Method for real-time adjustment of gait training parameters

Abstract

A method for real-time adaptation of gait training parameters (10) applicable to a gait training device (100), wherein the gait training device (100) comprises a detection unit, a training unit and a control unit, wherein the control unit is electrically connected to the detection unit and the training unit and controls the operation of the training unit, the method for real-time adaptation of gait training parameters (10) comprising the steps of: (a) Collecting, by the acquisition unit, the muscle relaxation gait data of at least one first user measured during gait training in a muscle relaxation state and the active power output gait data of the at least one first user measured during gait training in the active power output state of the first user; and then, creating, by the control unit, a standard motion model based on the ratio of the active power output gait data of the first user to the muscle relaxation gait data of the first user; (b) Obtaining a movement model of a second user comprising the muscle relaxation gait data of the second user measured during gait training under the muscle relaxation state of the second user, and then estimating a plurality of personalized training models of different difficulty levels by combining the muscle relaxation gait data of the second user with the standard movement model with 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100% difficulty curves, so that personalized training models with different difficulty levels are created and the difficulty curves can be adjusted according to requirements by the control unit; and (c) determining whether the actual training status of the second user matches the standard of the one of the personalized training models, and then adjusting the one of the personalized training models and providing a replacement training model by the control unit, wherein the gait training comprises at least one gait cycle and the gait cycle is divided into a center of gravity shift interval (F1), a hip flexion interval (F2) and a knee extension interval (F3), wherein, in step (b), the calculation method of the control unit for estimating the personalized training model from the muscle relaxation gait data of the second user comprises, when the muscle relaxation state of the second user is obtained, obtaining a maximum value of the force in a muscle relaxation state and a minimum value of the force in a muscle relaxation state in the gait cycle, whereupon, when the second user is actively outputting force, estimating a predicted maximum value of the active force output and a predicted minimum value of the active force output in the gait cycle of the second user by the standard motion model, whereupon, when the second user is in the active force output state, using the value obtained by the detection unit to obtain an actual maximum value of the active force output and an actual minimum value of the active force output in the gait cycle,When the second user actively exerts force, the actual maximum value of the active force output, the predicted maximum value of the active force output, and the maximum value of the force in the muscle relaxation state are input into a calculation formula for the center of gravity shift interval; and the actual minimum value of the active force output, the predicted minimum value of the active force output, and the minimum value of the force in the muscle relaxation state are input into a calculation formula for the hip flexion interval to obtain the force output level of the center of gravity shift interval (F1) and the force output level of the hip flexion interval (F2), after which the lower force output level is used as these personalized training models of different difficulty levels for the second user.

Description

BACKGROUND OF THE INVENTION 1. Technical field

[0001] The present invention relates to gait training technology and, more particularly, to a method for real-time adjustment of gait training parameters. 2. State of the art

[0002] Generally, gait training uses gait training devices to assist users.

[0003] Reference is made to an electric orthosis as in US 8 147 436 B2. As in Fig. As shown in Figure 7 of the said patent, it primarily collects the gait data of six normal people with active effort as an ideal gait model and plans the tunnel-type allowable error space at the edge of the gait path. Thus, the user obtains a training effect close to the ideal gait training model.

[0004] However, as in the above-mentioned US 8 147 436 B2, when applying the ideal model to all users for training, the individual differences of the users are not taken into account when planning the movement model, so that the ideal model planned in this patent is difficult to suit different users.

[0005] Furthermore, reference is made to an adaptive active training system described in Chinese patent CN 1 13 244 084 A. As described in Fig. As shown in Figure 6 of this patent, it mainly includes a sensor module, a control module, and a movement module. By recording the physiological signals of each section when the user's muscle strength is relaxed and driven by the exoskeleton, the physiological signal threshold of each section is calculated, and the exercise difficulty is adjusted in real time according to the user's physiological state signal during exercise.

[0006] However, as in the aforementioned Chinese patent CN 1 13 244 084 A, only the user's own gait data is used as a reference, and no reference is made to the gait data of ordinary people or other users. Therefore, the training model proposed in this patent cannot achieve the optimal training effect.

