TWI619488B - Lower extremity exoskeleton auxiliary device and method thereof - Google Patents

Abstract

A lower extremity exoskeleton assisting device is provided for at least one lower limb of a user. The lower extremity exoskeleton assisting device includes a muscle inductance measuring end, a knee joint angle sensing unit, a neural network computing unit, a lower extremity exoskeleton motor and a lower extremity assistive device. . The muscle inductance measuring terminal is used to measure the muscle electrical signal of the thigh. The knee angle sensing unit is used to measure the knee angle signal of the knee joint. The neural network calculation unit calculates the output value based on the electromyogram signal, knee joint angle signal, and muscle mechanical model, and the output value is converted into a lower extremity exoskeleton motor control signal. The lower extremity exoskeleton motor is adjacent to the knee joint and generates auxiliary torque according to the input of the lower extremity exoskeleton motor control signal. The lower limb assist is driven by an assisting torque to assist in supporting the lower limb.

Description

Lower extremity exoskeleton auxiliary device and method

The invention relates to an exoskeleton assisting device and a method thereof, and in particular to an exoskeleton assisting device and a method thereof applied to a lower limb of a human body.

As countries have stepped into an aging society, the social environment and life style have also changed, and the demand for medical care and rehabilitation has continued to rise. At the same time, it has also led to the development of medical care and medical equipment industries. For the elderly and people with disabilities, if there is a smart machine assist device such as the extremity exoskeleton assist device to assist the elderly and people with disabilities with some movement deficiency, although the speed cannot be as fast as normal people, but for the Medical rehabilitation or home activities can be very helpful.

However, in order to achieve the effect of supporting the lower limbs, the conventional lower extremity exoskeleton assisting device is usually equipped with multiple sensors, which makes the traditional lower extremity exoskeleton assisting device inconvenient and comfortable to wear and expensive. Therefore, the market is eager to develop a lower extremity exoskeleton assisting device and method that are convenient to wear and affordable, in order to increase the willingness to wear and further increase the penetration rate of wear.

The invention provides a lower extremity exoskeleton assisting device and a lower extremity exoskeleton assisting method. A neural network-like calculation unit is used to generate an assist torque for the lower extremity exoskeleton motor according to the myoelectric signal, the knee joint angle signal, and the muscle mechanical model. Driving the lower extremity assistive device can have the advantages of simple device and good support for the lower extremity, and can increase the user's willingness to wear.

According to the present invention, a lower extremity exoskeleton auxiliary device is provided for at least one lower limb of a user. The lower extremity exoskeleton auxiliary device includes a muscle inductance measuring end, a knee joint angle sensing unit, a neural network-like calculation unit, and a lower extremity exoskeleton motor. And lower limb assistive devices. The muscle inductance measuring end is arranged corresponding to the thigh of the lower limb, and the muscle inductance measuring end is used to measure the muscle electrical signal of the thigh. The knee joint angle sensing unit is set corresponding to the knee joint of the lower limb, and the knee joint angle sensing unit is used to measure the knee joint angle signal of the knee joint. The neural network-like calculation unit is electrically connected to the muscle inductance measuring end and the knee joint angle sensing unit. The neural network-like calculation unit calculates the output value based on the myoelectric signal, knee angle signal, and muscle mechanical model, and the output value is converted into Lower extremity exoskeleton motor control signal. The lower extremity exoskeleton motor is adjacent to the knee joint and is electrically connected to a neural network-like calculation unit. The lower extremity exoskeleton motor generates an auxiliary torque according to the input of the lower extremity exoskeleton motor control signal. The lower limb assist is driven by an assisting torque to assist in supporting the lower limb. Therefore, the lower extremity exoskeleton auxiliary device of the present invention can simplify the structure of the lower extremity exoskeleton auxiliary device and has a good effect of assisting the lower limbs.

