External thrust self-adaptive exoskeleton and safety support cooperative control method

Abstract

The invention relates to the field of robot motion control, in particular to a cooperative control method of an exoskeleton and a safety support self-adaptive to external thrust. A pressure center is represented by decoupling a dynamic model with external force in a sagittal plane; the dynamic constraint is set by introducing a component eta only containing motion of a human and an exoskeleton system in the z direction, meanwhile, a plurality of breaking line segments are adopted to approximately represent nonlinear arc setting kinematics constraint, a discrete time linear model is constructed accordingly, on this basis, a linear MPC problem is constructed, and the motion of the human and exoskeleton system in the z direction is calculated. And optimizing to obtain the optimal centroid motion trail and foothold position output of the human and exoskeleton system. Compared with the prior art, the cooperative control precision of the patient-exoskeleton-safety support system is improved.

Description

Technical Field

[0001] The present invention relates to the field of robot motion control, and in particular to a method for cooperatively controlling an exoskeleton and a safety bracket with adaptive external thrust. Background Art

[0002] Paraplegic patients who use lower limb exoskeletons to walk generally need the help of crutches to maintain their body balance, but for patients with weak upper body strength, it is very difficult to maintain balance with crutches, so a movable safety bracket is usually used to help patients maintain balance while walking. In addition, due to the heavy weight of the safety bracket, another person is often required to push it. During the pushing process, if the thrust is too large, it will cause the patient's body to lean forward, and if the thrust is insufficient, it may cause the patient's body to lean back. Therefore, when the thrust applied by the pusher is unknown, how to achieve coordinated motion control among the patient, exoskeleton, and safety bracket has become a crucial issue.

[0003] In order to help patients with weak muscle strength maintain balance during gait training, a removable safety support is essential for assisting people and exoskeletons. In recent years, many safety supports have been used to help people with gait training. CPWalker is a robotic platform for the rehabilitation of children with cerebral palsy. It allows free movement and integrates physical and cognitive interfaces into the treatment. Children with cerebral palsy can walk autonomously in a rehabilitation environment through this platform. Ai-robot is designed to help paraplegic patients walk safely and effectively, but its use needs to be learned under the guidance of experienced medical staff. Mobile Robot Balance Assistant is a device that combines the functions of a gait-assisting robot and an electric wheelchair. It can evaluate instability detection algorithms by calculating its sensitivity and specificity in identifying normal walking and simulating falls. However, these safety supports have not yet provided a satisfactory solution to the problem of human-machine collaborative motion.

[0004] At present, model predictive control (MPC) is a popular method for solving the problem of collaborative motion control of robots. MPC is usually applied to lower limb exoskeletons and humanoid robots to generate the motion trajectory and gait of the center of mass. For example, Iyer et al. used an MPC-based collaborative control method in the scene of climbing stairs to ensure that the knee joint follows the knee joint trajectory generated by optimizing the impedance model parameters; Wang et al. used MPC to generate online joint trajectories based on gait parameters such as step length, swing duration and walking speed; Jammeli et al. proposed a model predictive control framework for lower limb exoskeletons for assistance and rehabilitation. However, the above methods only consider the situation where the human-machine motion state is known, and cannot predict and control the collaborative motion of the human-machine in the case of unknown external forces. Summary of the invention

[0005] The purpose of the present invention is to provide a method for collaborative control of an exoskeleton and a safety support with adaptive external thrust, thereby improving the collaborative control accuracy of a patient-exoskeleton-safety support system.

[0006] To achieve the above object, the present invention adopts the following technical solution:

[0007] A method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust, applicable to a human-exoskeleton-safety support system, and the implementation method includes the following steps:

[0008] Step 1: construct a spring damping model of the human-exoskeleton-safety bracket system; based on the spring damping model, simplify the dynamic model of the human and exoskeleton system into a spring inverted pendulum model with horizontal external force in the XOZ plane;

[0009] Step 2: Set the dynamic constraints and kinematic constraints of the spring inverted pendulum model to complete the linearization of the kinematic model of the human and the exoskeleton system;

[0010] Step 3: Based on the assumption that the derivative of acceleration is constant within a set time, a discrete time linear model is established in the sagittal plane to represent the dynamic constraints and kinematic constraints obtained in step 2;

[0011] Step 4: construct an adaptive collaborative control task target of the MPC function of the human and the exoskeleton system according to the expected speed, center of mass height, center of pressure, and derivative of acceleration;

[0012] Step 5. Use the discrete-time linear model obtained in step 3 to convert the MPC function to be solved into a linear MPC problem model. Input the current center of mass position of the person and the exoskeleton system and the current foothold position into the linear MPC problem, predict and optimize the acceleration derivative of the center of mass motion of the person and the exoskeleton system at the next moment and the foothold position, generate the center of mass motion trajectory and foothold position of the person and the exoskeleton system, and realize the coordinated control of the person-exoskeleton-safety support system.

