Stable walking development and control method for self-balancing lower limb exoskeleton

Abstract

The invention provides a stable walking development and control method for a self-balancing lower limb exoskeleton, and the method comprises the steps: generating an initial mass center track meeting the stable walking requirement of the exoskeleton based on an exoskeleton dynamics simplified model of a table-trolley model; a pitching momentum controller based on centroid linear acceleration adjusts the centroid trajectory to obtain a real-time centroid trajectory; generating a swing leg motion track based on a third-order Bezier curve, changing step length and step height of an exoskeleton foothold in real time by adjusting coordinates of a control point, and generating an initial foothold track; a yaw momentum controller based on step length adjustment adjusts the initial foothold trajectory to obtain a real-time foothold trajectory; and performing stable walking control on the self-balancing lower limb exoskeleton according to the real-time centroid track and the real-time foothold track. According to the method, the stability of the self-balancing lower limb exoskeleton gait is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robot technology, and in particular to a stable walking development and control method for a self-balancing lower extremity exoskeleton. BACKGROUND

[0002] Lower extremity exoskeleton robots are mainly divided into two categories: non-self-balancing lower extremity exoskeletons and self-balancing lower extremity exoskeletons. Non-self-balancing lower extremity exoskeletons cannot achieve self-balancing walking assistance by themselves due to the lack of sufficient active degrees of freedom, and the wearer must rely on external devices such as wheelchairs or crutches to maintain balance, thereby ensuring the stability of the human-exoskeleton system and avoiding falling. Therefore, non-self-balancing lower extremity exoskeletons are not suitable for patients with quadriplegia and can only be applied to patients with hemiplegia or lower extremity paraplegia with intact upper extremities. Self-balancing lower extremity exoskeletons have the ability to maintain the balance of the human-exoskeleton system during walking, freeing the wearer from the need to rely on external support devices to maintain balance, and can be widely used in the daily rehabilitation training and walking assistance of patients with hemiplegia, paraplegia, and quadriplegia, expanding the application field. Due to the complexity of the system and the high difficulty of control, there are fewer teams researching self-balancing lower extremity exoskeletons at present, and most of them are in a relatively preliminary stage. Different research teams have developed several typical self-balancing lower extremity exoskeleton prototypes / products to promote patient gait rehabilitation, such as ALEX-I, AutoLEE-G1, REX, and Atalante. The mechanism configuration of the above devices is a simple series mechanism, resulting in a large axis deviation between the hip joint and the ankle joint and the human body. Joint axis deviation will cause additional resistance force / moment to the human joints and torso during the robot's walking, which not only affects the wearer's comfort but also poses a certain safety hazard.

[0003] The core control goal of a self-balancing lower limb exoskeleton is to maintain the balance of the human-exoskeleton system and keep stable walking while adapting to different wearers. Generally, the methods of stable walking for a self-balancing lower limb exoskeleton (similar to a biped robot) can be divided into two categories. The first category is a model-based method, and the most commonly used model is a simplified point mass model, such as an inverted pendulum model, a spring-loaded inverted pendulum model, and a table-cart model, which describes and analyzes the robot through analytical differential equations. The advantage of the control method based on the simplified model is its robustness and scalability, which is not sensitive to the physical parameters of different wearers and can be quickly deployed to other self-balancing lower limb exoskeletons. Other commonly used models include: center of mass dynamics model, hybrid zero dynamics model, and full-body dynamics model, etc. The second category is a non-model-based method, including a periodic gait generation strategy based on an adaptive oscillator, a gait generation strategy based on motion prediction, and a gait generation strategy based on deep reinforcement learning, etc. The advantage of these methods is that they do not depend on the accuracy of the model, so they have strong robustness to model errors. However, the disadvantage is that the control strategy successfully applied to a certain robot hardware system cannot be directly transplanted to other platforms.

[0004] There are still some problems to be solved in the stable walking of existing self-balancing lower limb exoskeleton robots. Although the gait generator based on the simplified point mass model has strong robustness and scalability, the existing research ignores the rotational momentum of the actual multi-rigid body system during walking in the actual application process, so the travel trajectory often deviates from the planned gait trajectory. In the rotational inertia, the yaw moment needs special attention, especially when the robot is walking quickly or on a smooth surface, the yaw rotational inertia cannot be ignored. The rotational momentum in the sagittal plane can be controlled through the feedback of the ground reaction force and the body pitch angle. The yaw momentum is usually balanced by the rotational friction of the foot surface normal. For this, there are two common strategies in the field of humanoid robots: one is to offset the yaw moment through the rotation of the waist joint; the other is based on arm swing, that is, the arm rotates in the opposite direction. However, since the self-balancing lower limb exoskeleton usually has no structure above the waist and mainly faces patients with motor function disorders, the strategies based on arm swing and waist rotation are not applicable in such devices. SUMMARY

[0005] Therefore, the present application provides a stable walking development and control method for a self-balancing lower limb exoskeleton to solve the above problems.

[0006] The application provides a stable walking development and control method for a self-balancing lower extremity exoskeleton, which comprises the following steps: generating an initial center of mass trajectory meeting the stable walking requirement of the exoskeleton based on a table-trolley model exoskeleton dynamics simplified model; adjusting the center of mass trajectory based on a pitch momentum controller of linear acceleration of the center of mass to obtain a real-time center of mass trajectory; generating a swing leg motion trajectory based on a third-order Bezier curve, and changing the step length and step height of a landing point of the exoskeleton in real time by adjusting the coordinates of control points to generate an initial landing point trajectory; adjusting the initial landing point trajectory based on a yaw momentum controller of step length adjustment to obtain a real-time landing point trajectory; and controlling the stable walking of the self-balancing lower extremity exoskeleton according to the real-time center of mass trajectory and the real-time landing point trajectory.

