Exoskeleton control method based on man-machine dynamics model and lower limb exoskeleton device

Abstract

The invention relates to an exoskeleton control method based on a man-machine dynamics model and a lower limb exoskeleton device. The method comprises the following steps: selecting a program mode of an exoskeleton; in the following mode, joint movement data and active torque of a wearer are collected to construct a database; in the power assisting mode, a human leg-trouser body-exoskeleton dynamic model is established based on the database, and online parameter identification is carried out; kinematics parameters of all links are calculated through forward kinematics; establishing a system state space equation by combining dynamics and kinematics parameters; judging the real-time motion state of the wearer by using a deep learning neural network, and generating a corresponding expected control rule; and the exoskeleton is driven and cyclically executed, so that self-adaptive control in continuous motion is realized. The exoskeleton device comprises a thigh supporting part, a shank supporting part and a flexible buffering belt which are combined with the pants body, and dynamic fitting and force transmission between a man and a machine are achieved. High precision of the exoskeleton is achieved, and the naturalness and efficiency of assistance are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, in particular to an exoskeleton control method based on human-machine dynamics model and lower limb exoskeleton device. BACKGROUND

[0002] As a wearable device, exoskeleton robots aim to enhance or assist human movement ability, and have shown great potential in the fields of rehabilitation medicine, military, industrial handling, etc. However, the control system of existing exoskeleton robots still faces many challenges.

[0003] The existing control method of exoskeleton robots usually adopts preset trajectory control or impedance control based on force sensors, which requires hard specification of the motion trajectory of the exoskeleton, and cannot adapt to individual differences and actual movement intentions of the wearer in real time, resulting in poor coordination between man and machine, and poor comfort and naturalness of wearing. For example, when the wearer needs to temporarily change the movement direction or speed, the preset trajectory control cannot respond in time, which may cause safety hazards or additional energy consumption In addition, when the traditional exoskeleton is connected with the wearer's leg, it is often fixed by a simple strap. During the movement process, the friction and pressure between the strap and the skin may cause discomfort, and even affect blood circulation. Moreover, when the exoskeleton main body moves relative to the wearer, the force transmission efficiency is low, so that the assistance of the exoskeleton cannot be fully and effectively applied to the human body.

[0004] Therefore, we propose an exoskeleton control method based on human-machine dynamics model and lower limb exoskeleton device. SUMMARY

[0005] The main purpose of the present application is to solve the deficiencies of the existing exoskeleton control method based on human-machine dynamics model and lower limb exoskeleton device in real-time, adaptability and wearing comfort.

[0006] To achieve the above purpose, the present application provides an exoskeleton control method based on human-machine dynamics model, comprising the following steps: S1, selecting a program mode of the exoskeleton, the mode including a following mode and an assistance mode; S2, when the exoskeleton is in the following mode, automatically acquiring joint motion data and corresponding joint active torque when the wearer performs standard target movement, and constructing a model database including standard kinematics and dynamics parameters; when the exoskeleton is in the assistance mode, automatically collecting joint motion data and joint active torque under the current movement state of the wearer, and matching the pre-stored model parameters corresponding to the current movement state based on the model database, to initialize and establish a human leg-exoskeleton dynamics model, and estimate and update the model parameters by using an online parameter identification algorithm to obtain the estimated value of the dynamics model parameters; S3, obtaining the center of mass position, velocity and acceleration of each link of the lower limbs by forward kinematics calculation based on the joint active torque, joint motion data and leg information of the wearer obtained in the above steps, and obtaining the kinematics model parameter estimation value in combination with a preset human body link inertia parameter model; S4, establishing a state space equation of the exoskeleton system based on the dynamics model parameter estimation value and the kinematics model parameter estimation value; S5, judging the current motion state of the wearer based on the real-time collected motion data of the wearer according to a deep learning neural network model, and generating an exoskeleton expected control law corresponding to the current motion state in combination with the state space equation of the exoskeleton system; S6, inputting the exoskeleton expected control law to a driving device of the exoskeleton to provide motion assistance to the wearer, and repeatedly executing steps S2 to S6 to realize adaptive assistance control in a continuous motion process.

