A hip joint exoskeleton structure capable of assisting carrying and walking at the same time and a control method thereof

Abstract

The present invention discloses a hip joint exoskeleton structure that can assist in carrying and walking at the same time, including a back fixing part, a waist fixing part and a thigh fixing part; a joint drive module is arranged on each side of the waist fixing part, and the joint drive module is connected to the corresponding thigh fixing part through a flexion and extension leg rod; an auxiliary bar block is arranged on the back fixing part, and the joint drive module is connected to the auxiliary bar block through a traction cable; a control and clutch module for adjusting the force of the traction cable is arranged on the waist fixing part; the present invention discloses a hip joint exoskeleton control method that can assist in carrying and walking at the same time. The structure of the present invention integrates carrying and walking assistance functions, which can simultaneously solve the needs of walking assistance and carrying bending assistance. In addition, the method of the present invention can continuously and accurately identify the carrying action and walking action and their subdivided motion states, and provide decision-making for power assistance under various complex working conditions.

Description

Technical Field

[0001] The present invention relates to the field of robotics technology, and in particular to a hip joint exoskeleton structure capable of assisting in carrying and walking at the same time and a control method thereof. Background Art

[0002] Exoskeletons can assist the human body in walking, carrying and other movements. External force assistance can improve work efficiency and prevent muscle fatigue and injury. They play an important role in logistics, industrial manufacturing, elderly care, construction and other fields and have strong practical significance.

[0003] The problems with existing exoskeletons are as follows:

[0004] (1) Most existing exoskeleton devices are hip, knee and ankle multi-joint robot structures. The entire system is complex and bulky, making the assisted motion operation not flexible and light enough, reducing the user's assistance experience and actual application effect.

[0005] (2) Existing products can only realize one function, that is, carrying assistance or walking assistance. They are single-function products and cannot realize the assistance of carrying assistance and walking assistance and can switch between them continuously, which cannot meet the complex working conditions of actual scenes. At the same time, they cannot realize the intention recognition of subdivided states in the process of carrying assistance and walking well. There are problems with the accuracy and switching speed of pattern recognition, and they cannot realize the assistance of each subdivided state and the switching process between them in the two modes.

[0006] (3) The existing methods are unable to achieve parametric planning and optimization of the power assist curves in each process of power assist and handling. During the power assist and handling process, more precise power assist control is required for each subdivided state to improve the comfort and smoothness of the assistance.

[0007] The patent specification with announcement number CN110802603B discloses a method for controlling an active waist-assisted exoskeleton, including a sensor module, a brushless DC motor, a signal acquisition and processing module, a transport mode switching motion control module, a motor drive module, and a human-machine interaction interface. The sensors include: an electromyographic sensor and a motor encoder. The data collected by the electromyographic sensor and the motor encoder are processed and used as the basis for switching the transport state, thereby realizing the fully automatic switching function of the waist exoskeleton robot's transport state and achieving the effect of human-machine integration.

[0008] The patent specification with announcement number CN107088872B discloses a mechanism with dual functions of carrying and assisting walking. When assisting walking, a rocker slider mechanism composed of a support rod and a sliding component is used to drive the spring to stretch, and energy is stored and released through the flexion and extension of the spring to provide assistance; when carrying, the spring fixed on the planar connecting rod mechanism is used to offset part of the torque during the loading process, and the stepper motor drives the gear transmission to control the angle of the connecting rod mechanism.

[0009] Although the above two solutions have their own advantages, they are still unable to achieve parametric planning and optimization of the power assist curves in each process of power assist and transportation conditions. During the power assist and transportation process, more precise power assist control is required for each subdivided state to improve the comfort and smoothness of the assistance. Summary of the invention

[0010] The purpose of the present invention is to provide a hip joint exoskeleton structure that can assist in carrying and walking at the same time, integrating the carrying and walking assistance functions, and only by driving the flexion and extension of the hip joint and the arc bending freedom of the waist through a motor, it can simultaneously meet the needs of walking assistance and carrying and bending assistance.

[0011] A hip joint exoskeleton structure capable of assisting in carrying and walking simultaneously, comprising a back fixing part, a waist fixing part and a thigh fixing part;

[0012] A joint drive module is arranged on each side of the waist fixing part, and the joint drive module is dynamically connected to the thigh fixing part on the corresponding side through a leg flexion and extension rod; an auxiliary strip block is arranged on the back fixing part, and the joint drive module is dynamically connected to the auxiliary strip block through a traction wire; a control and clutch module for adjusting the force of the traction wire is arranged on the waist fixing part.