[0007] In another implementation mode of the application, the step of generating an initial center of mass trajectory meeting the stable walking requirement of the exoskeleton based on a table-trolley model exoskeleton dynamics simplified model comprises the following steps: simplifying an AutoLEE-G3 exoskeleton dynamics model based on a table-trolley model to obtain an exoskeleton dynamics simplified model; and combining model predictive control and zero moment point theory to generate an initial center of mass trajectory meeting the stable walking requirement of the exoskeleton.

[0008] In another implementation mode of the application, the center of mass position in the initial center of mass trajectory and the planned landing point position meet the relationship of the ZMP equation as shown in the following formula:

[0009]

[0010] Wherein, H C is the height of the center of mass of the exoskeleton, g is the gravity constant, is the acceleration of the center of gravity.

[0011] In another implementation mode of the application, the pitch momentum dynamics formula when the sole of the self-balancing lower extremity exoskeleton rotates around the front edge or the rear edge is expressed as:

[0012]

[0013] Wherein, I p is the moment of inertia of the exoskeleton around the front edge or the rear edge of the sole, M is the total mass of the human-exoskeleton system, g is the gravity acceleration, θ p is the angle between the upper body and the vertical direction, l foot is the length of the exoskeleton sole, H is the height of the center of mass to the sole, d is the distance from the center of mass to the rotating edge, is the linear acceleration of the center of mass in the x-axis direction of the center of mass fixed on the foot.

[0014] In another implementation mode of the application, the linear acceleration of the center of mass is expressed as:

[0015]

[0016] where θ p is the pitch angle; is the angular velocity.

[0017] In another implementation of the present application, the horizontal and vertical axis decomposition of the initial foot trajectory is expressed as two decoupled components:

[0018]

[0019] where b x,i and b z,i are the weight coefficients of the polynomial on the x and z axes, respectively, (b x,i , b z,i ) is the i-th control point, (b x,0 , b z,0 ) and (b x,3 , b z,3 ) are the start and end points, respectively, and (b x,1 , b z,1 ) and (b x,3 , b z,3 ) are the control points of the trend of the exoskeleton foot trajectory.

[0020] In another implementation of the present application, the dynamics equation of the self-balancing lower limb exoskeleton generated by the rotational friction torque of the swing leg inertia and the upper body rotational inertia perpendicular to the supporting foot is expressed as:

[0021]

[0022] where τ static is the static torque when the supporting foot does not rotate, τ swing is the rotational torque generated by the swing leg inertia, τ rotational is the rotational torque generated by the base link rotational inertia of the exoskeleton, Δτ is the incremental rotational friction torque perpendicular to the surface of the supporting foot caused by the ground reaction force, is the effective mass of the swing leg, I y is the rotational mass inertia of the base link of the exoskeleton around the ZMP, is the angular acceleration of the center of mass, D is the torque arm length between the swing leg plane and the ZMP of the supporting leg, is the acceleration of the ankle joint of the swing leg in the horizontal direction, is the incremental acceleration of the ankle joint of the swing leg in the horizontal direction, M thigh , M shank , M footThese represent the masses of the exoskeleton's thigh, calf, and foot, respectively. These represent the accelerations at the center of mass of the exoskeleton's thigh and lower leg, respectively.

[0023] In another aspect, the present invention provides a stable walking development and control system for a self-balancing lower limb exoskeleton, comprising: a center-of-gravity trajectory generation module: generating an initial center-of-gravity trajectory that meets the requirements for stable walking of the exoskeleton based on a simplified exoskeleton dynamics model of a table-cart model; adjusting the center-of-gravity trajectory based on a pitch momentum controller with linear acceleration of the center-of-gravity to obtain a real-time center-of-gravity trajectory; a footpoint trajectory generation module: generating a swing leg motion trajectory based on a third-order Bézier curve, and changing the step length and step height of the exoskeleton's footpoint in real time by adjusting the coordinates of the control points to generate an initial footpoint trajectory; adjusting the initial footpoint trajectory based on a yaw momentum controller with step length adjustment to obtain a real-time footpoint trajectory; and a control module: performing stable walking control of the self-balancing lower limb exoskeleton based on the real-time center-of-gravity trajectory and the real-time footpoint trajectory.

[0024] In another aspect, the present invention provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a method for developing and controlling stable walking for a self-balancing lower limb exoskeleton as described in any of the preceding claims.

[0025] In another aspect, the present invention provides a computer storage medium, characterized in that the computer storage medium stores a computer program, which, when executed by a processor, implements the steps in a method for developing and controlling stable walking for a self-balancing lower limb exoskeleton as described in any of the preceding claims.

[0026] The present invention provides a method for the development and control of stable walking for a self-balancing lower limb exoskeleton. It adopts a multi-loop stable walking control framework based on a table-cart model of model predictive control, which has strong robustness to the physical parameters of the human-exoskeleton system. By designing pitch and yaw momentum controllers as a supplement to model predictive control, the stability of the human-exoskeleton system in the pitch and yaw directions is improved. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. By reading the detailed description of the embodiments below, the advantages and benefits of the solutions will become clear to those skilled in the art. The accompanying drawings are only for illustrating preferred embodiments and are not intended to limit the present invention. In the accompanying drawings:

[0028] Figure 1The flow chart of the development and control process of stable walking of the self-balancing lower extremity exoskeleton of an embodiment of the present application.