[0007] Preferably, the human leg-pant body-exoskeleton dynamics model comprises: ; wherein, is a knee joint motion angle, is a knee joint angular velocity, is a knee joint angular acceleration, is an inertia matrix, is a centrifugal force and Coriolis force matrix (Coriolis force), is a gravity matrix, is a friction and interference force of the pant body on the exoskeleton, is an auxiliary torque of the motor, is an active torque of the human knee joint.

[0008] Preferably, in step S2, the construction of the database sample is specifically: obtaining joint motion data by an encoder and an inertial measurement unit arranged at an exoskeleton joint, and obtaining an electromyographic signal by a surface electromyographic sensor arranged at a corresponding muscle group of the lower limbs of the wearer, and estimating an active torque of the joint after processing; after time alignment and normalization processing of the joint motion data and the estimated active torque of the joint collected during multiple executions of the target motion, storing as a database sample.

[0009] Preferably, in step S2, the online parameter identification algorithm adopts a recursive least squares method or an extended Kalman filter algorithm.

[0010] Preferably, in step S3, the forward kinematics calculation is based on a standard Denavit-Hartenberg parameter method to establish a lower limb kinematics model, for calculating the poses of the thigh, shank and foot links of the lower limb in an inertial coordinate system according to the joint motion data, and using rigid body kinematics formulas to calculate the positions, linear velocities and linear accelerations of the mass centers of the links.

[0011] Preferably, in step S4, the deep learning neural network model is a classification or regression model pre-trained by data in multiple user and multiple motion modes, for identifying at least one of the motion states of walking, ascending and descending stairs, and squatting up.

[0012] To achieve the above object, the application provides a lower limb exoskeleton device based on a man-machine dynamics model, which comprises a set of symmetrically arranged leg exoskeleton units and a matching trousers body. The shank supporting part comprises a fixed connecting rod arranged below the exoskeleton main body, and the fixed connecting rod is connected with two shank binding devices at both ends, and the shank binding devices are embedded in the shank part of the trousers body of the wearer. The region of the trousers body corresponding to the thigh supporting part and the shank supporting part is provided with a flexible buffer belt, and the flexible buffer belt is located on the inner side of the trousers body, so as to form a flexible interface between the exoskeleton main body and the leg of the wearer.

[0013] Preferably, the trousers body is provided with a reinforcing structure, and the reinforcing structure comprises a waistband and a plurality of reinforcing ribs connecting the waistband and the thigh supporting part, so as to transmit the weight of the exoskeleton device to the waist of the wearer.

[0014] Preferably, the flexible buffer belt is a carbon fiber sheet, the outside of which is wrapped with silica gel, and the inside of which is provided with a magnet, so as to realize magnetic attraction connection with the exoskeleton main body, and the flexible buffer belt is spliced by a plurality of movable modules.

[0015] The technical scheme of the application has the following beneficial effects: By collecting a large amount of human-machine motion data in the "follow mode" to construct a database sample, and based on this sample, establishing a human leg-pant body-exoskeleton dynamics model in the "assistance mode" and using an online parameter identification algorithm to obtain the model parameter estimate value, the dynamics of the human-machine coupling system is accurately described and the online adaptive update of the parameters is realized, so that the exoskeleton can more accurately understand the motion intention and body state of the wearer.

[0016] Combined with the parameter estimate value of the dynamics model and the parameter estimate value of the kinematics model, the state space equation of the exoskeleton system is established, and further a deep learning neural network model is used to judge the current motion state of the wearer, so as to generate the corresponding exoskeleton expected control law, so that the exoskeleton can provide personalized assistance according to the real-time needs of the wearer in a complex and variable motion environment, significantly improving the assistance accuracy and the adaptability of the system to different motion modes.