[0013] When assisting walking, the joint drive module outputs an active torque that acts on the thigh through the thigh fixing part, thereby achieving the purpose of assisting walking; when carrying, the torque output by the joint drive module is transmitted to the back fixing part and the waist fixing part through the traction wire, thereby achieving the purpose of saving effort in carrying.

[0014] Preferably, the thigh fixing member comprises a thigh connecting plate and a thigh strap fixed to the thigh connecting plate; one end of the flexion and extension leg rod is connected to the thigh connecting plate, and the other end is connected to the joint driving module via a damping hinge.

[0015] In this solution, the joint drive module realizes the active degree of freedom of flexion and extension, and the damping hinge realizes the passive degree of freedom of adduction and abduction.

[0016] Preferably, the auxiliary strip block is composed of a number of trapezoidal blocks arranged in a line, extending from the waist to the back, with the spacing between adjacent trapezoidal blocks increasing continuously and the width of the top surface of the trapezoidal blocks gradually decreasing; the traction wire passes through each trapezoidal block in turn and is fixed; further preferably, the traction wire adopts a Bowden wire.

[0017] In this solution, when the human body bends over to carry, the Bowden cable is not effective and the gap between the trapezoidal blocks continues to increase; when the person is carried and standing up, the pulling force is transmitted to each trapezoidal block through the Bowden cable, pulling the back fixings to assist the human body's waist and back to stand upright. The width of the top surface of the trapezoidal blocks is different, which can adapt to the different ranges of motion of the human waist and back. The range of motion of the waist is smaller when it is lifted, and the trapezoidal blocks will touch first when pulled. The trapezoidal blocks that have not been touched will transmit a larger pulling force to lift the human body, and the assistance can be better transmitted to the required position of the waist and back. In extreme cases, such as when the waist is suddenly pulled by external force, it is easy to cause sprains. The trapezoidal blocks also play a role in limiting protection to prevent sprains of the waist.

[0018] A hip joint exoskeleton control method capable of assisting carrying and walking simultaneously comprises the following steps:

[0019] (1) Obtain the angle values ​​of the human body trunk, left leg and right leg in the sagittal plane through inertial sensors located at the waist, the front of the left thigh and the front of the right thigh respectively;

[0020] (2) The angle data is sent to the microprocessor after passing through the signal acquisition and processing module;

[0021] (3) The microprocessor uses the motion state recognition model obtained through pre-offline training to recognize the human body's motion intention;

[0022] (4) Generate an assist torque curve based on the human body's movement intention and current state, control the output torque of the joint module, and achieve assistive force during human movement.

[0023] This method can effectively identify human movement intentions, realize multi-mode power assistance functions for carrying and walking, improve workers' work efficiency, prevent body strain, and also play a role in strengthening the body; it uses fewer sensors to effectively realize multi-state intention recognition and judgment of human movement, and improves the convenience of wearing and maintenance.

[0024] Preferably, the step (3) is specifically:

[0025] The flexion and extension angles of the left and right hip joints and the trunk inclination angle in the sagittal plane collected by inertial sensors are used as the training set input, and the actual movement state is used as the output. The support vector machine model is used for classification model training, and the trained classification model is used to recognize human movement intention.

[0026] Preferably, the method further comprises optimizing the boost torque curve in step (4);

[0027] The optimization of the assist torque curve includes parameterization of the assist torque curve, offline optimization of the assist torque curve based on physiological signal feedback, and packaging of the optimized assist torque curve.

[0028] Preferably, the boost torque curve parameterization is specifically as follows:

[0029] A power assistance curve is set in which the power assistance value varies with the gait phase, and an expected value is obtained from the power assistance curve according to the gait phase.

[0030] Preferably, the offline optimization of the assist torque curve based on physiological signal feedback comprises the following steps:

[0031] ① Obtain the electrical signal of the corresponding muscle through the electromyographic signal sensor;

[0032] ② The electrical signal is transmitted to the host computer via a microprocessor to perform myoelectric signal preprocessing and activation degree characterization calculation;

[0033] ③ Using the algorithm to optimize the torque curve parameters, a new set of torque curve parameters is generated;

[0034] ④ Send a new set of torque curve parameters to the microprocessor, and the microprocessor controls the exoskeleton joint module to apply the torque to the human body; after repeated iterations, a set of torque curve parameters that can ensure the lowest muscle activation is optimized.