[0013] Furthermore, the implementation method of step 1 includes the steps of:

[0014] 1.1. The spring damping model of the human-exoskeleton-safety bracket system is established as follows:

[0015]

[0016] Where F represents the force generated by the spring damping system, k k represents the stiffness coefficient of the spring-damper system, k c represents the damping coefficient of the spring damping system, x represents the position of the center of mass of the human and the exoskeleton system in the x direction, represents the speed of the person and the exoskeleton system in the x direction, x d represents the position of the safety bracket in the x direction, represents the speed of the safety bracket in the x direction;

[0017] 1.2. Based on the spring damping model, the dynamic model of the human and exoskeleton system is simplified to a spring inverted pendulum model with horizontal external force in the XOZ plane (sagittal plane):

[0018]

[0019] in, represents the constraint range of the pressure center, x represents the position of the center of mass of the human and the exoskeleton system in the x direction, z represents the position of the center of mass of the human and the exoskeleton system in the z direction, represents the acceleration of the center of mass of the human and the exoskeleton system in the z direction, g represents the acceleration due to gravity, and m represents the mass of the human and the exoskeleton system. represents the acceleration of the center of mass of the human and the exoskeleton system in the x direction, P cop Indicates the location of the center of pressure of the human and the exoskeleton system.

[0020] Furthermore, the implementation method of step 2 includes the steps of:

[0021] By introducing the component η, the motion of the dynamic model of the human and the exoskeleton system in the x-direction and the z-direction is decoupled to realize the constraints of the spring inverted pendulum model in the x-direction and the z-direction;

[0022] The dynamic constraints between the human and the exoskeleton system in the x-direction are expressed as:

[0023]

[0024] Where x represents the position of the center of mass of the person and the exoskeleton system in the x direction; η represents the component that only includes the movement of the person and the exoskeleton system in the z direction, The constraint range of η is: Where η represents the upper limit of the movement of the human and the exoskeleton system in the z direction, represents the lower bound of the movement of the human and the exoskeleton system in the z direction; P cop Indicates the position of the heel. Indicates the position of the toes;

[0025] The dynamic constraints between the human and the exoskeleton system in the z direction are expressed as:

[0026]

[0027] in, represents the acceleration of the center of mass of the human and the exoskeleton system in the z direction, and g represents the acceleration due to gravity;

[0028] The kinematic constraints are implemented using multiple segmented line segments, and the dynamic constraints are expressed as:

[0029] A c (cp)≤b

[0030] Among them, A c represents the matrix of linear kinematic constraints, b represents the vector of linear kinematic constraints, c represents the center of mass position of the human and the exoskeleton system; p represents the position of the contact point between the human and the exoskeleton system and the ground.

[0031] Furthermore, the implementation method of step 3 includes the steps of:

[0032] Based on the fact that the derivative of acceleration per unit time is constant, the motion state of the human and the exoskeleton system is spatially discretized to obtain a discrete-time linear model for representing the dynamic and kinematic constraints of the human and the exoskeleton system. The discrete-time linear model of the human and the exoskeleton system in the x direction is as follows:

[0033] The discrete-time state at the k+1th time step:

[0034]

[0035]

[0036]

[0037] Where t represents the discrete time period, x k represents the position of the person and the exoskeleton system in the x-direction at the kth time step, represents the speed of the human and the exoskeleton system in the x-direction at the kth time step, represents the acceleration of the human and the exoskeleton system in the x-direction at the kth time step;

[0038] The state matrix (position, velocity, acceleration) of N time steps:

[0039]

[0040] The matrix of acceleration derivatives for N time steps:

[0041]

[0042] The result obtained by N-step iterative calculation of the discrete time state of the k+1th time step is:

[0043]

[0044] Among them, A represents the state transfer matrix, B represents the control input vector, represents the control input from the kth moment to the k+N-1th moment, Indicates the current status;

[0045] The discrete-time state coefficient matrix P ps , P pu , P vs , P vu , P as and P au , are calculated from the matrix of acceleration derivatives of N time steps of N iterations;

[0046] The discrete-time linear model of the human and the exoskeleton system in the z direction is expressed in the same form as the discrete-time linear model of the human and the exoskeleton system in the x direction.