[0017] By embedding the thigh support part and the calf support part into a specially designed pant body, and cooperating with the pant body through the main binding belt, the auxiliary binding belt and the calf binding device, the exoskeleton is more closely combined with the human body. In particular, the design of the flexible buffer belt forms a flexible interface between the human and the machine, improving the wearing comfort and ensuring the effective transmission of power. Not only does it reduce the discomfort and local compression that may be caused by traditional binding belts, but it also improves the transmission efficiency of exoskeleton assistance to the human body, making the assistance effect more significant and natural.

[0018] By setting a reinforcing structure, including a waist belt and multiple reinforcing ribs connecting the waist belt and the thigh support part, the weight of the exoskeleton device is effectively transmitted to the wearer's waist instead of being concentrated on the legs, significantly reducing the weight burden on the lower limbs and improving the comfort of long-term wearing.

[0019] The flexible buffer belt is wrapped with silicone inside a carbon fiber sheet main body, with magnets inside connected to the exoskeleton main body by magnetic attraction, and is composed of multiple movable modules. The flexible buffer belt can better adapt to the curve changes of the human leg, improving the fit and wearing experience, and facilitating disassembly and maintenance. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 FIG. 1 is a structural schematic diagram of a lower limb exoskeleton device in an embodiment of the present application; Figure 2 FIG. 2 is an installation schematic diagram of a lower limb exoskeleton device in an embodiment of the present application; Figure 3 FIG. 3 is a structural schematic diagram of a pant body in an embodiment of the present application; Figure 4 FIG. 4 is a reinforcing structure schematic diagram of a pant body in an embodiment of the present application.

[0021] REFERENCE SIGNS: 100 - exoskeleton body, 110 - thigh support part, 111 - main strap, 112 - sub strap, 120 - calf support part, 121 - fixed connecting rod, 122 - calf binding device, 200 - pants body, 210 - flexible buffer belt, 220 - reinforcing structure, 221 - waistband, 222 - reinforcing rib. DETAILED DESCRIPTION

[0022] The embodiments of the present application are described in detail below with reference to the accompanying drawings. The embodiments described below are examples for explaining the present application and are not intended to limit the present application, and all other embodiments obtained by those skilled in the art based on the embodiments in the present application without creative labor fall within the scope of the present application.

[0023] In addition, if the description of "first", "second" and the like in the present application is only for the purpose of description, such as for distinguishing the same or similar elements, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implying the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one feature. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the fact that the technical solutions can be realized by those skilled in the art, and when the combination of technical solutions appears contradictory or cannot be realized, it should be considered that the combination of technical solutions in the present embodiment does not exist, nor within the scope of protection required by the present application.

[0024] The present application provides an exoskeleton control method based on human-machine dynamics model, comprising the following steps: S1, selecting a program mode of the exoskeleton, the mode including a following mode and an assisting mode.

[0025] When it is necessary to collect individual characteristic data of the wearer, learn the motion mode or establish the dynamics model under a specific task, the exoskeleton applies the following mode. In this mode, the exoskeleton mainly serves as a data collection platform and provides no or minimal intervention assistance.

[0026] When it is necessary to provide actual motion assistance for the wearer, the exoskeleton applies the assisting mode. In this mode, the exoskeleton will actively output assistance according to the predicted wearer motion intention and dynamics model. For example, when a new user uses it for the first time or the user performs a new training task, the system can automatically enter the following mode to complete data collection and model establishment; in daily walking or repetitive tasks, the system enters the assisting mode to provide continuous assistance.

[0027] S2, when the exoskeleton is in the following mode, automatically obtaining joint motion data and corresponding joint active torque when the wearer performs a standard target motion, and constructing a model database including standard kinematics and dynamics parameters; when the exoskeleton is in the assistance mode, automatically collecting joint motion data and joint active torque under the current motion state of the wearer, and based on the model database, matching the pre-stored model parameters corresponding to the current motion state, to initialize and establish the human leg-exoskeleton dynamics model, and using an online parameter identification algorithm to estimate and update the model parameters to obtain the estimated value of the dynamics model parameters.