[0007] Furthermore, US 2021 / 0 283 001 A1 discloses a system for the rehabilitation of the musculoskeletal system, in particular a robotic platform suitable for optimizing gravity-dependent trunk movements and thus enabling people with walking disabilities, spinal cord injuries, and strokes to move above the ground. At the same time, a lasting improvement in motor skills is promoted when this is done as part of gait rehabilitation using electrical spinal cord stimulation. US 2006293617 A1 discloses another rehabilitation device with at least three degrees of freedom of movement. SUMMARY OF THE INVENTION

[0008] The present invention has been realized under the given circumstances. The main object of the present invention is to provide a method for real-time adjustment of gait training parameters, which can plan a personalized movement model according to the condition of different users and recommend an appropriate training difficulty according to the user's output force data during training, so as to achieve the effect of adjusting the training difficulty in real time according to the actual performance during training.

[0009] To achieve this and other objectives of the present invention, the invention provides a method for real-time adjustment of gait training parameters applicable to a gait training device comprising a sensing unit, a training unit, and a control unit. The control unit is electrically connected to the sensing unit and the training unit and controls the operation of the training unit. The method for real-time adjustment of gait training parameters comprises the following steps: (a) The acquisition unit collects the muscle relaxation gait data of at least one first user measured during gait training in a muscle relaxation state and the active power output gait data of the at least one first user measured during gait training in the active power output state of the first user, whereupon the control unit creates a standard motion model based on the ratio of the active power output gait data of the first user to the muscle relaxation gait data of the first user. (b) The control unit obtains a movement model of a second user including the muscle relaxation gait data of the second user measured during gait training under the muscle relaxation state of the second user, and then estimates at least one personalized training model of different difficulty levels by combining the muscle relaxation gait data of the second user with the standard movement model having 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100% difficulty curves, so that personalized training models with different difficulty levels are created and the difficulty curves can be adjusted according to requirements. (c) The control unit determines whether the actual training state of the second user matches the standard of the at least one personalized training model, and adjusts the at least one personalized training model and provides a replacement training model. The gait training includes at least one gait cycle divided into a center of gravity shift interval, a hip flexion interval, and a knee extension interval. In step (b), the control unit's calculation method for estimating the personalized training model from the second user's muscle relaxation gait data includes, when the second user's muscle relaxation state is obtained, obtaining a maximum value of the force in a muscle relaxation state and a minimum value of the force in a muscle relaxation state in the gait cycle, whereupon, when the second user actively exerts force,a predicted maximum value of the active force output and a predicted minimum value of the active force output in the gait cycle of the second user are estimated by the standard motion model, whereupon, when the second user is in the active force output state, the value obtained by the detection unit is used to obtain an actual maximum value of the active force output and an actual minimum value of the active force output in the gait cycle when the second user is actively exerting force. Furthermore, step (b) comprises substituting the actual maximum value of the active force output, the predicted maximum value of the active force output, and the maximum value of the force in the muscle relaxation state into a calculation formula for the center of gravity shift interval, and incorporating the actual minimum value of the active force output, respectively.the predicted minimum value of active force output and the minimum value of force in the muscle relaxation state into a calculation formula for the hip flexion interval to obtain the force output level of the center of gravity shift interval and the force output level of the hip flexion interval, after which the lower force output level than these will be used for personalized training models of different difficulty levels for the second user.

