set of gait data.

[0128] Furthermore, in another possible implementation of this application embodiment, the aforementioned low-dimensional point set of gait data includes a low-dimensional point set of gait data on the affected side and a low-dimensional point set of gait data on the healthy side. The low-dimensional point set of gait data on the affected side includes one-dimensional reduced points of the hip joint trajectory of multiple hip joints during the affected side gait cycle, and the low-dimensional point set of gait data on the healthy side includes one-dimensional reduced points of the hip joint trajectory of multiple hip joints during the healthy side gait cycle.

[0129] Furthermore, in another possible implementation of this application embodiment, the low-pass filter is a fourth-order Butterworth low-pass filter.

[0130] Furthermore, in another possible implementation of this application embodiment, the aforementioned gait distribution curve includes the gait distribution curve of the affected side and the gait distribution curve of the unaffected side. The determination of assist parameters based on the gait distribution curve includes:

[0131] Determine the Wasserstein distance between the gait distribution curves of the affected side and the healthy side, whereby the Wasserstein distance is used to characterize the gait difference between the affected and healthy limbs of the subject undergoing correction.

[0132] The current boost parameters are updated based on the Wasserstein distance to obtain the updated boost parameters.

[0133] Furthermore, in another possible implementation of this application embodiment, updating the current assist parameters based on the Wasserstein distance to obtain the updated assist parameters includes:

[0134] Based on the Wasserstein distance, the current assist parameters are continuously updated using a Bayesian optimization algorithm to obtain the updated assist parameters.

[0135] Furthermore, in another possible implementation of this application embodiment, an inertial measurement unit is respectively set on the affected limb and the unaffected limb of the above-mentioned correction object to acquire the gait data of the correction object, including:

[0136] Gait data is acquired by using inertial measurement units set on the affected and healthy limbs of the patient undergoing correction.

[0137] Furthermore, in another possible implementation of the embodiments of this application, the above-mentioned assistive device is an exoskeleton.

[0138] The computer-readable storage medium disclosed in the above embodiments of this application first acquires the gait data of the correction subject; then, it uses isometric mapping to reduce the dimensionality of the gait data, obtaining a low-dimensional point set of gait data; next, it uses kernel density estimation to process the low-dimensional point set of gait data, obtaining a walking gait distribution curve; finally, it determines the assist parameters based on the walking gait distribution curve. Thus, by constructing the probability distribution of the correction subject's motion trajectory through isometric mapping and kernel density estimation, a probabilistic representation of the joint distribution of spatiotemporal and dynamic parameters of walking gait is achieved without losing high-dimensional information. Using a Bayesian optimization method based on Wasserstein distance, adaptive adjustment of assist parameters and synergistic optimization of multi-dimensional gait symmetry are achieved. Optimizing the assist control strategy by dynamically adjusting assist parameters based on the hip joint angle of stroke patients helps to achieve personalized adaptive control of the lower limb assistive exoskeleton, improving the effectiveness of gait correction.

[0139] The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0140] The methods described in this application can be implemented in whole or in part by a computer program product. When the computer program product is run on a terminal device, the terminal device can perform the following: acquire gait data of the correction object; reduce the dimensionality of the gait data using isometric mapping to obtain a low-dimensional point set of gait data; process the low-dimensional point set of gait data using kernel density estimation to obtain a walking gait distribution curve; determine the assist parameters based on the walking gait distribution curve; and use an assist device to assist the correction object in walking according to the assist parameters.

[0141] Furthermore, in one possible implementation of this application embodiment, the above-mentioned correction object is a stroke patient, and the correction object includes the affected limb and the unaffected limb.

[0142] Furthermore, in another possible implementation of this application embodiment, the above-mentioned dimensionality reduction of gait data using isometric mapping to obtain a low-dimensional point set of gait data includes:

[0143] Noise reduction of gait data is achieved using a low-pass filter;

[0144] For the denoised gait data, multiple gait cycles are identified using an adaptive thresholding method.

[0145] Based on multiple gait cycles, multiple gait cycle samples of the affected side and multiple gait cycle samples of the healthy side of the subject were determined;

[0146] We used isometric mapping to reduce the dimensionality of multiple gait cycle samples from the affected side and multiple gait cycle samples from the healthy side, respectively, to obtain a low-dimensional point set of gait data.

[0147] Furthermore, in another possible implementation of this application embodiment, the aforementioned low-dimensional point set of gait data includes a low-dimensional point set of gait data on the affected side and a low-dimensional point set of gait data on the healthy side. The low-dimensional point set of gait data on the affected side includes one-dimensional reduced points of the hip joint trajectory of multiple hip joints during the affected side gait cycle, and the low-dimensional point set of gait data on the healthy side includes one-dimensional reduced points of the hip joint trajectory of multiple hip joints during the healthy side gait cycle.

[0148] Furthermore, in another possible implementation of this application embodiment, the low-pass filter is a fourth-order Butterworth low-pass filter.

[0149] Furthermore, in another possible implementation of this application embodiment, the aforementioned gait distribution curve includes the gait distribution curve of the affected side and the gait distribution curve of the unaffected side. The determination of assist parameters based on the gait distribution curve includes:

[0150] Determine the Wasserstein distance between the gait distribution curves of the affected side and the healthy side, whereby the Wasserstein distance is used to characterize the gait difference between the affected and healthy limbs of the subject undergoing correction.

[0151] The current boost parameters are updated based on the Wasserstein distance to obtain the updated boost parameters.