[0029] Figure 2 The structure framework of the multi-level stable walking control method based on the AutoLEE-G3 exoskeleton of an embodiment of the present application.

[0030] Figure 3 The schematic diagram of the exoskeleton dynamics simplified model based on the three-dimensional table-trolley model of an embodiment of the present application.

[0031] Figure 4 The schematic diagram of the exoskeleton angular momentum controller of an embodiment of the present application.

[0032] Figure 5 The schematic diagram of the walking of four subjects with different body feature parameters respectively wearing the AutoLEE-G3 exoskeleton on the flat ground of an embodiment of the present application.

[0033] Figure 6 The three-dimensional diagram of the expected motion trajectory of the AutoLEE-G3 exoskeleton in the Cartesian space of an embodiment of the present application.

[0034] Figure 7 The schematic diagram of the generation of the center of mass trajectory of the stable walking of the exoskeleton based on the model predictive control and the pitch momentum control of an embodiment of the present application.

[0035] Figure 8 The schematic diagram of the comparison of the yaw angle changes of the AutoLEE-G3 exoskeleton under the yaw angle momentum controller and without the yaw angle momentum controller under different load conditions on the flat ground of an embodiment of the present application.

[0036] Figure 9 The schematic diagram of the ZMP trajectory changes of the AutoLEE-G3 exoskeleton in the stable walking process under different conditions of an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to make personnel in the art better understand the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and in detail below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art should belong to the scope of protection of the present application.

[0038] Figure 1 The flow chart of the development and control method for the stable walking of the self-balancing lower extremity exoskeleton provided by the embodiments of the present application is as follows: Figure 1The embodiment shown mainly includes:

[0039] S101, based on the table-trolley model, the initial center of mass trajectory satisfying the stable walking requirement of the exoskeleton is generated.

[0040] S102, based on the pitch momentum controller of the linear acceleration of the center of mass, the center of mass trajectory is adjusted to obtain the real-time center of mass trajectory.

[0041] S103, the swing leg motion trajectory is generated based on the third-order Bezier curve, and the step length and step height of the exoskeleton foot point are changed in real time by adjusting the coordinates of the control points to generate the initial foot point trajectory.

[0042] S104, the yaw momentum controller based on the step length adjustment is used to adjust the initial foot point trajectory to obtain the real-time foot point trajectory.

[0043] S105, the self-balancing lower limb exoskeleton is controlled to walk stably according to the real-time center of mass trajectory and the real-time foot point trajectory.

[0044] The stable walking development and control method for the self-balancing lower limb exoskeleton of the application adopts a multi-loop stable walking control framework based on model prediction control of the table-trolley model, which has strong robustness to the physical parameters of the human-exoskeleton system; by designing the pitch and yaw momentum controllers as a supplement to the model prediction control, the stability of the human-exoskeleton system in the pitch and yaw directions is improved.

[0045] In another implementation manner of the application, the exoskeleton dynamics simplified model based on the table-trolley model generates an initial center of mass trajectory satisfying the stable walking requirement of the exoskeleton, which comprises: simplifying the AutoLEE-G3 exoskeleton dynamics model based on the table-trolley model to obtain an exoskeleton dynamics simplified model; combining model prediction control and zero moment point theory to generate an initial center of mass trajectory satisfying the stable walking requirement of the exoskeleton.

[0046] In another implementation manner of the application, the center of mass position in the initial center of mass trajectory and the planned foot point position satisfy the relationship of the ZMP equation as shown below:

[0047]

[0048] Wherein, H C is the height of the center of mass of the exoskeleton, g is the gravitational constant, is the acceleration of the center of gravity.

[0049] Exemplarily, as Figure 2As shown, the AutoLEE-G3 exoskeleton dynamics model needs to be simplified in the centroid trajectory planner. Common simplification methods are mainly based on the inverted pendulum model, the linear inverted pendulum model, and the table-cart model. The inverted pendulum and linear inverted pendulum models adjust the landing point position according to balance requirements during gait planning, which causes a deviation between the planned gait trajectory and the original planned landing point trajectory. Conversely, the table-cart model adjusts the centroid trajectory according to balance requirements during gait planning, theoretically allowing it to perfectly follow the original planned landing point trajectory. Therefore, this invention uses the table-cart model for centroid trajectory planning.

[0050] like Figure 3 As shown, when the AutoLEE-G3 exoskeleton stands on one leg at the k-th planned foothold... At that time, i.e., the reference position of the k-th zero moment point (ZMP). Its dynamic model can be simplified to a three-dimensional table-cart model, where all the weight of the exoskeleton is concentrated at the center of the cart. Assume the height of the exoskeleton's center of mass is H. C In order to achieve balance while walking, the position of the center of mass P com,k =(x k ,y k ) and the planned landing point location (i.e., ZMP location) P k =(p x,k ,p y,k The following ZMP equation relationship needs to be satisfied:

[0051]

[0052] In the formula, g is the gravitational constant.

[0053] To achieve online planning of the centroid trajectory, a centroid trajectory planning algorithm based on MPC theory was adopted. For equation (1), a new variable s = (s... x ,s y ), s x and s y The jerk (first derivative of acceleration) representing the center of gravity is used as input. Since the solution process for the gait trajectory is the same in the x and y directions, the following will only describe the specific solution process for the trajectory in the x direction.