[0010] Thus, the present invention provides a method for real-time adjustment of gait training parameters. Depending on the condition of the second user, a personalized training model can be planned for the second user. During training, the method can recommend suitable replacement training models according to the force output data of the second user, thus achieving the effect of adjusting the training difficulty in real time according to the actual performance during training. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a flowchart of a general method for real-time adjustment of gait training parameters. Fig. Figure 2 is a schematic diagram of the use state of the method when used in a gait training device. Fig. Figure 2a is a schematic diagram showing the gait cycle. Fig. Figure 2b is a graph showing the first user's muscle relaxation data measured during gait training in the first user's muscle relaxation state. Fig. Figure 2c is a graph g showing the gait data of the first user measured during gait training in an active power output state. Fig. Figure 2d is a graph showing the relationship between the gait data of the first user with active force output and the gait data of the first user with muscle relaxation. Fig. Figure 2e is a graph showing the ratio and average value of the data measured by several initial users in the state of active force output and muscle relaxation. Fig. Figure 2f is a graph showing the interval of center of gravity displacement and the interval of hip flexion. Fig. Figure 2g is a graph showing the interval of center of gravity shift and the interval of knee extension. Fig. 2h is a schematic diagram according to a preferred embodiment of the present invention showing the implementation state of the upper sensor element and the lower sensor element of the knee pressure sensor. Fig. Figure 3 is a diagram showing a personalized training model. Fig. Figure 3a is a diagram showing several personalized training models. Fig. Figure 3b is a graph showing the muscle relaxation state data of the second user. Fig. Figure 3c is a graph showing the predicted maximum active force output and the predicted minimum active force output in the personalized training model of the second user. Fig. Figure 3d is a graph showing the actual maximum value of active force output and the actual minimum value of active force output in the actual training state of the second user. Fig. Figure 3e is a graph showing a surrogate training model obtained after adjusting the personalized training model. DETAILED DESCRIPTION OF THE INVENTION

[0011] In order to describe the technical features of the present invention in detail, the following preferred embodiment will be described with reference to the drawings as follows: As shown in the Fig. 1-2, the method for real-time adjustment of gait training parameters 10 of the present invention is mainly used in conjunction with a gait training device 100. The gait training device 100 mainly includes a sensing unit, a training unit, and a control unit. The sensing unit includes two sole force sensors and two knee pressure sensors. The sole force sensors are a left foot force sensor 101 and a right foot force sensor 102, respectively. The knee pressure sensors are a left knee pressure sensor 103 and a right knee pressure sensor 104, respectively. The training unit includes two pedals 11 and other components for driving the user's lower limbs for training. The left foot force sensor 101 is arranged on one of the pedals 11, and the right foot force sensor 102 is arranged on the other pedal 11. The control unit is electrically connected to the sensor unit and the training unit and controls the operation of the training unit.The control unit has analysis and computing capabilities and may be, but is not limited to, a central processing unit (CPU) or other information processing elements with analysis and computing capabilities. When a first or a second user uses the gait training device 100, the gait training device 100 provides the control, calculation, and operation required for the method for real-time adjustment of the gait training parameters 10. The method for real-time adjustment of the gait training parameters 10 essentially comprises steps (a), (b), and (c). In this preferred embodiment, the left foot force sensor 101 and the right foot force sensor 102 are load cells; the left knee pressure sensor 103 and the right knee pressure sensor 104 are film pressure sensors. It is worth noting that the user can select the appropriate sensor depending on their actual needs, but is not limited to this. As shown in FIG. Fig. As shown in Figure 2a, the gait training according to the invention comprises at least one gait cycle. The gait cycle corresponds to the gait trajectory of one of the feet. The gait trajectory simulates the process of human walking from the beginning of the right heel striking the ground to the lifting of the left toe from the ground, from the striking of the left heel to the lifting of the right toe from the ground, and finally back to the striking of the right heel on the ground. The horizontal axis of the Fig. 2b, Fig. 2c, Fig. 2d, Fig. 2nd, Fig. 2f, Fig. 2g, Fig. 3, Fig. 3a, Fig. 3b, Fig. 3c, Fig. 3D, Fig. 3e corresponds to the gait cycle of Fig. 2a. The data on the graph is divided into 100 equal parts, and the position where the heel of the user's individual foot touches the ground corresponds to the starting point of the gait cycle (i.e., the data point marked 0 on the horizontal axis), and the position before the heel of the same foot touches the ground again corresponds to the 99th data point on the horizontal axis.

[0012] As in the Fig. 1, Fig. 2b, Fig. 2c, Fig. 2d and Fig. 2e, the acquisition unit collects in step (a) the muscle relaxation gait data of a first user measured during gait training in a muscle relaxation state (as shown in Fig. 2b), and the active power output gait data of the first user measured during gait training in the active power output state of the first user (as shown in Fig. 2c). 2c), the control unit creates a standard movement model based on the ratio of the first user's active force output gait data to the first user's muscle relaxation gait data (as shown in Fig. 2d). The "muscle relaxation state" here means that the user does not need to exert any effort during gait training and the training unit is only driven by the control unit of the gait training device 100, causing the user's feet to swing; the "active force output state" refers to the fact that the user's feet must actively exert force during the operation of the training unit.