[0028] Specifically, when the exoskeleton is in the following mode, the real-time “joint motion data” (including but not limited to joint angle, angular velocity, angular acceleration) and “corresponding joint active torque” of the wearer when performing a specific “target motion” (for example, walking, going up and down stairs, squatting, etc.) are automatically obtained by the built-in sensors (such as joint encoders, inertial measurement units (IMU), etc.) or additional sensors (such as surface electromyography sensors) integrated with the exoskeleton, preprocessed, and formed into a “database sample” for the wearer under a specific motion mode. The database sample completely records the natural motion characteristics and biomechanical responses of the wearer without significant intervention of the exoskeleton.

[0029] When the exoskeleton is in the assistance mode, the “joint motion data” and “joint active torque” of the wearer are collected in real time. At this time, the exoskeleton will establish a “human leg-pant body-exoskeleton dynamics model” for the current wearer and the current motion state based on the “database sample” constructed in the following mode.

[0030] S3, according to the joint active torque, joint motion data and leg information of the wearer obtained in the above steps, the mass center position, velocity and acceleration of each link of the lower limb are calculated through forward kinematics, and the kinematics model parameter estimation value is obtained in combination with a pre-set human body link inertia parameter model. S4, based on the dynamics model parameter estimation value and the kinematics model parameter estimation value, a state space equation of the exoskeleton system is established; S5, based on a deep learning neural network model, the current motion state of the wearer is judged according to the real-time collected motion data of the wearer, and the exoskeleton expected control law corresponding to the current motion state is generated in combination with the state space equation of the exoskeleton system; S6, the exoskeleton expected control law is input to the driving device of the exoskeleton to assist the motion of the wearer, and steps S2 to S6 are repeatedly executed to realize adaptive assistance control in a continuous motion process.

[0031] In one of the embodiments, the forward kinematics calculation is based on a standard Denavit-Hartenberg parameter method to establish a lower limb kinematics model, for calculating the poses of the thigh, shank and foot segments of the lower limb in an inertial coordinate system according to the joint motion data, and using rigid body kinematics formulas to calculate the positions, linear velocities and linear accelerations of the mass centers of the segments.

[0032] In one of the embodiments, the deep learning neural network model is a classification or regression model pre-trained by data in multiple users and multiple motion modes, for identifying at least one of the motion states of walking, ascending or descending stairs, and squatting.

[0033] In one of the embodiments, in step S2, the online parameter identification algorithm uses a recursive least squares method or an extended Kalman filter algorithm.

[0034] In one of the embodiments, the human leg-pant-exoskeleton dynamics model, different from a conventional exoskeleton man-machine model, takes into account the influence of the pant on the exoskeleton. Although the pant itself does not change the resultant force of the whole system, it changes the distribution of force on the local interface, and affects the skin local pressure and slip, etc. According to the Lagrange equation, we have: ; wherein, is the knee joint motion angle, is the knee joint angular velocity, is the knee joint angular acceleration, is the inertia matrix, is the centrifugal force and Coriolis force matrix (Coriolis force), is the gravity matrix, is the friction and the interference force of the pant on the exoskeleton, is the auxiliary torque of the motor, is the active torque of the human knee joint.

[0035] The force analysis is as follows: The force analysis when standing, the overall analysis of the exoskeleton is subjected to the pant tension, friction and its own gravity, while the bandage itself has elasticity, and the bandage tension meets Hooke's law. When standing, the friction on each part of the bandage is relatively uniform, and the calculation formula is as follows: ; ; wherein, T is the bandage tension, k is the elastic coefficient, which depends on the material, is the change in the length of the bandage, w is the width of the bandage, R is the leg radius, and p is the average pressure on the leg, is the total pressure of the bandage on the leg, The tensile force of the exoskeleton to the trousers, The friction force, The weight of the exoskeleton, the weight of one side of a single leg is about 1.5-3 kg, The friction coefficient.

[0036] Further, when moving, the exoskeleton is subjected to the tensile force of the trousers, the friction force and the gravity of the exoskeleton in addition to the output torque of the motor, because the torque of the motor is ultimately transmitted to the bandage to act on the leg, so the total pressure of the bandage on the leg is the pressure of the bandage on the leg at rest plus the force of the exoskeleton to the bandage, because the output force of the exoskeleton body is not only to the bandage, but also to the flexible belt module, the right side of the bandage is subjected to a larger positive pressure, and the friction force is also larger, and vice versa, so the friction force should be the sum of each place.