[0035] Beneficial effects of the present invention:

[0036] (1) The hip joint exoskeleton structure of the present invention is simple, lightweight and flexible, integrating carrying and walking assistance functions. The flexion and extension freedom of the hip joint is driven only by a motor, which can simultaneously meet the needs of walking assistance and carrying and bending assistance.

[0037] (2) The motion pattern recognition algorithm can continuously and accurately identify carrying movements, walking movements and their subdivided motion states, providing decision-making for power assistance under various complex working conditions.

[0038] (3) The power assist curves of the alternating multi-state processes of walking and carrying are controlled parametrically so that the power assist meets the actual requirements of each motion condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a schematic diagram of the skeleton structure of the present invention;

[0040] Figure 2 It is a connection diagram of the joint drive module, the flexion-extension connecting plate, the flexion-extension leg rod connecting rod and the flexion-extension leg rod;

[0041] Figure 3 The diagram is a schematic diagram of the structure of the auxiliary strip block. The middle one is when it is bent, and the right one is when it is lifted under tension.

[0042] Figure 4 It is a structural diagram of the control and clutch module;

[0043] Figure 5 It is a structural block diagram of the control system of the present invention;

[0044] Figure 6 The structural block diagram for offline training and online application of human motion intention recognition model;

[0045] Figure 7 Schematic diagram for transport assistance;

[0046] Figure 8 Schematic diagram for assisting walking;

[0047] Fig. 9 Schematic diagram of the parameterization of the torque curve;

[0048] Fig.10 Iterative optimization algorithm flow for boosting torque based on physiological signal feedback. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] like Figure 1 and 2 As shown, a hip joint exoskeleton structure that can assist in carrying and walking at the same time includes a back fixing part 1, a waist fixing part 2 and a thigh fixing part 3; in this embodiment, the back fixing part 1 is a top, the waist fixing part 2 is a belt, and the top is connected to the belt, and the thigh fixing part 3 includes a thigh connecting plate 31 and a thigh strap 32 fixed on the thigh connecting plate 31.

[0051] A joint driving module 4 is arranged on each side of the waist fixing part 2. The joint driving module 4 adopts a combination of a motor 41 and a reducer. The joint driving module 4 is fixed on the waist fixing part 2 through a flexion and extension connecting plate 5. A hip joint cover plate 6 is fixed on the inner side of the joint driving module 4 to facilitate contact with the human body; a flexion and extension leg rod connecting rod 7 is connected to the output shaft of the joint driving module 4, and the end of the flexion and extension leg rod connecting rod 7 is connected to one end of the flexion and extension leg rod 9 through a damping hinge 8, and the other end of the flexion and extension leg rod 9 is fixed to the thigh connecting plate 31.

[0052] The back fixing member 1 is provided with an auxiliary bar block 10, and the joint driving module 4 is connected to the auxiliary bar block 10 by a traction cable 11; specifically, the auxiliary bar block 10 includes two parallel arranged ones; Figure 3As shown, the auxiliary bar block 10 is composed of a number of trapezoidal blocks arranged in a line, extending from the waist to the back, the distance between adjacent trapezoidal blocks continues to increase, and the width of the top surface of the trapezoidal blocks gradually decreases; the traction wire 11 specifically adopts a Bowden wire, one end of the Bowden wire passes through each trapezoidal block in turn and is fixed, and the other end is connected to the output shaft of the joint drive module 4.

[0053] The waist fixing member 2 is provided with a control and clutch module for adjusting the force of the traction wire 11; Figure 4 As shown, the clutch part adopts an electromagnetic clutch 12, and the electromagnetic clutch 12 has a first wire-laying wheel 121 on one side and a second wire-laying wheel 121 on the other side. Bowden wires are respectively wound around the first wire-laying wheel 121 and the second wire-laying wheel 122. The lead-out end of the Bowden wire on the first wire-laying wheel 121 is connected to the output shaft of the motor 41, and the lead-out end of the Bowden wire on the second wire-laying wheel 122 is connected to the auxiliary strip block 10; the control part includes a controller 13, and the controller 13 is used to control the on and off of the electromagnetic clutch 12.

[0054] The working principle of the clutch is specifically as follows: when the motor 41 is driven, it pulls the Bowden cable and outputs torque to the second cable wheel 122 of the electromagnetic clutch 12 to generate torque. If the current opportunity requires power to be transmitted to the waist and back, the electromagnetic clutch 12 is energized to transmit the torque to the Bowden cable of the waist and back through the first cable wheel 121. Otherwise, the power is cut off and the torque will not be uploaded to the waist and back.