[0013] To improve the stability of the various data, in this preferred embodiment, multiple samples of the first user are taken. According to the ratio of the average value of the gait data of multiple active force outputs of the first users to the average value of the gait data of the muscle relaxation of the first users' muscles, the standard movement model is then created (as shown in Fig. 2e). In other preferred embodiments, the number of first users may also be assumed to be one if the active output gait data of the first users and the muscle relaxation gait data of the first users are sufficiently representative. Therefore, the number of first users is not limited to this preferred embodiment.

[0014] In this preferred embodiment, the gait cycle is as shown in the Fig. 2a, Fig. 2f and Fig. 2g, it is mainly divided into a center of gravity shift interval F1, a hip flexion interval F2, and a knee extension interval F3. The center of gravity shift interval F1 is located in the 0-40 equal parts of the gait cycle, the hip flexion interval F2 is located in the 45-70 equal parts of the gait cycle, and the knee extension interval F3 is located in the 80-99 equal parts of the gait cycle.

[0015] Taking the right foot of the first user as an example (the assessment method for the left foot is also the same and will not be repeated here), as shown in Fig. 2f, the value detected by the right foot of the first user stepping on the right foot force sensor 102 is greater than a model threshold (the prediction value of the 100% active force output of the first user), when in the hip flexion interval F2, the value detected by the right foot stepping on the right foot force sensor 102 is smaller than the model threshold, as shown in Fig. 2g when the right knee pressure sensor 104 is less than the model threshold in the knee extension interval F3.

[0016] In this preferred embodiment, as in Fig. As shown in Figure 2h, the left knee pressure sensor 103 and the right knee pressure sensor 104 each have an upper sensor element 105 and a lower sensor element 106 (because the upper and lower sensor elements 105 and 106 of the left and right knee pressure sensors 103 and 104 are the same elements and have the same configuration relationship, only one diagram is used as a schematic representation of the left knee pressure sensor 103 and the right knee pressure sensor 104). Assuming that the pressure value P measured by the upper sensor element 105 K1 is, the pressure value P measured by the lower sensor element 106 K2 is the shortest distance between the center of the upper sensor element 105 and the lower end surface of the lower sensor element 106 X2 (in this embodiment, 100 mm), then the pressure center position of the left knee pressure sensor 103 (or the right knee pressure sensor 104) =x1×PK1+X2×PK2PK1+PK2.

[0017] As in the Fig. 1 and Fig. 3, the control unit receives a movement model of a second user in step (b). The movement model of the second user includes the data of the second user's muscle-relaxed gait, which were measured during gait training under the second user's muscle-relaxed state. By combining the data of the second user's muscle-relaxed gait with the standard movement model, a personalized training model is estimated (see Fig. 3). As in Fig. 3a, 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100% difficulty curves are estimated with the standard movement model and the personalized training model (only the 100% difficulty curve L1 and the 10% difficulty curve L12 are shown in Fig. 3a). In this way, personalized training models with different levels of difficulty are created, and the difficulty curves can be adapted according to requirements, so that the difficulty curves are not limited to this preferred embodiment.

[0018] Since the center of gravity transfer interval F1 is the phase in which the right foot descends, the higher the value measured by the force sensor 102 of the right foot, the higher the difficulty level of the personalized training model for the right foot. If the value measured by the right foot force sensor 102 is lower, this means that the difficulty of the personalized training model is lower. Therefore, approximately between the 0-40 gait cycle, the top curve is the 100% difficulty curve L1, and the bottom curve is the 10% difficulty curve L12. The hip flexion interval F2 and the knee extension interval F3 are the right foot raising phases.Therefore, if the value measured by the right foot force sensor 102 is lower, it means that the difficulty of the personalized training model is higher, and if the value measured by the right foot force sensor 102 is higher, it means that the difficulty of the personalized training model is lower. Therefore, approximately between the gait cycles 45-100, the lowest curve is the 100% difficulty curve L1, and the top curve is the 10% difficulty curve L12.