[0037] When moving, the bandage is also subjected to the transverse force of the exoskeleton, which is specifically manifested as a non-uniform pressure distribution on the bandage, at this time, the pressure on the leg is: ; ; Wherein, The unit pressure of the bandage on the leg is related to its position, and the force analysis of the bandage, the trousers and the exoskeleton shows that ; Wherein, The force of the exoskeleton to the bandage, The proportionality coefficient, The total pressure on the leg, The tangential force of the exoskeleton on the bandage, The length of the thigh support plate.

[0038] Through analysis, it is found that when moving, the tensile force of the exoskeleton to the trousers is the largest, so the selected trousers material should meet this tensile strength, if the fabric of the trousers and the bandage is too soft or not wrapped in place, the system cannot provide enough friction force and tensile force, and when the force applied by the exoskeleton to the knee joint suddenly increases, it will cause the relative transverse displacement (slip) of the trousers and the bandage with the leg, even tear the trousers, to a certain extent, affect the assisting effect of the exoskeleton, and also cause uneven force on the leg, affect the wearing experience of the exoskeleton system.

[0039] Therefore, the application also provides a lower limb exoskeleton device based on a man-machine dynamics model, which comprises a group of symmetrically arranged leg exoskeleton units. The leg exoskeleton unit is designed to combine with the leg of a wearer to provide motion support and assistance. Each leg exoskeleton unit mainly comprises: an exoskeleton body 100, a thigh support part 110 and a calf support part 120 detachably mounted on the exoskeleton body 100. Referring toFigures 1-2 .

[0040] The thigh support part 110 includes a main bandage 111 and a secondary bandage 112. The main bandage 111 is detachably connected to the exoskeleton body 100 in a rigid structure and cooperates with the trousers 200 of the wearer to be embedded in the thigh part of the trousers 200.

[0041] The rigidity of the main bandage 111 ensures that it can effectively transmit the power of the exoskeleton to the thigh of the wearer, and its embedding in the trousers 200 reduces the friction and pressure of direct contact with the human skin, improving the wearing comfort and aesthetics. The secondary bandage 112 is arranged in a flexible structure outside the exoskeleton body 100 and between the calf support part 120 and the main bandage 111, so as to be fitted with the knee area of the trousers 200. The flexibility of the secondary bandage 112 makes it better adapt to the complex movement of the knee joint and provide dynamic constraint and auxiliary force transmission when the wearer moves. It avoids the problem that the traditional rigid bandage may limit the natural activity of the knee joint, and improves the freedom and comfort of the joint.

[0042] The calf support part 120 includes a fixed connecting rod 121 arranged below the exoskeleton body 100, and the fixed connecting rod 121 is connected with two calf banding devices 122 at both ends. The calf banding device 122 is embedded in the calf part of the trousers 200 of the wearer. Similar to the thigh support part, the embedding of the calf banding device 122 in the trousers 200 is also to provide stable fixation, efficient force transmission and better wearing experience. The fixed connecting rod 121 ensures the structural integrity and strength of the calf support part.

[0043] The trousers 200 are provided with a flexible buffer belt 210 corresponding to the areas of the thigh support part 110 and the calf support part 120.

[0044] Further, the trousers 200 adopt a multi-layer design, wherein the first layer of trousers is a skin-friendly layer directly contacting the skin of the wearer, and the second layer of trousers is a functional structure layer combined outside the first layer of trousers and internally shaped with an embedded space for accommodating and positioning the main bandage 111, the secondary bandage 112 and the calf banding device 122 of the thigh support part 110, that is, to improve the supporting capacity and meet the requirements of embedding the thigh main bandage and the calf support part.

