Feedback control device and method for human lower limb muscle strength training

Abstract

The present invention discloses a feedback control device and method for human lower limb muscle strength training. The device comprises: a force acquisition module, a lower limb motion acquisition module, a training evaluation module and a feedback calculation module; the force acquisition module is used to acquire a time series of force values ​​emitted by a user's foot when the user uses a human lower limb muscle strength training device to perform human lower limb muscle strength training; the lower limb motion acquisition module is used to acquire a set of key point spatial position matrices when the user uses the human lower limb muscle strength training device to perform human lower limb muscle strength training; the training evaluation module is used to evaluate and process the force value time series and the key point spatial position matrix set to obtain an evaluation sequence value; the feedback calculation module is used to calculate and process the force value time series, the key point spatial position matrix set and the evaluation sequence value to obtain a feedback value sequence.

Description

Technical Field

[0001] The invention relates to the field of exoskeleton robots and feedback control, and in particular to a feedback control device and method for human lower limb muscle strength training. Background Art

[0002] In order to solve the problem of lower limb muscle strength training for users with lower limb movement disorders, a variety of lower limb strength training equipment has been developed. For typical users with lower limb movement disorders, including users with lower limb dysfunction caused by common diseases such as spinal cord injury, stroke, brain trauma, fracture surgery, users after fracture surgery, osteoarthritis, rheumatoid arthritis, joint replacement surgery, and elderly people or users who are bedridden for a long time and need to increase exercise. The current lower limb strength training equipment mainly changes the patient's training angle under weight loss to achieve squatting training from lying to standing posture; strengthen the lower limb flexors and extensors, bones, ligaments, joints and proprioception, while promoting the coordinated control ability and stability of the hip, knee and ankle, laying a good foundation for future walking training.

[0003] When constructing a training environment, current lower limb strength training equipment only constructs the user's sitting posture scenario and footstep force module. It lacks the evaluation of the user's force information and movement standards, and lacks feedback based on the evaluation results, such as modifying the amount of force applied to the user, which affects the training efficiency and effectiveness. Summary of the invention

[0004] The present invention mainly solves the problem that existing lower limb muscle strength training equipment lacks the ability to evaluate the user's force information and movement standards, and provide feedback based on the evaluation results, such as modifying the amount of force applied to the user, which affects the training efficiency and effect. The present invention discloses a feedback control device and method for lower limb muscle strength training of the human body.

[0005] In a first aspect of the embodiment of the present application, a feedback control device for lower limb muscle strength training of a human body is disclosed, comprising: a strength acquisition module, a lower limb motion acquisition module, a training evaluation module and a feedback calculation module;

[0006] The force acquisition module is used to acquire a time series of force values ​​generated by the user's feet during the process of using the human lower limb muscle strength training device to perform human lower limb muscle strength training; the force value time series includes force values ​​at several acquisition moments;

[0007] The lower limb motion acquisition module is used to acquire a set of key point spatial position matrices of the user during the process of performing lower limb muscle strength training using the human lower limb muscle strength training device;

[0008] The training evaluation module is connected to the strength acquisition module, the lower limb movement acquisition module and the feedback calculation module respectively, and is used to evaluate the strength value time series and the key point space position matrix set to obtain an evaluation sequence value; the evaluation sequence value includes a training effect evaluation value at each acquisition moment;

[0009] The feedback calculation module is used to calculate and process the force value time series, the key point spatial position matrix set and the evaluation sequence value to obtain a feedback value sequence; the feedback value sequence includes the feedback value corresponding to each acquisition moment; the feedback value is the adjustment value of the force applied to the user by the human lower limb muscle strength training device during the user using the human lower limb muscle strength training device to perform human lower limb muscle strength training.

[0010] The lower limb motion acquisition module is used to acquire real-time image information of the training scene, perform human body key point recognition on the real-time image information of the training scene, and obtain a key point spatial position matrix set;

[0011] The key point spatial position matrix set includes the key point spatial position matrix at each acquisition moment; the row vector of the key point spatial position matrix is ​​the position coordinate of each key point of the lower limbs of the human body; the real-time image information of the training scene includes images of the user performing lower limb muscle strength training using the lower limb muscle strength training device.

[0012] The training evaluation module is used to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value, including:

[0013] The training evaluation module performs a joint evaluation process on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment;

[0014] The evaluation sequence value is constructed using the training effect evaluation values ​​at all acquisition moments.

[0015] The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including:

[0016] The feedback calculation module performs feedback calculation processing on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment;

[0017] The feedback values ​​at all acquisition moments are fused to obtain a feedback value sequence.