According to the lower extremity exoskeleton auxiliary device described in the previous paragraph, the muscle mechanical model can be constructed according to the parameters of the Hill muscle model and the lower extremity exoskeleton motor. Stand. The muscle mechanical model can adjust the parameters of the muscle mechanical model according to the time varying information of the myoelectric signal and the time varying information of the knee joint angle signal to determine the action intention of the user's lower limbs. The lower extremity exoskeleton assisting device can be worn by two lower limbs of a user, in which the number of muscle inductance measuring ends, knee joint angle sensing units, lower extremity exoskeleton motors and lower limb assists are two and correspond to two lower limbs, respectively. The neural network-like calculation unit may include the establishment of a neural network, and the neural network includes an input layer, a first layer, a second layer, a third layer, and an output layer. The input layer is based on the musculature signals of the two lower limbs and the angle signals of the two knee joints as variables to input fuzzy rules. In the first layer, the fitness rules of fuzzy rules are calculated to obtain the strength of complex rules. The second layer is a normalization operation of the rule strength to obtain a complex normalization value. The third layer multiplies the normalized value with the Sugeno fuzzy mode. The output layer is the sum of the operation results of the third layer to obtain the output value. The knee joint angle sensing unit may include an encoder, which is connected to the lower extremity exoskeleton motor, and the encoder converts the knee joint angle signal from an analog form to a digital form. With the technical features of the points mentioned above, it helps the lower extremity exoskeleton motor to assist in supporting the lower limb in a more timely manner.

According to the present invention, a lower limb exoskeleton assisting method is provided, which provides assisting torque to at least the lower limbs of the user. The lower extremity exoskeleton assisting method includes a muscle inductance measurement step, a knee angle sensing step, a neural network-like calculation step, and a lower limb Exoskeleton motor actuation step and lower limb assistive device step. The muscle inductance measurement step measures the muscle electrical signals of the thighs of the lower limbs. The knee angle sensing step measures the knee angle signal of the lower limb. The neural network-like calculation step is based on the electromyogram signal, knee joint angle signal, and muscle mechanical model to calculate the output value, and convert the output value to the lower extremity exoskeleton motor control signal. under The extremity exoskeleton motor actuation step is based on the input of the lower extremity exoskeleton motor control signal to generate auxiliary torque. The step of driving the lower limb assistive device is that the lower limb assistive device is driven by an assisting torque to assist in supporting the lower limb. Therefore, the method for assisting lower extremity exoskeleton according to the present invention can simplify the method for assisting lower extremity exoskeleton and has a good effect of assisting lower extremity.

According to the lower extremity exoskeleton assistance method described in the previous paragraph, the muscle mechanical model can be established based on the Hill muscle model and the parameters of the lower extremity exoskeleton motor. The lower extremity exoskeleton assist method can provide two auxiliary torques to the user's two lower limbs, respectively. The neural network-like calculation step may include the establishment of a neural-like network. The neural-like network includes an input layer, a first layer, a second layer, a third layer, and an output layer. The input layer is based on the musculature signals of the two lower limbs and the angle signals of the two knee joints as variables to input fuzzy rules. In the first layer, the fitness rules of fuzzy rules are calculated to obtain the strength of complex rules. The second layer is a normalization operation of the rule strength to obtain a complex normalization value. The third layer multiplies the normalized value with the Sugeno fuzzy mode. The output layer is the sum of the operation results of the third layer to obtain the output value. With the technical features of the points mentioned above, the accuracy of the lower extremity exoskeleton assisting method of the present invention to assist in supporting lower limbs can be further improved.

40‧‧‧ users

51, 52‧‧‧ lower limbs

61, 62‧‧‧ thighs

71, 72‧‧‧ Knee

100, 200‧‧‧ Lower limb exoskeleton assist device

131, 231, 232‧‧‧ muscle inductance test terminal

141, 241, 242‧‧‧Knee joint angle sensing unit

150, 250‧‧‧ neural network calculation units

161, 261, 262‧‧‧ Lower limb exoskeleton motor

171, 271, 272‧‧‧ Lower limb assistive devices

480‧‧‧Hill muscle model

600‧‧‧ class neural network

605‧‧‧input layer

610‧‧‧First floor

620‧‧‧Second floor

630‧‧‧third floor

640‧‧‧output layer

A1, A2‧‧‧myoelectric signals

B1, B2‧‧‧‧ Knee joint angle signal

Q _0,1 , Q _0,2 , Q _0,3 , Q _0,4 , Q _1,1 , Q _1,2 , Q _2,1 , Q _2,2 , Q _3,1 , Q _3,2 ‧ ‧ Operation Value