[0045] Referring to Figure 3The second layer of the pants is provided with a flexible buffer belt, which is located inside the pants 200 and is used to form a flexible interface between the exoskeleton body 100 and the wearer's leg. Specifically, the flexible buffer belt 210 serves as a soft and deformable medium layer, effectively avoiding direct contact between the exoskeleton rigid components and the human skin, significantly improving the wearing comfort, and can uniformly disperse the pressure of the exoskeleton on the leg, while also preventing skin abrasion and pressure sores.

[0046] Referring to Figure 4 In a preferred embodiment, the pants structure is optimized according to a dynamic model, and a reinforcing structure 220 is provided on the pants 200 through force analysis. The reinforcing structure 220 includes a waistband 221 and a plurality of reinforcing ribs 222 connecting the waistband 221 and the thigh support portion, for transmitting the weight of the exoskeleton device to the wearer's waist. The weight of conventional exoskeleton devices is often concentrated on the legs, increasing the burden on the lower limbs. By providing the waistband 221 and the reinforcing ribs 222, part or all of the weight of the exoskeleton can be effectively transferred to the waist of the human body, which is more suitable for bearing weight, through the connection of the pants and the thigh support portion, thereby reducing the weight-bearing burden on the wearer's legs and improving the comfort of long-term wearing.

[0047] In order to improve the wearing comfort of the exoskeleton and optimize the force on the pants, in another preferred embodiment, the main body of the flexible buffer belt 210 is a carbon fiber sheet, which is wrapped with silicone on the outside and provided with a magnet on the inside, for realizing magnetic attraction connection with the exoskeleton body 100, and the flexible buffer belt 210 is composed of a plurality of movable modules.

[0048] In this embodiment, the carbon fiber sheet is used as the main body of the flexible buffer belt 210, which not only ensures sufficient support strength, but also maintains lightweight. The silicone wrapped on the outside provides a soft and skin-friendly touch and good friction, further improving comfort and stability. The magnet provided inside can realize magnetic attraction connection with the corresponding magnetic components on the exoskeleton body, simplifying the installation and disassembly process, and also enabling the flexible buffer belt 210 to be more closely and quickly attached to the exoskeleton body 100. The flexible buffer belt 210 is composed of a plurality of movable modules, which can better adapt to the geometric shape of different wearer's legs and dynamic deformation during movement, providing better fit and force transmission effect, and facilitating individual replacement and maintenance for different parts of damage or wear.

[0049] It should be noted that, in the present text, the terms "comprises", "comprising", or any other variant thereof, are intended to cover a non-exclusive inclusion, such that a process, device, article or exoskeleton control method based on human-robot dynamics model comprising a list of elements does not only include those elements, but is also capable of preventing additional elements not expressly listed or inherent to such process, device, article or exoskeleton control method based on human-robot dynamics model. Without further limitations, an element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, device, article or exoskeleton control method based on human-robot dynamics model comprising the element.

[0050] The above description is merely preferred embodiments of the present application, and is not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation using the content of the present application specification and drawings, or direct or indirect application in other related technical fields, is also included in the patent protection scope of the present application.

Claims

1. An exoskeleton control method based on a human-machine dynamics model, characterized in that, Includes the following steps: S1. Select the exoskeleton program mode, which includes follow mode and assist mode; S2. When the exoskeleton is in follow mode, it automatically acquires joint motion data and corresponding joint active torques when the wearer performs a standard target movement, and constructs a model database including standard kinematic and dynamic parameters. When the exoskeleton is in assist mode, it automatically collects joint motion data and joint active torques in the wearer's current movement state, and matches the pre-stored model parameters corresponding to the current movement state based on the model database, thereby initializing and establishing a human leg-exoskeleton dynamic model, and using an online parameter identification algorithm to estimate and update the model parameters to obtain the estimated values ​​of the dynamic model parameters. S3. Based on the joint active torque, joint motion data and wearer's leg information obtained in the above steps, calculate the center of mass position, velocity and acceleration of each segment of the lower limb through forward kinematics, and obtain the estimated values ​​of kinematic model parameters by combining the preset human segment inertial parameter model. S4. Based on the estimated values ​​of the dynamic model parameters and the estimated values ​​of the kinematic model parameters, establish the state-space equation of the exoskeleton system; S5. Based on a deep learning neural network model, determine the wearer's current motion state according to the real-time collected motion data, and combine the state space equation of the exoskeleton system to generate the expected control law of the exoskeleton corresponding to the current motion state. S6. Input the desired control law of the exoskeleton into the drive device of the exoskeleton to assist the wearer in movement, and repeat steps S2 to S6 to achieve adaptive assist control during continuous movement.