[0018] In a second aspect of the embodiment of the present invention, a feedback control method for human lower limb muscle strength training is disclosed, which is implemented by using the feedback control device for human lower limb muscle strength training, and includes:

[0019] S1, using the force collection module to collect a time series of force values ​​generated by the user's feet during the lower limb muscle strength training of the human body using the human lower limb muscle strength training device; the force value time series includes force values ​​at several collection moments;

[0020] S2, using the lower limb motion acquisition module to acquire a set of key point spatial position matrices of the user during the lower limb muscle strength training using the human lower limb muscle strength training device;

[0021] S3, using the training evaluation module to evaluate the force value time series and the key point spatial position matrix set to obtain an evaluation sequence value;

[0022] S4, using the feedback calculation module, calculating and processing the force value time series, key point spatial position matrix set and evaluation sequence value to obtain a feedback value sequence.

[0023] The method of using the training evaluation module to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value includes:

[0024] Using the training evaluation module, a joint evaluation process is performed on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment;

[0025] The evaluation sequence value is constructed using the training effect evaluation values ​​at all acquisition moments.

[0026] The joint evaluation process of the force value and the key point spatial position matrix at each acquisition moment to obtain the training effect evaluation value at the acquisition moment includes:

[0027] Obtaining a standard spatial position matrix and a standard strength value; the row vector of the standard spatial position matrix is ​​the position coordinates of each key point of the lower limb of a normal user during the process of performing lower limb strength training using the lower limb strength training device; the standard strength value is the strength value generated by the normal user during the process of performing lower limb strength training using the lower limb strength training device;

[0028] Perform deviation calculation on the key point space position matrix and the standard space position matrix at a collection moment to obtain position deviation value and angle deviation value;

[0029] Perform deviation calculation on the force value at a collection moment to obtain the force deviation value;

[0030] The position deviation value, angle deviation value and force deviation value are fused and calculated to obtain a training effect evaluation value at the collection moment.

[0031] The calculation expression of the position deviation value is:

[0032] a1=|x0-x1| / |x0+x1|+|1-y0 / y1|+|exp(z0 / z1)-p1| / |z0+z1|, where p1 is the calculation parameter, a1 is the position deviation value, (x 0i ,y 0i ,z 0i ) is the row vector of the i-th human lower limb key point in the standard spatial position matrix, (x i ,y i ,z i ) is the row vector of the i-th human lower limb key point in the key point spatial position matrix;

[0033] The calculation expression of the angle deviation value is:

[0034] c1=arctan(|x0-x1| / |y0-y1|)

[0035] c2=arctan(|x0-x1| / |z0-z1|),

[0036] Among them, c1 and c2 are angle deviation values;

[0037] The calculation expression of the force deviation value is:

[0038] β=sin(|f1-f0| / |f1+f0|),

[0039] Among them, β is the force deviation value.

[0040] The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including:

[0041] Using the feedback calculation module, feedback calculation processing is performed on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment;

[0042] The feedback values ​​at all acquisition moments are fused to obtain a feedback value sequence.

[0043] The feedback calculation process is performed on the force value, key point spatial position matrix and training effect evaluation value at each acquisition moment to obtain the feedback value at the acquisition moment, including:

[0044] Subtract the key point spatial position matrix at a collection moment from the standard spatial position matrix to obtain a difference matrix;

[0045] Performing deviation calculation processing on the difference matrix to obtain a deviation vector;

[0046] The expression of the deviation calculation process is:

[0047]

[0048] Among them, v i is the i-th element of the deviation vector, is the mean of all elements in the ith column of the difference matrix, a ki is the element in the kth row and ith column of the difference matrix, π is the pi constant, and m and n are the row and column dimensions of the difference matrix respectively;

[0049] Calculate the median value χ and variance value γ of the deviation vector;

[0050] Performing fusion calculation processing on the strength value and the training effect evaluation value at a collection moment to obtain the feedback value at the collection moment;

[0051] The expression of the fusion calculation process is:

[0052]

[0053] Among them, p is the feedback value and θ is the training effect evaluation value.

[0054] The beneficial effects of the present invention are:

[0055] In the process of constructing the training environment of the current lower limb muscle strength training equipment, the present invention evaluates the user's force information and movement standard, and establishes feedback based on the evaluation results, which is expected to improve the training efficiency and effect of users with lower limb movement disorders.