[0055] like Figure 5 and 6 As shown, a hip joint exoskeleton control method that can simultaneously assist in carrying and walking includes the following steps:

[0056] (1) The angle values ​​of the human body trunk, left leg and right leg in the sagittal plane are obtained through inertial sensors (IMU) located at the waist, the front of the left thigh and the front of the right thigh respectively;

[0057] (2) The angle data is sent to the microprocessor after passing through the signal acquisition and processing module, and serves as the basis for logical judgment;

[0058] (3) The microprocessor uses the motion state recognition model obtained through pre-offline training to recognize the human body's motion intention;

[0059] (4) Generate an assist torque curve based on the human body's movement intention and current state, control the output torque of the joint module, and achieve assistive force during human movement.

[0060] like Figure 7 and 8As shown, in the carrying assistance state, the microprocessor controls the left and right hip joints to output extension torques in the same direction to achieve carrying assistance for the waist, while in the walking assistance state, the microprocessor controls the left and right hip joints to alternately achieve flexion / extension torques in opposite directions to achieve walking assistance for the thigh.

[0061] Among them, the specific methods for recognizing the intention of assisting walking and carrying human body movements are as follows:

[0062] The present invention uses a data-driven machine learning classification algorithm to identify the carrying and walking modes and their corresponding subdivided states (each subdivided state includes bending, squatting, lifting, standing, stepping with the left leg, stepping with the right leg, etc.). The flexion and extension angles of the left and right hip joints and the inclination angle of the trunk in the sagittal plane collected by the inertial sensor IMU are used as the training set input, and the actual movement state is used as the output. The classic support vector machine model is used for classification model training, and the classification model obtained by training is used to identify the human body's movement intention, and finally the judgment and identification of the alternating switching of each subdivided state are realized. The specific process is as follows: Figure 6 shown.

[0063] The power-assistance torque curve determines the power-assistance effect of the hip joint exoskeleton. There are obvious differences in the power-assistance torque curve for different working conditions such as carrying and walking. The power-assistance torque curve needs to be optimized offline first, so that the optimal mode can be selected in combination with the motion state during actual use. The power-assistance torque curve optimization mainly includes the parameterization of the power-assistance torque curve, the offline optimization of the power-assistance torque curve based on physiological signal feedback, and the packaging of the optimized power-assistance torque curve.

[0064] Assist torque curve parameterization:

[0065] The parameterization of the assist torque curve can be used to achieve the smoothest assist torque planning and the most effective assist torque by adjusting parameters for different walking or carrying task stages, combined with the motion state and assist task requirements of each stage. The specific assist torque parameterization is as follows: Fig. 9 shown.

[0066] By presetting the power assist curve to provide power assist torque values ​​in certain gait phases (i.e., the percentage of each time phase in a walking cycle or a carrying cycle in the entire cycle time) and smoothly connecting the two curves, the formula for calculating the power assist curve a can be listed as follows:

[0067]

[0068] The above τ E_p They represent the gait phase or carrying phase corresponding to the peak value of the stretch assist torque curve, the rising gait phase interval of the stretch assist curve, and the stretch assist peak value; f a, a1, a2, a3 represent the stretching assistance curve a and the three parameters in its polynomial expression. The polynomial parameters of the assistance curve a are finally obtained as follows:

[0069]

[0070] Similarly, the parameter value of the assist curve b can be obtained:

[0071]

[0072] Parameter value of the power curve c:

[0073]

[0074] Parameter value of the power assist curve d:

[0075]

[0076] Offline optimization of the assist torque curve based on physiological signal feedback:

[0077] With the help of the above-mentioned parameterization of the power assist torque curve, the adjustment of the power assist distance curve is realized through a set of parameters, and the optimization of this set of parameters is obtained by iterative optimization through physiological signal feedback with the help of Bayesian optimization algorithm. The process is as follows: Fig.10 As shown. Taking the walking-assisting condition as an example, by sticking the surface electromyography signal sensors on the rectus femoris, vastus medialis and vastus lateralis muscles, the weak electrical signals of the muscles can be obtained to characterize the strength of the human muscles. The signals are transmitted to the host PC via Bluetooth through the microprocessor for electromyography signal preprocessing and activation characterization calculation. The power-assisting torque curve parameters are then optimized with the help of the Bayesian optimization algorithm. With the goal of reducing muscle activation, a new set of power-assisting torque curve parameters is generated and sent to the microprocessor via Bluetooth. The microprocessor controls the exoskeleton joint module to apply the power-assisting torque to the human body. After repeated iterations, a set of power-assisting torque curve parameters that can ensure the lowest muscle activation is optimized, thereby achieving an optimized power-assisting torque curve in each working condition.