2. The exoskeleton control method based on a human-machine dynamics model according to claim 1, characterized in that, The human leg-trouser body-exoskeleton dynamic model includes: ； in, This refers to the range of motion of the knee joint. The angular velocity of the knee joint. This refers to the angular acceleration of the knee joint. The inertia matrix, The matrix of centrifugal force and Coriolis force (Coriolis force). For the gravity matrix, This is due to friction and the interference of the trousers with the exoskeleton. This is the auxiliary torque for the motor. This is the main driving torque of the human knee joint.

3. The exoskeleton control method based on a human-machine dynamics model according to claim 2, characterized in that, In step S2, the construction of the database sample specifically involves: Joint motion data is acquired by encoders and inertial measurement units installed at the joints of the exoskeleton, and electromyographic signals are acquired by surface electromyography sensors installed on the corresponding muscle groups of the wearer's lower limbs. The active torque of the joint is estimated after processing. The joint motion data collected during multiple executions of the target motion and the estimated joint active torque are time-aligned and normalized, and then stored as database samples.

4. The exoskeleton control method based on a human-machine dynamics model according to claim 3, characterized in that, In step S2, the online parameter identification algorithm employs either recursive least squares or extended Kalman filtering.

5. The exoskeleton control method based on a human-machine dynamics model according to claim 1, characterized in that, In step S3, the forward kinematics calculation is based on the standard Denavit-Hartenberg parametric method to establish a lower limb kinematic model, which is used to calculate the pose of the thigh, calf and foot segments of the lower limb in the inertial coordinate system according to the joint motion data, and to calculate the position, linear velocity and linear acceleration of the center of mass of each segment using rigid body kinematics formulas.

6. The exoskeleton control method based on a human-machine dynamics model according to claim 1, characterized in that, In step S4, the deep learning neural network model is a classification or regression model pre-trained with data from multiple users and multiple motion modes, used to identify at least one motion state among walking, going up and down stairs, and squatting.

7. A lower limb exoskeleton device according to any one of claims 1-6, comprising a set of symmetrically arranged leg exoskeleton units and matching trousers, wherein the leg exoskeleton unit comprises: An exoskeleton body, on which a thigh support and a lower leg support are detachably mounted, characterized in that the thigh support includes a main strap and a secondary strap, the main strap being rigidly connected to the lower part of the exoskeleton body and fitting into the wearer's pants, thus embedding itself into the thigh portion of the pants, the secondary strap being flexible and located on the outside of the exoskeleton body between the lower leg support and the main strap, thus fitting against the knee area of ​​the pants, for providing dynamic constraint and auxiliary force transmission during wearer movement; The lower leg support includes a fixed connecting rod, which is located below the main body of the exoskeleton. Both ends of the fixed connecting rod are connected to two lower leg binding devices, which are embedded in the lower leg portion of the wearer's trousers. The pants have flexible cushioning bands in the areas corresponding to the thigh support and calf support. The flexible cushioning bands are located on the inside of the pants and are used to form a flexible interface between the exoskeleton body and the wearer's legs.

8. A lower limb exoskeleton device according to claim 7, characterized in that, The pants also have a reinforcing structure, which includes a waistband and multiple reinforcing ribs connecting the waistband and the thigh support, for transferring the weight of the exoskeleton device to the wearer's waist.

9. A lower limb exoskeleton device according to claim 7, characterized in that, The flexible buffer strip is made of carbon fiber sheet, which is wrapped with silicone and has magnets inside for magnetic connection with the exoskeleton body. The flexible buffer strip is composed of multiple movable modules.

Images