[0056] The present invention obtains an evaluation sequence value by evaluating the force value time series and the key point spatial position matrix set. The present invention fully extracts the force characteristics and position characteristics of the user at each training moment by jointly evaluating the force value and the key point spatial position matrix at each acquisition moment, thereby ensuring the accuracy and comprehensiveness of the evaluation result. In the process of position evaluation, the accuracy of position evaluation is improved by evaluating the position offset and the angle offset respectively.

[0057] In the process of feedback calculation, the present invention establishes a special calculation model for deviation calculation processing and fusion calculation processing, thereby ensuring the accuracy and real-time performance of the feedback amount. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is a block diagram of the composition of the device of the present invention, Figure 2 It is a flow chart for implementing the method of the present invention. DETAILED DESCRIPTION

[0059] In order to better understand the content of the present invention, an embodiment is given here.

[0060] Figure 1 is a block diagram of the composition of the device of the present invention, Figure 2 It is a flow chart for implementing the method of the present invention.

[0061] In view of the problem that existing lower limb muscle strength training equipment lacks the ability to evaluate the user's force information and movement standards, and provide feedback based on the evaluation results, such as modifying the amount of force applied to the user, which affects the training efficiency and effect, the present invention discloses a feedback control device and method for lower limb muscle strength training of the human body.

[0062] In a first aspect of the embodiment of the present application, a feedback control device for lower limb muscle strength training of a human body is disclosed, comprising: a strength acquisition module, a lower limb motion acquisition module, a training evaluation module and a feedback calculation module;

[0063] The force acquisition module is used to acquire a time series of force values ​​generated by the user's feet during the process of using the human lower limb muscle strength training device to perform human lower limb muscle strength training; the force value time series includes force values ​​at several acquisition moments;

[0064] The lower limb motion acquisition module is used to acquire a set of key point spatial position matrices of the user during the process of performing lower limb muscle strength training using the human lower limb muscle strength training device;

[0065] The training evaluation module is connected to the strength acquisition module, the lower limb movement acquisition module and the feedback calculation module respectively, and is used to evaluate the strength value time series and the key point space position matrix set to obtain an evaluation sequence value; the evaluation sequence value includes a training effect evaluation value at each acquisition moment;

[0066] The feedback calculation module is used to calculate and process the force value time series, the key point spatial position matrix set and the evaluation sequence value to obtain a feedback value sequence; the feedback value sequence includes the feedback value corresponding to each acquisition moment; the feedback value is the adjustment value of the force applied by the human lower limb muscle strength training device to the user during the user using the human lower limb muscle strength training device to perform human lower limb muscle strength training;

[0067] The feedback control device for human lower limb muscle strength training of the present invention is used for a human lower limb muscle strength training device.

[0068] The lower limb action acquisition module is used to acquire a real-time image of a training scene, perform human key point recognition on the real-time image of the training scene, and obtain a key point spatial position matrix set; the key point spatial position matrix set includes a key point spatial position matrix at each acquisition moment; the row vector of the key point spatial position matrix is ​​the position coordinate of each key point of the lower limb of the human body;

[0069] The training evaluation module is used to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value, including:

[0070] The training evaluation module performs a joint evaluation process on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment;

[0071] Using the training effect evaluation values ​​at all acquisition moments, an evaluation sequence value is constructed;

[0072] The joint evaluation process of the force value and the key point spatial position matrix at each acquisition moment to obtain the training effect evaluation value at the acquisition moment includes:

[0073] Obtaining a standard spatial position matrix and a standard strength value; the row vector of the standard spatial position matrix is ​​the position coordinates of each key point of the lower limb of a normal user during the process of performing lower limb strength training using the lower limb strength training device; the standard strength value is the strength value generated by the normal user during the process of performing lower limb strength training using the lower limb strength training device;

[0074] Perform deviation calculation on the key point space position matrix and the standard space position matrix at a collection moment to obtain position deviation value and angle deviation value;

[0075] Perform deviation calculation on the force value at a collection moment to obtain the force deviation value;

[0076] The position deviation value, angle deviation value and force deviation value are fused and calculated to obtain a training effect evaluation value at the collection moment.