[0054]

[0055] By taking s as the system input variable in equation (1), it can be transformed into a control problem. The state equation of the control system can be expressed in the following form:

[0056]

[0057] The continuous system equation (3) is discretized using a sampling time ΔT:

[0058]

[0059] where, s x,k ≡s x (kΔT),p x,k ≡p x (kΔT),

[0060] C≡[1 0-H C / g].

[0061] Given a series of reference ZMP trajectories and assume the corresponding system input as S x,k = [s x,k ,s x,k+1 ,…,s x,k+N ] T .

[0062] According to equation (4), a new state equation is obtained by N-step iteration:

[0063]

[0064] P x,k+1 is of length N, which also represents the prediction interval. In order to make the system output p x,k as accurate as possible to track the target Equation (5) is expressed as a QP problem, and the objective function is:

[0065]

[0066] Q and R represent the reference tracking error and the weight of minimizing system input, respectively. This QP problem can be solved by analytical method, so that:

[0067]

[0068] where, I N×N is a unit matrix. Substituting the above formula into the state equation (3) can obtain the optimal solution of s x,k as follows:

[0069] s x，k = e ΔT S x，k (7)

[0070] where,

[0071] The above equation gives the optimal input s x,k at the kth sampling time step, which is calculated based on the predicted state space model within the next N time steps.

[0072] In another implementation of the present application, the pitch momentum dynamics formula when the self-balancing lower extremity exoskeleton foot rotates around the front or rear edge is expressed as:

[0073]

[0074] where I p is the moment of inertia of the exoskeleton around the front or rear edge of the foot, M is the total mass of the human-exoskeleton system, g is the gravitational acceleration, θ p is the angle between the upper body and the vertical direction, l foot is the length of the exoskeleton foot, H is the height of the center of mass to the foot, d is the distance of the center of mass to the rotating edge, is the linear acceleration of the center of mass in the x-axis direction of the center of mass fixed on the foot.

[0075] Exemplarily, in the process of walking, the self-balancing exoskeleton is inevitably affected by various uncertain factors, such as the uncertainty of mathematical models based on different assumptions, the uncertainty of human-exoskeleton system parameters due to structural size, material properties, manufacturing and assembly errors, the random uncertainty of driving torque caused by actuator noise and joint friction, and the uncertainty of external environmental disturbance and ground conditions in the process of walking, which causes the model predictive control based on the table-cart model to fail to guarantee the rotational stability of the exoskeleton.

[0076] As shown in Figure 4 , an exoskeleton angular momentum controller is designed to supplement the model predictive control to guarantee the stability of the AutoLEE-G3 exoskeleton in the process of walking. In the process of realizing walking of the AutoLEE-G3 exoskeleton, the upper body thereof is always kept upright. Therefore, when the exoskeleton appears to be tilted forward or backward, the supporting foot will rotate around the front or rear edge of the foot, and the angle between the foot and the ground is equivalent to the pitch angle of the inertial measurement unit installed at the upper body of the exoskeleton, as shown in Figure 4 (A). The unified dynamics of this case can be expressed by the following formula:

[0077]

[0078] where I p is the moment of inertia of the exoskeleton around the front or rear edge of the foot, M is the total mass of the human-exoskeleton system, g is the gravitational acceleration, θ p is the angle between the upper body and the vertical direction, l footwhere L is the length of the foot of the exoskeleton, H is the height of the center of mass to the foot, and d is the distance of the center of mass to the rotating edge, is the linear acceleration of the center of mass in the x-axis direction.

[0079] In another implementation of the present application, the center of mass linear acceleration is expressed as:

[0080]

[0081] where θ p is the pitch angle; is the angular velocity.

[0082] Illustratively, to control the momentum of the AutoLEE-G3 exoskeleton in the sagittal plane, to avoid the foot of the exoskeleton rotating around the front edge or the back edge, the pitch angle caused by the forward or backward acceleration of the upper body is designed as a feedback control of the pitch momentum controller of the center of mass linear acceleration. The control law adjusts the center of mass trajectory of the exoskeleton in real time according to the pitch angle θ p and the angular velocity as feedback:

[0083]

[0084] The pitch momentum controller is a supplement to the model predictive control to generate a stable walking center of mass trajectory of the human-exoskeleton system. When the ZMP is out of the support polygon range, the model predictive control stops working. At this time, the pitch momentum controller is activated to pull the ZMP back into the support polygon range by adjusting the center of mass trajectory. When the support leg of the exoskeleton restores to full contact with the ground, the model predictive control starts working again. The pitch momentum controller will only work when the model predictive control fails, and the two will not update the center of mass trajectory at the same time.

[0085] In another implementation of the present application, the initial foot landing point trajectory is decomposed into two decoupled components in the horizontal and vertical axes, expressed as:

[0086]

[0087] where b x,i and b z,i represent the weight coefficients of the polynomial in the x-axis and z-axis, respectively, (b x,i , b z,i ) represents the i-th control point, (b x,0 , b z,0 ) and (b x,3 , b z,3 ) represent the start and end points, respectively, and (b x,1 , b z,1 ) and (b x,3 , b z,3) represents the control point of the trend of the trajectory of the stance point of the exoskeleton, the step length L = L f (1) -L f (0) = b x,3 -b x,0 , t represents the sampling time.