[0152] Furthermore, in another possible implementation of this application embodiment, updating the current assist parameters based on the Wasserstein distance to obtain the updated assist parameters includes:

[0153] Based on the Wasserstein distance, the current assist parameters are continuously updated using a Bayesian optimization algorithm to obtain the updated assist parameters.

[0154] Furthermore, in another possible implementation of this application embodiment, an inertial measurement unit is respectively set on the affected limb and the unaffected limb of the above-mentioned correction object to acquire the gait data of the correction object, including:

[0155] Gait data is acquired by using inertial measurement units set on the affected and healthy limbs of the patient undergoing correction.

[0156] Furthermore, in another possible implementation of the embodiments of this application, the above-mentioned assistive device is an exoskeleton.

[0157] The computer program product disclosed in the above embodiments of this application first acquires the gait data of the subject being corrected; then, it uses isometric mapping to reduce the dimensionality of the gait data, obtaining a low-dimensional point set of gait data; next, it uses kernel density estimation to process the low-dimensional point set of gait data, obtaining a walking gait distribution curve; finally, it determines the assist parameters based on the walking gait distribution curve. Thus, by constructing the probability distribution of the movement trajectory of the subject being corrected through isometric mapping and kernel density estimation, a probabilistic representation of the joint distribution of spatiotemporal and dynamic parameters of walking gait is achieved without losing high-dimensional information. Using a Bayesian optimization method based on Wasserstein distance, adaptive adjustment of assist parameters and synergistic optimization of multi-dimensional gait symmetry are achieved. Optimizing the assist control strategy by dynamically adjusting assist parameters based on the hip joint angle of stroke patients helps to achieve personalized adaptive control of the lower limb assistive exoskeleton, improving the effectiveness of gait correction.

[0158] The embodiments described above are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A gait correction system, characterized in that, It includes a processing device and an assisting device, the processing device being used for: Gait data of the subject to be corrected is obtained. The subject to be corrected has an asymmetrical gait during walking. The subject to be corrected includes the affected limb and the unaffected limb. The gait data is reduced in dimensionality using isometric mapping to obtain a low-dimensional point set of gait data. The gait data low-dimensional point set is processed using the kernel density estimation method to obtain the walking gait distribution curve, which includes the walking gait distribution curve of the affected side and the walking gait distribution curve of the healthy side. Based on the gait distribution curve, assist parameters are determined; the determination of assist parameters based on the gait distribution curve includes: determining the Wasserstein distance between the gait distribution curve of the affected side and the gait distribution curve of the healthy side, wherein the Wasserstein distance is used to characterize the gait difference between the affected limb and the healthy limb of the subject being corrected; updating the current assist parameters based on the Wasserstein distance to obtain the updated assist parameters; The assist device is used to assist the corrected object in walking according to the assist parameters.

2. The gait correction system according to claim 1, characterized in that, The target group for the correction is stroke patients.

3. The gait correction system according to claim 2, characterized in that, The step of using isometric mapping to reduce the dimensionality of the gait data to obtain a low-dimensional point set of gait data includes: The gait data is denoised using a low-pass filter; For the denoised gait data, an adaptive thresholding method is used to identify multiple gait cycles; Based on the multiple gait cycles, multiple affected-side gait cycle samples and multiple unaffected-side gait cycle samples are determined for the corrective subject; The multiple gait cycle samples from the affected side and the multiple gait cycle samples from the healthy side are dimensionality reduced using isometric mapping to obtain a low-dimensional point set of the gait data.

4. The gait correction system according to claim 3, characterized in that, The low-dimensional point set of gait data includes a low-dimensional point set of gait data on the affected side and a low-dimensional point set of gait data on the healthy side. The low-dimensional point set of gait data on the affected side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the affected side, and the low-dimensional point set of gait data on the healthy side includes one-dimensional reduced points of the hip joint trajectory of multiple gait cycles on the healthy side.

5. The gait correction system according to claim 3, characterized in that, The low-pass filter is a fourth-order Butterworth low-pass filter.

6. The gait correction system according to claim 1, characterized in that, The step of updating the current assist parameters based on the Wassstein distance to obtain the updated assist parameters includes: Based on the Wasserstein distance, the current assist parameters are continuously updated using a Bayesian optimization algorithm to obtain the updated assist parameters.

7. The gait correction system according to claim 1, characterized in that, An inertial measurement unit is set up on both the affected and unaffected limbs of the subject being corrected. The acquisition of the subject's gait data includes: The gait data is acquired through the inertial measurement units set on the affected and unaffected limbs of the subject undergoing correction.

8. The gait correction system according to any one of claims 1-7, characterized in that, The assistive device is an exoskeleton.

9. A computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to achieve the following: Gait data of the subject to be corrected is obtained. The subject to be corrected has an asymmetrical gait during walking. The subject to be corrected includes the affected limb and the unaffected limb. The gait data is reduced in dimensionality using isometric mapping to obtain a low-dimensional point set of gait data. The gait data low-dimensional point set is processed using the kernel density estimation method to obtain the walking gait distribution curve, which includes the walking gait distribution curve of the affected side and the walking gait distribution curve of the healthy side. Based on the walking gait distribution curve, assist parameters are determined; the determination of assist parameters based on the walking gait distribution curve includes: determining the Wasserstein distance between the walking gait distribution curve of the affected side and the walking gait distribution curve of the healthy side, wherein the Wasserstein distance is used to characterize the walking difference between the affected limb and the healthy limb of the subject being corrected; updating the current assist parameters based on the Wasserstein distance to obtain the updated assist parameters.

Images