Q _output ‧‧‧ output value

800‧‧‧ Lower extremity exoskeleton assistance method

810, ‧‧‧ muscle inductance measurement steps

820‧‧‧Knee Joint Angle Sensing Procedure

830‧‧‧type neural network calculation steps

840‧‧‧ Lower limb exoskeleton motor operation steps

850‧‧‧ Steps for driving assistive device of lower limb

l ^m ‧‧‧length of muscle fiber

φ‧‧‧Knee joint angle

CE‧‧‧Contraction element

F _A ‧‧‧ The main power generated by the shrinking element

PE‧‧‧Elastic element

F _P ‧‧‧ Passive force generated by elastic element

The combined force of F ^m ‧‧‧ F _A and F _P ,

F ^mt ‧‧‧F ^m component force in the direction of knee joint angle

FIG. 1 is a block diagram of a lower extremity exoskeleton assisting device according to an embodiment of the present invention; FIG. 2 is a block diagram of a lower extremity exoskeleton assisting device according to an embodiment of the present invention; Fig. 3 is a schematic diagram showing the use of the lower extremity exoskeleton assisting device of Fig. 2; Fig. 4 is a diagram showing the Hill muscle model in the lower extremity exoskeleton assisting device of Fig. 2; A neural-like network in a lower extremity exoskeleton assistance device; and FIG. 6 is a flowchart illustrating a lower extremity exoskeleton assistance method according to another embodiment of the present invention.

With reference to FIG. 1, FIG. 1 is a block diagram illustrating a lower extremity exoskeleton assisting device 100 according to an embodiment of the present invention. It can be seen from FIG. 1 that the lower extremity exoskeleton assisting device 100 is provided for at least one lower limb of a user. The lower extremity exoskeleton assisting device 100 includes a muscle inductance measuring end 131, a knee joint angle sensing unit 141, a neural network-like computing unit 150, Lower extremity exoskeleton motor 161 and lower limb assistive device 171.

With reference to FIG. 2 and FIG. 3, FIG. 2 is a block diagram showing a lower extremity exoskeleton assisting device 200 according to an embodiment of the present invention, and FIG. 3 is a lower extremity exoskeleton assisting device of FIG. 2 A schematic diagram of the use of the device 200. As can be seen from FIG. 2 and FIG. 3, the lower extremity exoskeleton assisting device 200 of this embodiment is intended to be worn by the user 40 bis lower limbs 51 and 52. The lower extremity exoskeleton assisting device 200 includes muscle inductance measuring ends 231 and 232, and the knee angle Sensing units 241, 242, neural network calculation unit 250, lower extremity exoskeleton motors 261, 262, and lower extremity assist devices 271, 272, of which the muscle inductance measuring end 231, knee joint angle sensing unit 241, lower extremity exoskeleton motor 261 Lower limb assists 271 Corresponding to the lower limb 51, the muscle inductance measuring end 232, the knee joint angle sensing unit 242, the lower extremity exoskeleton motor 262, and the lower limb assisting device 272 correspond to the lower limb 52.

The muscle inductance measuring terminal 231 is provided corresponding to the thigh 61 of the lower limb 51, and the muscle inductance measuring terminal 231 is used to measure the electromyographic signal (EMG signal) A1 of the thigh 61, and the muscle inductance measuring terminal 232 is arranged corresponding to the thigh 62 of the lower limb 52 The muscle inductance measuring terminal 232 is used to measure the muscle electrical signal A2 of the thigh 62. The muscle inductance measuring terminals 231 and 232 can be non-invasive electrode patches and can be connected to an amplifier. The muscle inductance measuring terminals 231 and 232 can be respectively Set on the thighs 61 and 62 where there are the most muscle fibers, such as the thigh quadriceps, to improve the strength and accuracy of the EMG signals A1 and A2, and the frequency domain sampling range of the EMG signals A1 and A2 can be 10Hz ~ 1000Hz.

The knee joint angle sensing unit 241 is provided corresponding to the knee joint 71 of the lower limb 51, and the knee joint angle sensing unit 241 is used to measure the knee angle signal B1 of the knee joint 71. The knee joint angle sensing unit 242 corresponds to the knee of the lower limb 52 The joint 72 is set, and the knee joint angle sensing unit 242 is used to measure the knee joint angle signal B2 of the knee joint 72.

The neural network-like calculation unit 250 is electrically connected to the muscle inductance measuring terminals 231 and 232 and the knee joint angle sensing units 241 and 242. The neural network-like calculation unit 250 is based on the aforementioned myoelectric signals A1, A2, and knee joint angle signals B1 , B2 and the muscle mechanical model, the output value Q _{output is} calculated, and the output value Q _{output is} converted into a control signal of the exoskeleton motor of the lower limbs.