[0088] Exemplarily, a third-order Bezier curve is adopted to generate the swing leg motion trajectory, wherein two control points are used to adjust the step length and the step height of the stance point of the exoskeleton. The advantage of the Bezier curve lies in that it can conveniently modify the trajectory of the stance point of the exoskeleton online by adjusting the coordinates of the control points, and it is also easy to decompose the horizontal and vertical axes into two decoupled components.

[0089] Since the joints of the AutoLEE-G3 exoskeleton are based on position control, when planning the trajectory of the stance point of the exoskeleton, the swing leg should be made to land with as small a horizontal linear velocity as possible. If the exoskeleton is often disturbed by a relatively large horizontal impact force, the horizontal impedance will be relatively large, and it will be difficult to control stable walking.

[0090] In another implementation manner of the present application, the dynamics equation of the self-balancing lower limb exoskeleton under the rotational friction torque generated by the inertia of the swing leg and the inertia of the upper body perpendicular to the supporting foot is represented as:

[0091]

[0092] Wherein, τ static represents the static torque when the supporting foot does not rotate, τ swing represents the rotational torque generated by the inertia of the swing leg, τ rotational represents the rotational torque generated by the rotational inertia of the base link of the exoskeleton, and Δτ represents the incremental rotational friction torque perpendicular to the surface of the supporting foot caused by the ground reaction force, represents the effective mass of the swing leg, I y represents the rotational mass inertia of the base link of the exoskeleton around the ZMP, represents the angular acceleration of the center of mass, and D represents the torque arm length between the swing leg plane and the ZMP of the supporting leg, represents the acceleration of the ankle joint of the swing leg in the horizontal direction, represents the incremental acceleration of the ankle joint of the swing leg in the horizontal direction, M thigh , M shank , M foot respectively represent the mass of the thigh, the shank and the foot of the exoskeleton, respectively represent the acceleration of the center of mass of the thigh and the shank of the exoskeleton.

[0093] For example, the ZMP criterion can be used to determine whether the legs of the AutoLEE-G3 exoskeleton will rotate around the edge of the foot during walking. However, simply controlling the ZMP within the supporting polygon cannot prevent rotational movement of the exoskeleton in the direction perpendicular to the foot surface. Due to the low coefficient of friction between the exoskeleton's foot and the ground, and the lack of consideration for the dynamics of the swinging leg during the human-exoskeleton system modeling, the AutoLEE-G3 exoskeleton often fails to achieve straight-line walking along the planned trajectory. Figure 4 As shown in (B), when the upper body of the exoskeleton begins to rotate around the yaw axis, the rotational torque generated perpendicular to the supporting foot is influenced by the inertia of its swing leg and the rotational inertia of the upper body. The acceleration of each part of the swing leg can be obtained through forward kinematics calculations of the exoskeleton.

[0094] To control the momentum of the AutoLEE-G3 exoskeleton's cross-section and prevent it from deviating from the preset trajectory during straight-line motion, a yaw angle caused by the swing leg inertia and the baselink rotational inertia is designed as feedback.

[0095] A yaw momentum controller that controls the acceleration of the center of mass. The control law is based on the yaw angle θ. y and angular velocity This serves as feedback to minimize rotational friction torque.

[0096]

[0097] When the coordinates of the four control points of the exoskeleton's landing point are full b x,0 =-L / 2,b x,1 =(2b x,2 +b x,0 ) / 3,b x,2 =L / 2,b x,3 When = L / 2,

[0098]

[0099] Right now, Therefore, the control law of formula (14) can be equivalent to the following equation:

[0100]

[0101] While the accuracy of inertial moment estimation can be improved by constructing a more complex exoskeleton swing leg dynamics model, the simple and easy-to-deploy control law mentioned above is sufficient to enable the AutoLEE-G3 exoskeleton to maintain straight-line walking during actual movement.

[0102] In another implementation of the present application, a novel self-balancing lower extremity exoskeleton (AutoLEE-G3) is designed, which innovatively adopts a series-parallel hybrid configuration, and has 12 active degrees of freedom, which can meet all the movement requirements of human lower extremities. By designing a double-layer RCM mechanism, the problem of three axes being difficult to be concentric in the design of an active hip joint of an exoskeleton is solved. By adopting a quadrilateral remote transmission linkage mechanism in the knee joint and designing a novel decoupled parallel mechanism in the ankle joint, the problems of low gravity center and weak structural stiffness of the previous generation of exoskeleton systems are solved.

[0103] Example 1

[0104] In order to verify the feasibility and effectiveness of the proposed multi-level stable walking control method for realizing the stable walking of the AutoLEE-G3 exoskeleton, the formal experiment process mainly includes three parts:

[0105] 1) Verification experiment of the pitch momentum controller. Under the condition of controlling other conditions to be the same, the stability of the AutoLEE-G3 exoskeleton control system when walking on a flat ground with / without a pitch momentum controller is compared. The mass center and ZMP trajectories in the x-axis direction during the walking of the AutoLEE-G3 are collected, and the Kalman filter is used to obtain the mass center state (linear position and velocity) estimation in real time. A six-axis force torque sensor is used to assist in determining which leg is the supporting leg.

[0106] 2) Verification experiment of the yaw momentum controller. Under the condition of controlling other conditions to be the same, the yaw angle of the AutoLEE-G3 exoskeleton control system when walking on a flat ground with / without a yaw momentum controller is compared.