[0019] In this preferred embodiment, as in Fig. 3b, Fig. 3c, Fig. 3d and Fig. 3e, as ways of estimating the personalized training model from the personalized training models of different difficulties that is most suitable for the second user, when the second user is initially in a state of muscle relaxation, taking the right foot as an example, from the value obtained by the force sensor 102 of the right foot, and when the second user is in the state of muscle relaxation, a maximum value of a muscle relaxation state and a minimum value of a muscle relaxation state in the gait cycle are obtained (as in Fig. 3b). Then, when the second user actively exerts force, a predicted maximum value of the active force output and a predicted minimum value of the active force output in the gait cycle of the second user are estimated by the standard motion model (as shown in Fig. 3c). Using the value obtained by the right foot force sensor 102 under the condition of the active force output of the second user, and when the second user is actively outputting force, then an actual maximum value of the active force output and an actual minimum value of the active force output in the gait cycle are obtained (as shown in Fig. 3d). The actual maximum value of active force output, the predicted maximum value of active force output, and the maximum value of force at muscle relaxation are substituted into a calculation formula for the center of gravity shift interval, and the actual minimum value of active force output, the predicted minimum value of active force output, and the minimum value of force at the state of muscle relaxation are substituted into a calculation formula for the hip flexion interval to obtain the force output level of the center of gravity shift interval and the hip flexion interval. Then, the appropriate personalized training model is recommended based on the lower force output level (as shown in Fig. 3e). The specific calculation method is as follows: Formula for calculating the center of gravity transmission interval: Actual maximum value of active force output - maximum value of force in muscle relaxation state / predicted maximum value of active force output - maximum value of muscle relaxation state = initial level of the center of gravity transfer interval F1.

[0020] As in the Fig. 3c and Fig. 3d, in this preferred embodiment, the output level of the second user in the center of gravity transmission interval F1=35.651(kg)−22.066(kg)34.263(kg)−22.066(kg)=111.37%.

[0021] The calculation formula for the hip flexion interval: Actual minimum value of active force output - minimum value of force in the state of muscle relaxation / predicted minimum value of active force output - minimum value of force in the state of muscle relaxation = force output level of the hip flexion interval F2.

[0022] As in the Fig. 3c and Fig. 3d, in this preferred embodiment, the starting level of the second user in the hip flexion interval F2=16.885(kg)−23.68(kg)15.128(kg)−23.68(kg)=79.45%.

[0023] Since the force output level of the hip flexion interval F2 is smaller than the force output level of the center of gravity shift interval F1, the appropriate personalized training mode recommended by the force output level of the hip flexion interval F2 is selected, for example, the model represented by the 80% difficulty curve in Fig. 3a is chosen as a personalized training model.

[0024] As in the Fig. 1, Fig. 3c, Fig. 3D, Fig. 3e and Fig. 4, in step (c), the control unit determines whether the user's actual training state conforms to the standard of the personalized training model, and then adjusts the personalized training model and provides a replacement training model.

[0025] In this preferred embodiment, the control unit determines whether the actual training state of the second user conforms to the standard of the personalized training model, including a continuous determination method within a specific interval and a single-point trigger determination method within the specific interval. The continuous determination method within the specific interval consists of continuously determining whether the actual training state of the second user conforms to the standard of the personalized training model in any of the intervals of the gait cycle (e.g., the center of gravity shift interval F1, the hip flexion interval F2, the knee extension interval F3). If the second user meets the standard at the beginning of the training, the training session of the gait training device 100 maintains the originally set speed.If the user does not meet the standard, the control unit controls the training session to reduce the running speed (in this preferred embodiment, the running speed is set to be reduced by 12% each time, and the minimum speed is reduced to 25% of the originally set speed; this is not limited to this). If the second user meets the standard after the speed of the training session has been reduced, the control unit controls the running speed of the training session to increase by 38% each time until 100% of the originally set speed is reached. The one-point trigger determination consists in having all data in one of the intervals of the gait cycle (e.g.,the center of gravity shift interval F1, the hip flexion interval F2, and the knee extension interval F3) meet the standard of the personalized training model, and they are assumed to meet the standard of the personalized training model in order to avoid the second user requiring continuous force output to adjust the personalized training model.