The lower extremity exoskeleton motor 261 is adjacent to the knee joint 71 and is electrically connected to a neural network calculation unit 250. The lower extremity exoskeleton motor 261 generates auxiliary torque according to one of the input lower extremity exoskeleton motor control signals; the lower extremity exoskeleton motor 262 is adjacent to the knee joint 72 and is electrically connected to a neural network calculation unit 250. The lower extremity exoskeleton motor 262 generates an assist torque according to the input of another lower extremity exoskeleton motor control signal.

The lower limb assist 271 is driven by an assisting torque to assist in supporting the lower limb 51, and the lower limb assist 272 is driven by the assisting torque to assist in supporting the lower limb 52. Therefore, according to the lower extremity exoskeleton assisting device 200 of the present invention, the lower extremity exoskeleton motors 261 and 262 are made by the neural network calculation unit 250 according to the myoelectric signals A1, A2, knee joint angle signals B1, B2, and the muscle mechanical model. Generates auxiliary torque to drive the lower limb assists 271 and 272 respectively, which can simplify the installation position and number of sensors of the lower limb exoskeleton assist device 200, and has the advantages of supporting the lower limbs 51 and 52 well to improve the wear of the user 40 Will. In addition, the lower extremity exoskeleton motors 261 and 262 can be connected to a speed reducer (not shown in the figure), so that the user 40 feels smooth and comfortable about the assist torque provided by the lower extremity exoskeleton assist device 200.

With reference to FIG. 4, FIG. 4 shows a Hill muscle model 480 in the lower extremity exoskeleton assistance device 200 of FIG. 2. As can be seen from FIG. 4, in this embodiment, the muscle mechanical model can be established based on the parameters of the Hill muscle model 480 and the lower extremity exoskeleton motors 261 and 262, thereby improving the lower extremity exoskeleton motors 261 and 262 to assist in supporting the lower extremities 51 and 52 Accuracy. Among the parameters of the Hill muscle model 480 in Fig. 4, l ^m represents the length of muscle muscle fibers, φ represents the knee joint angle, F _A represents the main power generated by the contraction element CE, and F _P represents the passive power generated by the elastic element PE. , F ^m represents the combined force of F _A and F _P , and F ^mt represents the component force of F ^m in the direction of the knee angle φ.

The muscle mechanical model can use the time-varying information of the myoelectric signals A1 and A2 and the time-varying information of the knee angle signals B1 and B2 to determine the action intentions of the lower limbs 51 and 52 of the user 40, such as standing, sitting, and going up stairs , Going down the stairs, etc., and adjusting the parameters of the muscle mechanical model to help the lower extremity exoskeleton motors 261, 262 assist in supporting the lower limbs 51, 52 in a more timely manner.

With reference to FIG. 5, FIG. 5 illustrates a neural network 600 in the lower extremity exoskeleton assisting device 200 of FIG. 2. It can be seen from FIG. 5 that the neural-like network calculation unit 250 may include a neural-like network 600 to be established. The neural-like network 600 may be an Adaptive-Network-based Fuzzy Inference System (ANFIS) and include The input layer 605, the first layer 610, the second layer 620, the third layer 630, and the output layer 640. The input layer 605 is the electromyogram signals A1, A2 of the lower limbs 51, 52 and the knee joint angle signals B1, B2 as variable input fuzzy rules. The first layer 610 performs the fitness calculation of the fuzzy rules to obtain the strength of the complex rules. The second layer 620 performs a normalization operation on the rule strength to obtain a complex normalization value. The third layer 630 multiplies the normalized value and the Sugeno fuzzy mode. The output layer 640 is obtained by performing a total operation on the operation results of the third layer 630 to obtain an output value Q _output . This can further improve the accuracy of the lower extremity exoskeleton motors 261, 262 to assist in supporting the lower limbs 51, 52.