[0107] 3) Load-carrying stable walking experiment of different individuals. Four subjects (such as S2-S5) with different body characteristics are respectively dressed in the AutoLEE-G3 exoskeleton to walk on a flat ground, so as to verify the effectiveness and stability of the proposed multi-level stable walking control method under different load conditions. Figure 5

[0108] As shown in Figure 6 In the Cartesian space, the AutoLEE-G3 exoskeleton has a mass center trajectory generated based on the model prediction control planning of the table-cart model and a swing leg trajectory generated based on a three-order Bezier curve. Then, the reference motion trajectory of the exoskeleton in the joint space can be obtained through inverse kinematics solution. As shown in Figure 7 ​As shown, for a complete gait cycle containing left / right leg, assuming the initial position is set from the standing start, into half single support period, ZMP shifts from the center of the right leg to the right leg sole, while swinging the left leg; into the first double support period, ZMP shifts from the right leg sole to the left leg heel; then a single support period, ZMP shifts from the left leg heel to the left leg sole, while swinging the right leg; followed by the second double support period, ZMP shifts from the left leg sole to the right leg heel; and finally half single support period, ZMP shifts from the left leg heel to the left leg center, while swinging the right leg, returning to the standing posture. Thus, the entire walking cycle containing the AutoLEE-G3 exoskeleton left / right leg is completed.

[0109] The pitch momentum controller is a supplement to the model predictive control for generating a stable center of mass trajectory for the AutoLEE-G3 exoskeleton. When the center of mass trajectory generated based on the model predictive control causes the ZMP trajectory of the exoskeleton to deviate from the stable support polygon range, the pitch momentum controller starts to work and the model predictive control is not updated at this time. The pitch momentum controller is designed to adjust the center of mass trajectory to make its ZMP trajectory fall back into the stable support polygon range. Figure 7 The experimental results show that when the ZMP of the AutoLEE-G3 exoskeleton deviates from the stable support polygon range, the model predictive control stops working and the pitch momentum controller is activated to adjust the center of mass trajectory to make its ZMP trajectory fall back into the stable support polygon range. When the exoskeleton support leg is fully in contact with the ground again, the model predictive control starts to work again. The main purpose of the pitch momentum controller is to maintain the stability of the AutoLEE-G3 exoskeleton when its ZMP deviates from the stable support polygon range.

[0110] In order to verify the effectiveness of the yaw angle momentum controller based on the swing leg dynamics, two experiments were conducted:

[0111] 1) Under the condition of empty load, the change of yaw angle of the AutoLEE-G3 exoskeleton when walking on flat ground was compared when the rest of the control strategy and parameters were the same except with / without yaw angle momentum controller.

[0112] 2) Under different load conditions, the change of yaw angle of the AutoLEE-G3 exoskeleton when walking on flat ground was compared when the yaw angle momentum controller was used.

[0113] As shown in Figure 8 , under the condition of empty load (S1), the AutoLEE-G3 exoskeleton can achieve stable walking whether there is a yaw angle momentum controller or not, but when there is a yaw angle momentum controller, the walking trajectory of the exoskeleton is relatively straight, and when there is no yaw momentum control, the walking trajectory deviates greatly. This also reflects from the side that if the exoskeleton only stays in place, the swing leg cannot generate enough rotational inertia force to balance the yaw motion. Figure 8The results also show that the AutoLEE-G3 exoskeleton based on the yaw momentum controller can achieve stable walking under different load conditions (S2 to S5) and its walking trajectory has a small yaw angle.

[0114] ZMP response is a crucial indicator for evaluating the dynamic balance and stability of bipedal robots during walking tasks, directly reflecting the robot's ability to maintain a stable gait under complex motion conditions. For example... Figure 9 As shown, the ZMP response trajectories were generated under the conditions of unloaded walking and walking on a flat ground by four different subjects wearing the AutoLEE-G3 exoskeleton. The AutoLEE-G3 exoskeleton achieved a forward displacement of 1.56 meters in the X-axis direction, while exhibiting a significant center of gravity symmetry shift in the Y-axis direction. The shaded area represents the range of the support polygon determined by the walking mode parameters of the AutoLEE-G3 exoskeleton, and the reference ZMP is always located within this area, ensuring the stability and balance of the robot during movement. A detailed analysis is conducted using the ZMP response trajectory of S1 (unloaded) as an example. During the transition from the single-leg support phase to the bileg support phase, thanks to the regulation of the multi-level stable walking control method, although the actual ZMP trajectory deviated from the reference trajectory briefly, it quickly converged to the target range. Specifically, in the X-axis direction, the maximum deviation between the actual ZMP and the reference ZMP occurred at 198 seconds, with a deviation value of 0.0564 meters, while the minimum distance between the actual ZMP and the boundary of the support polygon was 0.0696 meters. In the Y-axis direction, the maximum deviation between the actual ZMP and the reference ZMP occurs at 150 seconds, with a deviation of 0.0456 meters, while the minimum distance between the actual ZMP and the boundary of the supporting polygon is 0.0479 meters. These sudden deviations are due to the premature contact between the swing leg and the ground, resulting in a large impact force between the foot and the ground. This premature contact is mainly caused by the difference between the model, the real robot, and the non-ideal environment. This impact force is unavoidable, and the manned walking exhibits the same situation as in S1. However, the actual ZMP trajectory under both unloaded and manned conditions remains within the supporting polygon area throughout the walking process, indicating that under multi-level stable walking control, subjects with different body characteristic parameters wearing the AutoLEE-G3 exoskeleton can achieve stable walking.