[0047] Furthermore, the implementation method of step 4 comprises the steps of:

[0048] According to the expected speed, center of mass height, center of pressure, and derivative of acceleration, the adaptive collaborative control task objective of the MPC function of the human and exoskeleton system is constructed as follows:

[0049]

[0050]

[0051] d3=||zh ref || 2

[0052]

[0053] in, is the desired walking speed, h ref is the target of the center of mass height control task in the support state, is the acceleration derivative of the center of mass of the human and the exoskeleton system in the x direction, It is the acceleration derivative of the center of mass of the human and the exoskeleton system in the z direction; d1 represents the speed control task target, d2 represents the pressure center control task target, and d3 represents the center of mass height control task target.

[0054] Furthermore, the implementation method of step 5 comprises the steps of:

[0055] The acceleration derivatives of the center of mass of the human and the exoskeleton system in the x-direction and the z-direction, and the position of the foothold are selected as decision variables;

[0056] The dynamic constraints and kinematic constraints obtained in step 2 are regarded as inequality constraints in the linear MPC problem; the objective functions d1, d2, d3 and d4 of the linear MPC problem are assigned corresponding weights α1, α2, α3 and α4; the control cycle and the prediction time step N of the MPC problem are set according to the application scenario; the linear MPC framework is obtained as follows:

[0057]

[0058] The discrete-time linear model obtained in step 3 is used to convert the MPC framework into the standard form of the quadratic programming solver. The conversion from the linear MPC framework to the standard quadratic programming solver form is expressed as:

[0059]

[0060] in, and All dynamic and kinematic constraints are included; decision variables It includes the acceleration derivatives in the x-direction and the z-direction of N time steps, as well as the position of the foothold; Q represents the Hessian matrix of the quadratic programming solver, is the original Hessian matrix, expressed as:

[0061]

[0062] in, represents the zero matrix, and the components of the original Hessian matrix Q0 are expressed as:

[0063]

[0064] R 13 = -α3(P pu -λP au ) T

[0065]

[0066] R 31 = -α3(P pu -λP au )

[0067] R 33 =α3I

[0068] in, is the identity matrix;

[0069] The gradient vector of the original objective function in the MPC framework is expressed as:

[0070] f=[f1,f2,f3] T

[0071] The components of the gradient vector f are expressed as:

[0072]

[0073] is a vector of all 1s, and

[0074] The center of mass position of the current person and exoskeleton system and the current foothold position are input into the standard form of the quadratic programming solver for optimization and solution, and the acceleration derivative of the center of mass motion of the person and exoskeleton system and the foothold position at the next moment are predicted and optimized; on this basis, the center of mass motion trajectory and foothold position of the person and exoskeleton system are generated to achieve coordinated control of the person-exoskeleton-safety support system.

[0075] The present invention provides a method for cooperative control of an exoskeleton and a safety support with adaptive external thrust, which represents the pressure center by decoupling the dynamic model with external force in the sagittal plane; sets the dynamic constraint by introducing a component η that only includes the motion of the human and the exoskeleton system in the z direction, and uses multiple broken line segments to approximate the nonlinear arc to set the kinematic constraint, and constructs a discrete time linear model based on this, and on this basis, constructs a linear MPC problem, and optimizes the optimal quality center motion trajectory and foothold position output of the human and the exoskeleton system. Compared with the prior art, the present invention improves the cooperative control accuracy of the patient-exoskeleton-safety support system. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] The present invention will be further described below with reference to the accompanying drawings, in which:

[0077] Figure 1 It is the overall framework diagram of the algorithm proposed in the present invention;

[0078] Figure 2 is a schematic diagram of a human-exoskeleton-safety support system in the present invention;

[0079] Figure 3 It is a simplified schematic diagram of the kinetic model of the present invention;

[0080] Figure 4 It is a schematic diagram of parameter linearization introduced in the present invention;

[0081] Figure 5 It is a simplified schematic diagram of the kinematic model of the present invention. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0083] like Figure 1 As shown, the present embodiment provides a method for cooperatively controlling an exoskeleton and a safety bracket with adaptive external thrust, comprising the following steps:

[0084] Step 1: Construct a spring damping model of the human-exoskeleton-safety bracket system. Based on the spring damping model, simplify the dynamic model of the human and exoskeleton system into a spring inverted pendulum model with horizontal external force in the XOZ plane. The implementation method is as follows:

[0085] 1.1. Construct the spring damping model of the human-exoskeleton-safety bracket system:

[0086] like Figure 2 In the human-exoskeleton-safety support system, the medical staff pushes the safety support from behind, and the patient wears the exoskeleton to complete gait training. The four wheels of the safety support are passive movements, and the vertical support joints are used to support the weight of the person and the exoskeleton system. Connected between the safety support and the exoskeleton is a spring damping system that can be extended and shortened to adjust the interaction force and relative position. In the entire system, on the one hand, the force generated by the spring damping system is affected by the movement of the person, the exoskeleton system, and the safety support; on the other hand, since the spring damping system always moves in the horizontal direction, the force it generates is always along the x-axis direction. The spring damping model is:

[0087]

[0088] In formula (1), F is the force generated by the spring damping system, k k is the stiffness coefficient of the spring-damper system, k c is the damping coefficient of the spring-damper system, x is the position of the person and the exoskeleton system in the x direction, is the speed of the person and the exoskeleton system in the x direction, x d is the position of the safety bracket in the x direction, are the speeds of the safety bracket in the x direction respectively.

[0089] 1.2. Based on the spring damping model, the dynamic model of the human and exoskeleton system is simplified to a spring inverted pendulum model with horizontal external force in the XOZ plane:

[0090] The simplified structure of the dynamic model of the human and exoskeleton system in this embodiment is as follows: Figure 3 As shown in the figure, for humans and exoskeleton systems, the dynamic equation of their center of mass can be expressed as:

[0091]

[0092] In formula (2) and formula (3), m is the mass of the human and the exoskeleton system, c is the center of mass of the human and the exoskeleton system, is the gravitational acceleration vector, is the ground reaction force, L is the angular momentum of the person and the exoskeleton system, and p is the position of the contact point with the ground.

[0093] According to equations (2) and (3), the external torque divided by the Newton equation in the z direction is expressed as:

[0094]

[0095] The superscript z indicates the component in the z direction, and G z It is g, usually taken as 9.81m / s.

[0096] Decomposing equation (4) on the sagittal plane, we get the following expression:

[0097]

[0098] In formula (5), the superscripts x and z represent the components in the x-direction and the z-direction, respectively; Indicates the center of pressure.

[0099] Considering that the foothold is always on the ground (only considering the movement on flat ground), we have p z = 0. So equation (5) can be rewritten as:

[0100]

[0101] In formula (6), P cop Indicates the location of the pressure center of the human and the exoskeleton system; Indicates the constraint range of the pressure center, which is used to ensure the stability of the movement of the human and the exoskeleton system.

[0102] So far, a spring inverted pendulum model with horizontal external force of the human and the exoskeleton system in the XOZ plane is obtained.

[0103] Step 2: Establish a dynamic constraint model and a kinematic constraint model of the spring inverted pendulum model to complete the linearization of the kinematic model of the human and exoskeleton system;

[0104] The dynamic constraints are established by introducing the parameter η:

[0105] As shown in Formula 6, for the inverted pendulum model constructed by a human and exoskeleton system with a spring damping system, it is necessary to ensure that its pressure center is within the range that satisfies the constraint conditions; and the dynamic model of the human and exoskeleton system is linear in the x direction and nonlinear in the z direction. Based on this, in order to ensure its linearization in the z direction, this embodiment defines a component η, which only includes the movement of the human and exoskeleton system in the z direction. The definition of η is

[0106] Since the condition for the human and the exoskeleton system to meet the stability constraint is that the pressure center must move within the range of the edge of the sole; according to formula (6), the change of η can reasonably limit the movement of the human and the exoskeleton system in the z direction to ensure that the pressure center meets the edge of the sole. So the constraint range of η can be expressed as where η and Respectively represent the upper and lower bounds; its linearization constraints within the plantar edge range are as follows Figure 4 shown.

[0107] The dynamic constraints between the human and the exoskeleton system in the x-direction can be expressed as:

[0108]

[0109] Among them, {P cop} represents the set of pressure center positions.

[0110] In equation (7), the center of pressure cannot be directly calculated. However, since the movement occurs in the sagittal plane, the position of the heel can be determined as cop and toe position And the pressure center is constrained in it. Formula (13) can be rewritten as:

[0111]

[0112] In addition, it can be noted that is always positive. Therefore, the dynamic constraints between the human and the exoskeleton system in the z direction can be expressed as:

[0113]

[0114] Formula (8) and Formula (9) are the dynamic constraint models of the human and the exoskeleton system. This embodiment successfully decouples the dynamic constraints in the x and z directions by introducing the parameter η. At the same time, the linearization of η also ensures that the movement of the human and the exoskeleton system in the z direction presents linear characteristics. Based on the above situation, the linearization of the dynamic constraints of the human and the exoskeleton system is finally completed.