[0077] The calculation expression of the position deviation value is:

[0078] ea i =|x 0i -x i | / |x 0i +x i |+|1-y 0i / y i |+|exp(z 0i / z i)-t1| / |z 0i +z i |,

[0079]

[0080] Among them, t1 is the preset calculation parameter, ea i is the i-th position deviation component, ea is the position deviation value, n is the number of key points of the lower limbs of the human body, (x 0i ,y 0i ,z 0i ) is the row vector of the i-th human lower limb key point in the standard spatial position matrix, (x i ,y i ,z i ) is the row vector of the i-th human lower limb key point in the key point spatial position matrix;

[0081] The calculation expression of the angle deviation value is:

[0082]

[0083] Among them, ec i is the first component of the ith angle deviation, ed i is the second component of the ith angle deviation, ef is the angle deviation value, and pi is the pi constant;

[0084] The calculation expression of the force deviation value is:

[0085] β=sin(|f1-f0| / |f1+f0|),

[0086] Among them, β is the force deviation value, f1 is the force value at a collection moment, and f0 is the standard force value;

[0087] The position deviation value, the angle deviation value and the force deviation value are subjected to fusion calculation processing to obtain the training effect evaluation value at the acquisition moment, wherein the expression of the fusion calculation processing is:

[0088]

[0089] Among them, θ is the training effect evaluation value at the time of collection.

[0090] The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including:

[0091] Performing feedback calculation processing on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment;

[0092] The feedback values ​​at all acquisition moments are used for fusion processing to obtain a feedback value sequence;

[0093] The feedback calculation process is performed on the force value, key point spatial position matrix and training effect evaluation value at each acquisition moment to obtain the feedback value at the acquisition moment, including:

[0094] Subtract the key point spatial position matrix at a collection moment from the standard spatial position matrix to obtain a difference matrix;

[0095] Performing deviation calculation processing on the difference matrix to obtain a deviation vector;

[0096] The expression of the deviation calculation process is:

[0097]

[0098] Among them, v i is the i-th element of the deviation vector, is the mean of all elements in the ith column of the difference matrix, a ki is the element in the kth row and ith column of the difference matrix, π is the pi constant, and m and n are the row and column dimensions of the difference matrix respectively;

[0099] Calculate the median value χ and variance value γ of the deviation vector;

[0100] Performing fusion calculation processing on the strength value and the training effect evaluation value at a collection moment to obtain the feedback value at the collection moment;

[0101] The expression of the fusion calculation process is:

[0102]

[0103] Among them, p is the feedback value, θ is the training effect evaluation value;

[0104] The human body key point recognition can be implemented by using SURF feature point detection or corner point detection algorithm, or OpenPose algorithm in OpenCV.

[0105] The key points of the human lower limbs include knee joints, ankle joints, feet, etc.

[0106] In a second aspect of the embodiment of the present application, a feedback control method for human lower limb muscle strength training is disclosed, which is implemented by using the feedback control device for human lower limb muscle strength training, including:

[0107] S1, using the force collection module to collect a time series of force values ​​generated by the user's feet during the lower limb muscle strength training of the human body using the human lower limb muscle strength training device; the force value time series includes force values ​​at several collection moments;

[0108] S2, using the lower limb motion acquisition module to acquire a set of key point spatial position matrices of the user during the lower limb muscle strength training using the human lower limb muscle strength training device;

[0109] S3, using the training evaluation module to evaluate the force value time series and the key point spatial position matrix set to obtain an evaluation sequence value;

[0110] S4, using the feedback calculation module, calculating and processing the force value time series, key point spatial position matrix set and evaluation sequence value to obtain a feedback value sequence.

[0111] The method of using the training evaluation module to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value includes:

[0112] Using the training evaluation module, a joint evaluation process is performed on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment;

[0113] The evaluation sequence value is constructed using the training effect evaluation values ​​at all acquisition moments.

[0114] The joint evaluation process of the force value and the key point spatial position matrix at each acquisition moment to obtain the training effect evaluation value at the acquisition moment includes:

[0115] Obtaining a standard spatial position matrix and a standard strength value; the row vector of the standard spatial position matrix is ​​the position coordinates of each key point of the lower limb of a normal user during the process of performing lower limb strength training using the lower limb strength training device; the standard strength value is the strength value generated by the normal user during the process of performing lower limb strength training using the lower limb strength training device;

[0116] Perform deviation calculation on the key point space position matrix and the standard space position matrix at a collection moment to obtain position deviation value and angle deviation value;

[0117] Perform deviation calculation on the force value at a collection moment to obtain the force deviation value;

[0118] The position deviation value, angle deviation value and force deviation value are fused and calculated to obtain a training effect evaluation value at the collection moment.