[0026] In this preferred embodiment, one of the methods for the control unit to determine whether the second user reaches hip flexion in the hip flexion interval F2, using the example of the right foot, is by combining with the parameters of the position of the center of pressure X K , measured by the right knee pressure sensor 104 when the second user is in the active force output state, the average position of the pressure center XK¯, which is measured by the second user in the relaxed state, the force P Fof the second user in the active force output state, which is detected by the right foot force sensor 102, and the force U FR of the second user in a relaxed state, which is detected by the right foot force sensor 102, to determine the personalized training model and the set difficulty. If it is judged that the second user has achieved hip flexion in the hip flexion interval F2, the following conditions must be met: (XK>XK¯+R%×URKX)∩(PF<(UFP−UFR)×R%+UFP), wherein the variation range of the pressure center of the second user in the relaxed state is half the difference between the maximum value and the minimum value of the pressure center position recorded by the hip flexion interval F2 of the second user in the relaxed state, wherein the average pressure center position XK¯ is the average value of the position of the center of pressure captured by the hip flexion interval F2.

[0027] In this preferred embodiment, one of the methods for the control unit to determine whether the second user reaches knee extension in the knee extension interval F3 is, using the right foot as an example, to determine the parameters of the pressure value P K , measured by the right knee pressure sensor 104 of the second user in the state of active force output, the pressure value U RKP , measured by the right knee pressure sensor 104 of the second user with relaxed muscles and the set difficulty R%. When assessing whether the second user has achieved knee extension in the knee extension interval F3, the following conditions must be met: PK<0.9−0.4×R%×URKP.

[0028] In this preferred embodiment, the assessment of whether the second user's actual training status meets the standard of the personalized training model using the continuous determination method within the interval is based on whether the second user's measurement data in the hip flexion interval F2 reaches 80% of the predicted value of the personalized training model. If they do not reach 80% of the predicted value of the personalized training model, this is considered a failure to meet the standard.

[0029] If they only achieve 50% of the personalized training model's predicted value, they are considered only a participant in gait training. If the second user completes five gait cycles in the personalized training model and four of them meet the personalized training model's standard, the control unit provides a replacement training model, increasing the difficulty level of the personalized training model. After the second user completes the gait training of the gait cycle five times, if four of the gait cycles do not meet the personalized training model's standard, the control unit will provide the replacement training model, reducing the difficulty level of the personalized training model.

[0030] In this preferred embodiment, as in Fig.3e, the 70% difficulty level of the personalized training model is used as an example of the substitute training model. In other preferred embodiments, the personalized training model is assessed and adjusted using the continuous determination method within the specific interval. According to the personalized training model, 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 80%, 90%, and 100% of the difficulty can be defined as the substitute training model. When assessing and adjusting the personalized training model using the one-point trigger determination within the specific interval, 20%, 40%, 60%, 80%, and 100% of the difficulty can be defined as the substitute training model according to the personalized training model.

[0031] Thereby, a personalized movement model associated with the second user according to the state of the second user can be planned by a method for real-time adjustment of gait training parameters 10 provided by the present invention, and suitable replacement training models can be recommended in training according to the data of the second user's output to achieve the effect of adjusting the training difficulty in real time according to the actual performance during training.

[0032] The above-mentioned preferred embodiments are intended to help understand the principles and methods of the present invention. The present invention is not limited to the above-mentioned preferred embodiments. All combinations and modifications that are within the scope and principle of the present invention are intended to be within the scope of the present invention.