[0115] A series of human-exoskeleton system walking control experiments were conducted, including walking on flat ground and walking by wearers with different body characteristic parameters (height, weight, etc.), to test and verify the feasibility and robustness of the multi-loop stable walking control framework.

[0116] The application can realize the adaptive stable walking of the exoskeleton system for different subjects without relying on external devices such as trolleys or crutches to maintain the stability of the human-exoskeleton system. Compared with the gait walking of the non-self-balancing lower extremity exoskeleton auxiliary user, the multi-level stable walking control method enables the self-balancing lower extremity exoskeleton to maintain the balance of the human-machine hybrid system during walking, and breaks the limitation that the wearer needs to rely on external support equipment to maintain balance, which can be widely applied to the daily rehabilitation training and walking assistance of patients with hemiplegia, paraplegia and quadriplegia, and expands the application field. The effectiveness and feasibility of the method have been verified by the verification experiment results. The multi-level stable walking control method is not only limited to the self-balancing lower extremity exoskeleton robot, but also applicable to the fields of lower extremity rehabilitation exoskeleton robot, prosthesis and humanoid robot.

[0117] In another aspect of the application, a stable walking development and control system for a self-balancing lower extremity exoskeleton is provided, comprising:

[0118] A center of mass trajectory generation module: based on a simplified exoskeleton dynamics model of the table-cart model, an initial center of mass trajectory satisfying the stable walking requirements of the exoskeleton is generated; based on the pitch momentum controller of the linear acceleration of the center of mass, the center of mass trajectory is adjusted to obtain a real-time center of mass trajectory.

[0119] A foot point trajectory generation module: based on a third-order Bezier curve, a swing leg motion trajectory is generated, and the step length and step height of the exoskeleton foot point are changed in real time by adjusting the coordinates of the control points to generate an initial foot point trajectory; based on the yaw momentum controller of the step length adjustment, the initial foot point trajectory is adjusted to obtain a real-time foot point trajectory.

[0120] A control module: according to the real-time center of mass trajectory and the real-time foot point trajectory, the self-balancing lower extremity exoskeleton is controlled to walk stably.

[0121] The stable walking development and control system for the self-balancing lower extremity exoskeleton of the application adopts a multi-loop stable walking control framework based on model predictive control of the table-cart model, which has strong robustness to the physical parameters of the human-exoskeleton system; by designing the pitch and yaw momentum controllers as a supplement to the model predictive control, the stability of the human-exoskeleton system in the pitch and yaw directions is improved.

[0122] In another aspect of the application, an electronic device includes a processor, a memory, and a communication bus, a communication interface (Communications Interface).

[0123] Wherein:

[0124] The processor, the memory and the communication interface complete the communication among each other through the communication bus.

[0125] a communication interface configured to communicate with other electronic devices or servers.

[0126] a processor configured to execute a program, which can be configured to execute the steps of any of the above embodiments of the method for developing and controlling stable walking of a self-balancing lower extremity exoskeleton.

[0127] In particular, the program can include program code comprising computer operation instructions.

[0128] The processor can be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement one or more embodiments of the present application. The one or more processors included in the smart device can be processors of the same type, such as one or more CPUs, or processors of different types, such as one or more CPUs and one or more ASICs.

[0129] a memory configured to store the program. The memory can include a high-speed RAM memory, and can also include a non-volatile memory, such as at least one disk memory.

[0130] The program can be configured to cause the processor to perform the steps described in any of the embodiments of the method for developing and controlling stable walking of a self-balancing lower extremity exoskeleton. The specific implementation of each step in the program can refer to the corresponding description of the steps and units performed by any of the above-described embodiments of the method for developing and controlling stable walking of a self-balancing lower extremity exoskeleton, which will not be described here. Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding process descriptions in the foregoing method embodiments.

[0131] The exemplary embodiments of the present application also provide a non-transitory computer readable storage medium storing computer instructions, wherein the computer instructions are configured to cause a computer to execute the method of any of the embodiments of the present application.

[0132] The above-described methods according to embodiments of the application can be implemented in hardware, firmware, or software, or any combination thereof, and can be implemented as software storable on a recording medium which is readable from a general use computer, a special processor or programmable or special hardware (such as ASIC, or FPGA) using a recording medium, and the method described herein can be processed by such software on a recording medium. It can be understood that the computer, processor, microprocessor controller or programmable hardware includes a storage component (for example, RAM, ROM, flash memory, etc.) which can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, processor or hardware, the method described herein is implemented. Furthermore, when the general-purpose computer accesses the code for implementing the method shown herein, the execution of the code will convert the general-purpose computer into a special-purpose computer for executing the method shown herein.

[0133] So far, specific embodiments of the present application have been described. Other embodiments are within the scope of the following claims. In some cases, the acts recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve the desired results.

[0134] It should be noted that all directional directions (such as up, down, left, right, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship between components and the like in a certain order (as shown in the drawings), and if the certain order changes, the directional directions also change accordingly.

[0135] In the description of the present application, the terms "first", "second" are only used for the convenience of describing different components or names, and cannot be understood as indicating or implying the order relationship, relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can be explicitly or implicitly included at least one of the features.

[0136] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0137] It should be noted that, although the specific embodiments of the present application are described in detail with reference to the accompanying drawings, it should not be understood as limiting the scope of protection of the present application. Various modifications and variations made by those skilled in the art within the scope described in the claims are still within the scope of protection of the present application.

[0138] The examples of the embodiments of the present application are intended to simply illustrate the technical features of the embodiments of the present application, so that those skilled in the art can directly understand the technical features of the embodiments of the present application, and are not improper limitations of the embodiments of the present application.