[0119] The calculation expression of the position deviation value is:

[0120] ea i =|x 0i -x i | / |x 0i +x i |+|1-y 0i / y i |+|exp(z 0i / z i )-t1| / |z 0i +z i |,

[0121]

[0122] Among them, t1 is the preset calculation parameter, ea i is the i-th position deviation component, ea is the position deviation value, n is the number of key points of the lower limbs of the human body, (x 0i ,y 0i ,z 0i ) is the row vector of the i-th human lower limb key point in the standard spatial position matrix, (x i ,y i ,z i ) is the row vector of the i-th human lower limb key point in the key point spatial position matrix;

[0123] The calculation expression of the angle deviation value is:

[0124]

[0125] Among them, ec i is the first component of the ith angle deviation, ed i is the second component of the ith angle deviation, ef is the angle deviation value, and pi is the pi constant;

[0126] The calculation expression of the force deviation value is:

[0127] β=sin(|f1-f0| / |f1+f0|),

[0128] Among them, β is the force deviation value, f1 is the force value at a collection moment, and f0 is the standard force value;

[0129] The position deviation value, the angle deviation value and the force deviation value are subjected to fusion calculation processing to obtain the training effect evaluation value at the acquisition moment, wherein the expression of the fusion calculation processing is:

[0130]

[0131] Among them, θ is the training effect evaluation value at the time of collection.

[0132] The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including:

[0133] Using the feedback calculation module, feedback calculation processing is performed on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment;

[0134] The feedback values ​​at all acquisition moments are fused to obtain a feedback value sequence.

[0135] The feedback calculation process is performed on the force value, key point spatial position matrix and training effect evaluation value at each acquisition moment to obtain the feedback value at the acquisition moment, including:

[0136] Subtract the key point spatial position matrix at a collection moment from the standard spatial position matrix to obtain a difference matrix;

[0137] Performing deviation calculation processing on the difference matrix to obtain a deviation vector;

[0138] The expression of the deviation calculation process is:

[0139]

[0140] Among them, v i is the i-th element of the deviation vector, is the mean of all elements in the ith column of the difference matrix, a ki is the element in the kth row and ith column of the difference matrix, π is the pi constant, and m and n are the row and column dimensions of the difference matrix respectively;

[0141] Calculate the median value χ and variance value γ of the deviation vector;

[0142] Performing fusion calculation processing on the strength value and the training effect evaluation value at a collection moment to obtain the feedback value at the collection moment;

[0143] The expression of the fusion calculation process is:

[0144]

[0145] Among them, p is the feedback value and θ is the training effect evaluation value.

[0146] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A feedback control method for lower limb muscle strength training of human body, characterized in that: The feedback control device for lower limb muscle strength training is used to achieve the above, including: S1, using the force collection module to collect a time series of force values ​​generated by the user's feet during the lower limb muscle strength training of the human body using the human lower limb muscle strength training device; the force value time series includes force values ​​at several collection moments; S2, using the lower limb motion acquisition module to acquire a set of key point spatial position matrices of the user during the lower limb muscle strength training using the human lower limb muscle strength training device; S3, using the training evaluation module to evaluate the force value time series and the key point spatial position matrix set to obtain an evaluation sequence value; S4, using the feedback calculation module to calculate and process the force value time series, the key point spatial position matrix set and the evaluation sequence value to obtain a feedback value sequence; The method of using the training evaluation module to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value includes: Using the training evaluation module, a joint evaluation process is performed on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment; Using the training effect evaluation values ​​at all acquisition moments, an evaluation sequence value is constructed; The joint evaluation process of the force value and the key point spatial position matrix at each acquisition moment to obtain the training effect evaluation value at the acquisition moment includes: Obtaining a standard spatial position matrix and a standard strength value; the row vector of the standard spatial position matrix is ​​the position coordinates of each key point of the lower limb of a normal user during the process of performing lower limb strength training using the lower limb strength training device; the standard strength value is the strength value generated by the normal user during the process of performing lower limb strength training using the lower limb strength training device; Perform deviation calculation on the key point space position matrix and the standard space position matrix at a collection moment to obtain position deviation value and angle deviation value; Perform deviation calculation on the force value at a collection moment to obtain the force deviation value; Performing fusion calculation processing on the position deviation value, the angle deviation value and the force deviation value to obtain a training effect evaluation value at the collection moment; The calculation expression of the position deviation value is: a1=|x0-x1| / |x0+x1|+|1-y0 / y1|+|exp(z0 / z1)-p1| / |z0+z1|, where p1 is the calculation parameter, a1 is the position deviation value, (x 0i ,y 0i ,z 0i ) is the row vector of the i-th human lower limb key point in the standard spatial position matrix, (x i ,y i ,z i ) is the row vector of the i-th human lower limb key point in the key point spatial position matrix; The calculation expression of the angle deviation value is: c1=arctan(|x0-x1| / |y0-y1|) c2=arctan(|x0-x1| / |z0-z1|), Among them, c1 and c2 are angle deviation values; The calculation expression of the force deviation value is: β=sin(|f1-f0| / |f1+f0|), Among them, β is the force deviation value, f1 is the force value at a collection moment, and f0 is the standard force value.