[0078] In order to ensure that the assist torque is optimal under different movement speeds, the assist distance curves under the carrying and walking conditions are optimized offline and divided into multiple gears according to the movement speed, so as to obtain the optimized curves in different speed ranges.

[0079] Optimized torque curve package:

[0080] The optimized power-assist torque curve is associated with the corresponding working condition and motion speed curve and packaged into a microprocessor. In actual application, the motion intention and power-assist working condition are identified through the motion intention recognition algorithm, and the motion speed is calculated. The microprocessor is used to look up the table to obtain the planned and packaged optimal power-assist torque curve as the control input of the exoskeleton.

[0081] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A hip joint exoskeleton control method that can assist in carrying and walking at the same time, characterized in that: Including hip exoskeleton structure and control method; The hip joint exoskeleton structure includes a back fixing part, a waist fixing part and a thigh fixing part; A joint drive module is provided on each side of the waist fixing part, and the joint drive module is dynamically connected to the corresponding thigh fixing part through the leg flexion and extension rod; an auxiliary bar block is provided on the back fixing part, and the joint drive module is dynamically connected to the auxiliary bar block through a traction cable; a control and clutch module for adjusting the force of the traction cable is provided on the waist fixing part; The control method comprises the following steps: (1) Obtain the angle values ​​of the human body trunk, left leg and right leg in the sagittal plane through inertial sensors located at the waist, the front of the left thigh and the front of the right thigh respectively; (2) The angle data is sent to the microprocessor after passing through the signal acquisition and processing module; (3) The microprocessor uses the motion state recognition model obtained through pre-offline training to recognize the human body's motion intention. The specific process is as follows: The flexion and extension angles of the left and right hip joints and the trunk inclination angle in the sagittal plane collected by inertial sensors are used as the input of the training set, and the actual movement state is used as the output. The support vector machine model is used for classification model training, and the trained classification model is used to identify the human body's movement intention. (4) Generate an assist torque curve based on the human body's movement intention and current state, control the output torque of the joint module, and achieve assistive force during human movement; It also includes optimizing the power-assisting torque curve, wherein the optimization includes parameterization of the power-assisting torque curve, offline optimization of the power-assisting torque curve based on physiological signal feedback, and packaging of the optimized power-assisting torque curve; The parameterization of the boost torque curve is specifically: A power assistance curve is set in which the power assistance value varies with the gait phase, and an expected value is obtained from the power assistance curve according to the gait phase. The power assistance curve includes: aThe formula of the power curve is: in, τ E_p They represent the gait phase or carrying phase corresponding to the peak value of the stretch assist torque curve, the rising gait phase interval of the stretch assist curve, and the stretch assist peak value; f a , a1, a2, a3 represent the stretching assistance curve a and the three parameters in its polynomial expression respectively; The polynomial parameters of the a-boosting curve are: Similarly, the parameter value of the assist curve b can be obtained: Parameter value of the power curve c: Parameter value of the power assist curve d: The offline optimization of the assist torque curve based on physiological signal feedback comprises the following steps: ① Obtain the electrical signal of the corresponding muscle through the electromyographic signal sensor; ② The electrical signal is transmitted to the host computer via a microprocessor to perform myoelectric signal preprocessing and activation degree characterization calculation; ③ Optimize the torque curve parameters with the help of algorithms to generate a new set of torque curve parameters; ④ Send a new set of torque curve parameters to the microprocessor, and the microprocessor controls the exoskeleton joint module to apply the torque to the human body; after repeated iterations, a set of torque curve parameters that can ensure the lowest muscle activation is optimized.

2. The hip joint exoskeleton control method capable of assisting carrying and walking at the same time according to claim 1, characterized in that: The thigh fixing member comprises a thigh connecting plate and a thigh strap fixed on the thigh connecting plate; one end of the flexion and extension leg rod is connected to the thigh connecting plate, and the other end is connected to the joint driving module through a damping hinge.

3. The hip joint exoskeleton control method capable of assisting carrying and walking at the same time according to claim 1, characterized in that: The auxiliary strip block is composed of a number of trapezoidal blocks arranged in a line, extending from the waist to the back, the distance between adjacent trapezoidal blocks increases continuously, and the width of the top surface of the trapezoidal blocks gradually decreases; the traction wire passes through each trapezoidal block in turn and is fixed.

Images