[0139] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for developing and controlling stable walking in a self-balancing lower limb exoskeleton, characterized in that, include: A simplified exoskeleton dynamics model based on the table-cart model is used to generate an initial centroid trajectory that meets the requirements for stable walking of the exoskeleton. A pitch momentum controller based on the linear acceleration of the center of mass is used to adjust the trajectory of the center of mass to obtain the real-time trajectory of the center of mass. The swing leg motion trajectory is generated based on the third-order Bézier curve, and the step length and step height of the exoskeleton landing point are changed in real time by adjusting the coordinates of the control points to generate the initial landing point trajectory. The yaw momentum controller based on step size adjustment adjusts the initial landing point trajectory to obtain the real-time landing point trajectory. The self-balancing lower limb exoskeleton is used for stable walking control based on the real-time center of mass trajectory and the real-time foot landing point trajectory.

2. The method according to claim 1, characterized in that, The simplified exoskeleton dynamics model based on the table-cart model generates an initial centroid trajectory that meets the requirements for stable walking of the exoskeleton, including: The AutoLEE-G3 exoskeleton dynamics model is simplified based on the table-cart model to obtain a simplified exoskeleton dynamics model. By combining model predictive control and zero-moment point theory, an initial centroid trajectory that meets the requirements for stable walking of the exoskeleton is generated.

3. The method according to claim 2, characterized in that, The position of the centroid in the initial centroid trajectory and the planned landing point position satisfy the following ZMP equation relationship: Among them, H C Let g be the height of the exoskeleton's center of mass, and g be the gravitational constant. This is the acceleration of the center of gravity.

4. The method according to claim 1, characterized in that, The pitch momentum dynamics formula for the self-balancing lower limb exoskeleton when the foot rotates around the front or rear edge is expressed as follows: Among them, I p Let M be the moment of inertia of the exoskeleton about the front or rear edge of the foot, g be the acceleration due to gravity, and θ be the moment of inertia of the exoskeleton about the front or rear edge of the foot. p The angle between the upper body and the vertical direction, l foot Here, H is the length of the exoskeleton foot, H is the height from the center of mass to the foot, and d is the distance from the center of mass to the edge of rotation. Let x be the linear acceleration of the center of mass along the x-axis with the coordinates fixed on the foot.

5. The method according to claim 4, characterized in that, The linear acceleration of the centroid is expressed as: Where, θ p The pitch angle; ω is the angular velocity.

6. The method according to claim 1, characterized in that, The initial landing point trajectory is decomposed into two decoupled components along its horizontal and vertical axes, as follows: Among them, b x,i and b z,i These represent the weighting coefficients of the polynomial on the x-axis and z-axis, respectively. x,i ,b z,i (b) represents the i-th control point, (b) x,0 ,b z,0 ) and (b x,3 ,b z,3 (b) represents the start and end points respectively. x,1 ,b z,1 ) and (b x,3 ,b z,3 () indicates the control point representing the trend of the exoskeleton's landing point trajectory.

7. The method according to claim 1, characterized in that, The dynamic equation for the rotational friction torque perpendicular to the supporting foot generated by the inertia of the swinging leg and the rotational inertia of the upper body on the self-balancing lower limb exoskeleton is expressed as follows: Where, τ static τ represents the static torque when the supporting leg is not rotating. swing τ represents the rotational torque generated by the inertia of the swing leg. rotational M represents the rotational torque generated by the rotational inertia of the base linkage of the exoskeleton, Δτ represents the incremental rotational frictional torque perpendicular to the surface of the supporting foot caused by the ground reaction force, and M represents the rotational torque generated by the rotational inertia of the base linkage of the exoskeleton. e I represents the effective mass of the swing leg. y This represents the rotational mass inertia of the base link of the exoskeleton about the ZMP. Let represent the angular acceleration of the center of mass, and D represent the torque arm length from the plane of the swing leg to the ZMP of the supporting leg. This represents the horizontal acceleration of the ankle joint of the swinging leg. M represents the incremental acceleration of the ankle joint in the swinging leg in the horizontal direction. thigh M shank M foot These represent the masses of the exoskeleton's thigh, calf, and foot, respectively. These represent the accelerations at the center of mass of the exoskeleton's thigh and lower leg, respectively.

8. A stable walking development and control system for a self-balancing lower limb exoskeleton, characterized in that, include: Centroid trajectory generation module: Based on the simplified exoskeleton dynamics model of the table-cart model, it generates an initial centroid trajectory that meets the requirements for stable walking of the exoskeleton; A pitch momentum controller based on the linear acceleration of the center of mass is used to adjust the trajectory of the center of mass to obtain the real-time trajectory of the center of mass. Landing point trajectory generation module: Generates the swing leg motion trajectory based on the third-order Bézier curve, and changes the step length and step height of the exoskeleton landing point in real time by adjusting the coordinates of the control point to generate the initial landing point trajectory; Based on the step length adjustment yaw momentum controller, the initial landing point trajectory is adjusted to obtain the real-time landing point trajectory. Control module: Performs stable walking control on the self-balancing lower limb exoskeleton based on the real-time center of mass trajectory and the real-time foot landing point trajectory.

9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method for developing and controlling stable walking for a self-balancing lower limb exoskeleton as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed by a processor, implements the steps in the stable walking development and control method for a self-balancing lower limb exoskeleton as described in any one of claims 1 to 7.

Images