2. The feedback control method for lower limb muscle strength training of a human body as claimed in claim 1, characterized in that: The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including: Using the feedback calculation module, feedback calculation processing is performed on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment; The feedback values ​​at all acquisition moments are fused to obtain a feedback value sequence.

3. The feedback control method for lower limb muscle strength training of a human body as claimed in claim 2, characterized in that: The feedback calculation process is performed on the force value, key point spatial position matrix and training effect evaluation value at each acquisition moment to obtain the feedback value at the acquisition moment, including: Subtract the key point spatial position matrix at a collection moment from the standard spatial position matrix to obtain a difference matrix; Performing deviation calculation processing on the difference matrix to obtain a deviation vector; The expression of the deviation calculation process is: Among them, v i is the i-th element of the deviation vector, a i is the mean of all elements in the ith column of the difference matrix, a ki is the element in the kth row and ith column of the difference matrix, π is the pi constant, m and n are the row and column dimensions of the difference matrix respectively; Calculate the median value χ and variance value γ of the deviation vector; Performing fusion calculation processing on the strength value and the training effect evaluation value at a collection moment to obtain the feedback value at the collection moment; The expression of the fusion calculation process is: Among them, p is the feedback value and θ is the training effect evaluation value.

4. A feedback control device for lower limb muscle strength training, characterized in that: A feedback control method for human lower limb muscle strength training according to any one of claims 1 to 3, comprising: a strength acquisition module, a lower limb motion acquisition module, a training evaluation module and a feedback calculation module; The force acquisition module is used to acquire a time series of force values ​​generated by the user's feet during the process of using the human lower limb muscle strength training device to perform human lower limb muscle strength training; the force value time series includes force values ​​at several acquisition moments; The lower limb motion acquisition module is used to acquire a set of key point spatial position matrices of the user during the process of performing lower limb muscle strength training using the human lower limb muscle strength training device; The training evaluation module is connected to the strength acquisition module, the lower limb movement acquisition module and the feedback calculation module respectively, and is used to evaluate the strength value time series and the key point space position matrix set to obtain an evaluation sequence value; the evaluation sequence value includes a training effect evaluation value at each acquisition moment; The feedback calculation module is used to calculate and process the force value time series, the key point spatial position matrix set and the evaluation sequence value to obtain a feedback value sequence; the feedback value sequence includes the feedback value corresponding to each acquisition moment; the feedback value is the adjustment value of the force applied to the user by the human lower limb muscle strength training device during the user using the human lower limb muscle strength training device to perform human lower limb muscle strength training.

5. The feedback control device for lower limb muscle strength training of a human body according to claim 1, characterized in that: The lower limb motion acquisition module is used to acquire real-time image information of the training scene, perform human body key point recognition on the real-time image information of the training scene, and obtain a key point spatial position matrix set; The key point spatial position matrix set includes the key point spatial position matrix at each acquisition moment; The row vectors of the key point spatial position matrix are the position coordinates of each key point of the lower limbs of the human body; the real-time image information of the training scene includes images of the user performing lower limb muscle strength training using the lower limb muscle strength training device.

6. The feedback control device for lower limb muscle strength training of a human body as claimed in claim 5, characterized in that: The training evaluation module is used to evaluate the force value time series and the key point space position matrix set to obtain an evaluation sequence value, including: The training evaluation module performs a joint evaluation process on the force value and the key point spatial position matrix at each acquisition moment to obtain a training effect evaluation value at the acquisition moment; The evaluation sequence value is constructed using the training effect evaluation values ​​at all acquisition moments.

7. The feedback control device for lower limb muscle strength training of a human body as claimed in claim 5, characterized in that: The feedback calculation module is used to calculate and process the force value time series, the key point space position matrix set and the evaluation sequence value to obtain a feedback value sequence, including: The feedback calculation module performs feedback calculation processing on the force value, key point spatial position matrix and training effect evaluation value at each collection moment to obtain the feedback value at the collection moment; The feedback values ​​at all acquisition moments are fused to obtain a feedback value sequence.

Images