Claims

[1] A method for real-time adaptation of gait training parameters (10) applicable to a gait training device (100), wherein the gait training device (100) comprises a detection unit, a training unit and a control unit, wherein the control unit is electrically connected to the detection unit and the training unit and controls the operation of the training unit, the method for real-time adaptation of gait training parameters (10) comprising the steps of: (a) Collecting, by the acquisition unit, the muscle relaxation gait data of at least one first user measured during gait training in a muscle relaxation state and the active power output gait data of the at least one first user measured during gait training in the active power output state of the first user; and then, creating, by the control unit, a standard motion model based on the ratio of the active power output gait data of the first user to the muscle relaxation gait data of the first user; (b) Obtaining a movement model of a second user comprising the muscle relaxation gait data of the second user measured during gait training under the muscle relaxation state of the second user, and then estimating a plurality of personalized training models of different difficulty levels by combining the muscle relaxation gait data of the second user with the standard movement model with 10%, 15%, 20%, 25%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100% difficulty curves, so that personalized training models with different difficulty levels are created, and the difficulty curves can be adjusted according to requirements by the control unit; and (c) determining whether the actual training status of the second user matches the standard of one of the personalized training models, and then adjusting the one of the personalized training models and providing a replacement training model by the control unit, wherein the gait training comprises at least one gait cycle and the gait cycle is divided into a center of gravity shift interval (F1), a hip flexion interval (F2) and a knee extension interval (F3), wherein, in step (b), the calculation method of the control unit for estimating the personalized training model from the muscle relaxation gait data of the second user comprises, when the muscle relaxation state of the second user is obtained, obtaining a maximum value of the force in a muscle relaxation state and a minimum value of the force in a muscle relaxation state in the gait cycle, whereupon, when the second user is actively outputting force, estimating a predicted maximum value of the active force output and a predicted minimum value of the active force output in the gait cycle of the second user by the standard movement model, whereupon, when the second user is in the active force output state, using the value obtained by the acquisition unit to obtain an actual maximum value of the active force output and an actual minimum value of the active force output in the gait cycle,When the second user actively exerts force, the actual maximum value of the active force output, the predicted maximum value of the active force output, and the maximum value of the force in the muscle relaxation state are input into a calculation formula for the center of gravity shift interval. Input the actual minimum value of the active force output, the predicted minimum value of the active force output, and the minimum value of the force in the muscle relaxation state into a calculation formula for the hip flexion interval to obtain the force output level of the center of gravity shift interval (F1) and the force output level of the hip flexion interval (F2), and then use the lower force output level than these personalized training models of different difficulty levels for the second user. [2] A method for real-time adjustment of gait training parameters (10) according to claim 1, wherein the detection unit comprises two knee pressure sensors and two sole force sensors; wherein, when in step (c) of the method for real-time adjustment of gait training parameters (10), the control unit determines that the second user reaches hip flexion in the hip flexion interval (F2), the conditions must be met: (XK>XK¯+R%×URKX)∩(PF<(UFP−UFR)×R%+UFP), where X K is the center of pressure position measured by one of the knee pressure sensors when the second user is in the active force output state, where XK¯ is the average center of pressure position measured by the knee pressure sensor when the second user is in a relaxed state, where U FR is the force value measured by one of the sole force sensors when the second user is in a relaxed state, where UFP is the personalized training model, where R% is the set difficulty. [3] A method for real-time adaptation of gait training parameters (10) according to claim 1, wherein the detection unit comprises two knee pressure sensors and two sole force sensors, wherein in step (c) of the method for real-time adaptation of gait training parameters (10), when the control unit determines that the second user reaches knee extension in the knee extension interval (F3), the conditions must be met: P K < 0.9 - 0.4 × R% × U RKP , where P K is the pressure value measured by one of the knee pressure sensors when the second user is in the active force output state, where U RKP is the pressure value measured by the knee pressure sensor when the second user's muscles are relaxed, where R% is the set difficulty. [4] A method for real-time adjustment of gait training parameters (10) according to claim 1, wherein in step (c) the method for judging whether the actual training state of the second user corresponds to the standard of the one of the personalized training models suitable for the user comprises a continuous determination path within a certain interval and a single-point trigger determination path within the certain interval. [5] A method for real-time adjustment of gait training parameters (10) according to claim 1, wherein in step (c), if the actual training state of the second user does not correspond to the standard of the one of the personalized training models suitable for the user, the control unit provides the replacement training model by reducing the difficulty level of the personalized training model. [6] A method for real-time adjustment of gait training parameters (10) according to claim 1, wherein in step (c), when the actual training state of the second user meets the standard of the one of the personalized training models suitable for the user, the control unit provides the replacement training model by increasing the difficulty of the personalized training model.

Images