In this embodiment, the neural network 600 described in the previous paragraph is based on the following formulas (1) to (6). Referring to the following formulas (1) and (2), the input layer 605 is the electromyogram signals A1, A2 of the lower limbs 51, 52 and the knee angle signals B1, B2 as variables. The fuzzy rule is input to obtain the calculation value Q _{0, i} . Where x and y represent the myoelectric signals A1, A2, and knee joint angle signals B1, B2, respectively, and their membership functions are Gaussian functions u _Ai and u _Bi , respectively, and the learning parameters are a _i , b _i , c _i , a _{i- 2} , b _i-2 and c _i-2 :

Referring to the following formula (3), the first layer 610 performs a fitness operation on the fuzzy rules (ie, w _i ) to obtain the strength of the complex rule, which is the operation value Q _{1, i} : , Where i = 1, 2, ... Equation (3).

Referring to the following formula (4), the second layer 620 performs a normalization operation on the rule strength to obtain a complex normalization value, that is, an operation value Q _{2, i} , where the operation value Q _{2, i is} between 0 and 1:

Referring to the following formula (5), the third layer 630 multiplies the normalized value and the Sugeno fuzzy mode (that is, f _i ) to obtain an operation value Q _{3, i} , where p _i , q _i and r _i are Sugeno fuzzy modes. The parameters:

Referring to the following formula (6), the output layer 640 calculates the output value Q _{output of the} neural network 600 by summing the operation values Q _{3, i} of the third layer 630:

The knee joint angle sensing units 241 and 242 may each include an encoder (not shown in the figure). One of the encoders is connected to the lower extremity exoskeleton motor 261 and the other is connected to the lower extremity exoskeleton motor 262. The encoder converts the knee joint. The angle signals B1 and B2 are changed from an analog form to a digital form to help reduce the complexity of the circuit design of the lower extremity exoskeleton assist device 200 and maintain the auxiliary support effect. In other embodiments, the knee joint angle sensing unit may include a Hall effect sensor (Hall Effect Sensor) or a gyroscope.

With reference to FIG. 6, FIG. 6 is a flowchart illustrating a lower extremity exoskeleton assistance method 800 according to another embodiment of the present invention. It can be seen from FIG. 6 that the lower extremity exoskeleton assistance method 800 provides assisting torque to at least the lower limbs of the user. The lower extremity exoskeleton assistance method 800 includes a muscle inductance measurement step 810, a knee joint angle sensing step 820, and a neural network-like calculation step. 830. The lower extremity exoskeleton motor actuating step 840 and the lower extremity assisting device step 850 are described in detail in FIG. 6.

In an example of another embodiment of the present invention, please refer to FIG. 2, FIG. 3 and FIG. 6 together. The lower extremity exoskeleton assisting method 800 can provide two assisting torques to the user 40 two lower limbs 51 and 52, respectively. . Further, the step 810 of the muscle inductance measurement measures the muscle signals A1 and A2 of the thighs 61 and 62 of the lower limbs 51 and 52, respectively. The knee joint angle sensing step 820 measures the knee joint angle signals B1 and B2 of the lower limbs 51 and 52, respectively. The neural network-like calculation step 830 is based on the aforementioned myoelectric signals A1, A2, knee joint angle signals B1, B2, and a muscle mechanical model, and calculates an output value Q _output , and converts the output value Q _output into a two-limb exoskeleton motor Control signal. The lower extremity exoskeleton motor actuation step 840 is that the lower extremity exoskeleton motors 261 and 262 generate an assist torque according to the input corresponding lower extremity exoskeleton motor control signal. Step 850 of driving the lower limbs assistive device 850, the lower limbs assistive devices 271, 272 are driven by corresponding assisting torques to assist in supporting the lower limbs 51, 52.

Please refer to FIG. 4 and FIG. 5 together. The muscle mechanical model of the lower extremity exoskeleton assistance method 800 can be established based on the parameters of the Hill muscle model 480 and the lower extremity exoskeleton motors 261 and 262. The neural-like network calculation step 830 may include the establishment of a neural-like network 600. The neural-like network 600 includes an input layer 605, a first layer 610, a second layer 620, a third layer 630, and an output layer 640. The input layer 605 is the electromyogram signals A1, A2 of the lower limbs 51, 52 and the knee joint angle signals B1, B2 as variable input fuzzy rules. The first layer 610 performs the fitness calculation of the fuzzy rules to obtain the strength of the complex rules. The second layer 620 performs a normalization operation on the rule strength to obtain a complex normalization value. The third layer 630 multiplies the normalized value and the Sugeno fuzzy mode. The output layer 640 is obtained by performing a total operation on the operation results of the third layer 630 to obtain an output value Q _output . Other details of the lower extremity exoskeleton assisting method 800 according to an embodiment of another embodiment of the present invention are the same as those of the lower extremity exoskeleton assisting device 200 according to an embodiment of the present invention, and are not described herein.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Any person skilled in the art can make various modifications and retouches without departing from the spirit and scope of the present invention. The scope shall be determined by the scope of the attached patent application.

Claims (6)

A lower extremity exoskeleton auxiliary device, which is worn by one user's two lower extremities, the lower extremity exoskeleton auxiliary device comprises: two muscle inductance measuring ends, each of the muscle inductance measuring ends corresponding to a thigh of each lower limb, and each muscle The inductive measuring end is used to measure a myoelectric signal of the thigh; two knee joint angle sensing units, each knee joint angle sensing unit is corresponding to a knee joint of each lower limb, and each knee joint angle sensing unit Used to measure a knee joint angle signal of one of the knee joints; a type of neural network calculation unit electrically connected to the biceps inductance measuring end and the two knee joint angle sensing unit, and the type of neural network calculation unit is based on the Two muscle electric signals, two knee joint angle signals and a muscle mechanical model are calculated to obtain an output value, which is converted into two lower limb exoskeleton motor control signals; two lower limb exoskeleton motors, each of which is adjacent to each other The knee joint is electrically connected to the neural network calculation unit, and each of the lower extremity exoskeleton motors generates an assist torque according to the input of each of the lower extremity exoskeleton motor control signals; and Lower limb assists, each lower limb assist is driven by each auxiliary torque to assist in supporting each lower limb; wherein the neural network calculation unit includes a type of neural network establishment, and the type of neural network includes: an input layer, A fuzzy rule is inputted using the biceps signal of the two lower limbs and the angle signal of the knee joint as variables; a first layer performs a fitness calculation on the fuzzy rule to obtain the strength of the complex rule; A second layer that performs normalization operations on the strength of the rules to obtain complex normalization values; a third layer that performs the operation of multiplying the normalization values with the Sugeno fuzzy mode; and an output layer that performs the third layer The result of the operation is summed to obtain the output value. The lower extremity exoskeleton assisting device according to item 1 of the patent application scope, wherein the muscle mechanical model is established based on the Hill muscle model and the parameters of the lower extremity exoskeleton motor. The lower extremity exoskeleton assistance device according to item 2 of the scope of patent application, wherein the muscle mechanical model is based on the time-varying information of the myoelectric signal and the time-varying information of the knee angle signal to determine the user's An action intention, and adjust the parameters of the muscle mechanical model. The lower extremity exoskeleton assisting device according to item 1 of the patent application scope, wherein the knee joint angle sensing unit includes an encoder, the encoder is connected to the lower extremity exoskeleton motor, and the encoder converts the knee joint angle signal From analog form to digital form. A method for assisting lower extremity exoskeleton, which provides two assist torques to two lower limbs of a user, respectively. The method for assisting lower extremity exoskeleton includes: A muscle inductance measurement step to measure a muscle electrical signal of each thigh of the lower limb; a knee angle sensing step to measure a knee angle signal of each of the lower limbs; a class of neural network calculation steps based on the two An electromyogram signal, the two knee joint angle signal and a muscle mechanical model are calculated to obtain an output value, and the output value is converted into two lower extremity exoskeleton motor control signals; the lower extremity exoskeleton motor operation steps, two lower extremity exoskeleton motors Each of the lower extremity exoskeleton motor control signals generates an assist torque; and a step of driving the lower limb assist device, each of the two lower limb assist devices is driven by each of the assist torque to assist in supporting each of the lower limb; The neural network-like calculation step includes the establishment of a type of neural network. The type of neural network includes: an input layer, using the two muscle signals of the two lower limbs and the angle signal of the knee joint as variables to input a fuzzy rule; In the first layer, the fitness of the fuzzy rules is calculated to obtain the strength of the plural rules; in the second layer, the intensity of the rules is normalized to obtain Number normalized value; a third layer, some of the normalized value and the multiplication Sugeno fuzzy pattern; and an output layer, the calculated result of the third layer is the sum of the calculation to obtain the output value. The method for assisting the lower extremity exoskeleton as described in item 5 of the scope of the patent application, wherein the muscle mechanical model is established according to the Hill muscle model and the parameters of the lower extremity exoskeleton motor.