[0115] Set up kinematic constraints:

[0116] The data collected from normal human walking show that when one leg is in a supporting state and the other leg is in a swinging state, the movement of the center of mass of the human and the exoskeleton system in the sagittal plane is close to a circular arc. The kinematic constraints of the human and the exoskeleton system can be expressed as {l|cp a}≤L0, where {l|cp a} represents the distance p from the center of mass c of the human and the exoskeleton system to the ankle joint position a , L0 is the length of the legs of the person and the exoskeleton system.

[0117] In order to linearize the kinematic constraints of the human and the exoskeleton system, multiple segmented line segments are used to approximate the nonlinear arc. In practical applications, the more segmented line segments there are, the closer they are to the actual constraint conditions, but the greater the impact on the subsequent MPC calculation amount. The linear kinematic constraints of the human and the exoskeleton system can be expressed as:

[0118] A c (cp)≤b, (10)

[0119] Among them, A c and b are the matrix and vector of linear kinematic constraints, respectively.

[0120] In order to ensure that the center of mass of the person and the exoskeleton system reaches the highest height when standing upright, this embodiment simplifies the kinematic model of the person and the exoskeleton system to constrain the center of mass within a five-segment line representation range, thereby completing the linearization of the kinematic model. The segmented representation of the kinematic constraints in this embodiment is as follows: Figure 5 As shown, Figure 5 Among them, l1, l2, l3, l4 and l5 are 5 segmented line segments, one of the five segmented line segments sets the lowest position of the centroid, and the remaining segmented line segments are set to even numbers.

[0121] Step 3: Based on the assumption that the derivative of acceleration is constant in a very short time, a discrete time linear model for representing dynamic constraints and kinematic constraints is established in the sagittal plane:

[0122] Assuming that the acceleration (derivative of acceleration) of the center of mass is constant within a set time, the time can be considered as a discrete time period. represents the state of the human and the exoskeleton system at the kth time step in the x direction, represents the derivative of acceleration, then the discrete-time state of the k+1th time step is It can be expressed as:

[0123]

[0124] In formula (11), t represents the discrete time period, x krepresents the position of the person and the exoskeleton system in the x-direction at the kth time step, represents the speed of the human and the exoskeleton system in the x-direction at the kth time step, Represents the acceleration of the human and the exoskeleton system in the x-direction at the kth time step.

[0125] The state matrix (position, velocity, acceleration) of N time steps is expressed as:

[0126]

[0127] The matrix representation of the acceleration derivatives for N time steps is:

[0128]

[0129] The result of N-step iterative calculation based on formula (11) can be expressed as:

[0130]

[0131] Discrete-time state coefficient matrix P ps , P pu , P vs , P vu , P as and P au , which is calculated by iterating N steps of equation (16). Similarly, the motion of the human and the exoskeleton system in the z direction can also be expressed in the same form of equations (11)-(16).

[0132] Step 4: Construct the adaptive collaborative control task target of the MPC function of the human and exoskeleton system, and set the expected speed, center of mass height, center of pressure, and derivative of acceleration:

[0133] In the MPC framework, the horizontal derivative of the acceleration of the center of mass is used and the vertical acceleration derivative Assume that the person behind the vehicle applies a variable range of force to push the safety support to walk, and hopes that the person and the exoskeleton system can move at a set speed, and this speed is adjusted within a variable range. The task goal can be expressed as:

[0134]

[0135] in, is the desired walking speed.

[0136] Dynamic constraints are used to ensure that the position of the pressure center cannot exceed the edge of the supporting sole; during the movement of the human-exoskeleton-safety bracket, the position of the pressure center of the human and the exoskeleton system is as close as possible to the middle position of the supporting sole edge to maintain the stability of the movement. The task goal can be expressed as:

[0137]

[0138] Considering that the exoskeleton is connected to the human, the task goal of making the walking motion more anthropomorphic can be expressed as:

[0139] d3=||zh ref || 2 , (19)

[0140] Among them, h ref is the height of the center of mass in the supported state; in this project mission objective, in order to make the exoskeleton in a human upright walking state rather than a robot's bent knee walking state, the height of the center of mass movement should be as high as possible.

[0141] Since the acceleration derivatives of the center of mass of the human and the exoskeleton system in the x and z directions are the decision variables of the MPC function, the acceleration derivative is required to be minimized to ensure the smoothness of the motion and prevent sudden changes in the motion. , the task objective can be expressed as:

[0142]

[0143] in, is the acceleration derivative of the center of mass of the human and the exoskeleton system in the x direction, is the acceleration derivative of the center of mass of the human and the exoskeleton system in the z direction.

[0144] The above d1, d2, d3 and d4 are the task objectives of the adaptive cooperative control of the MPC function.

[0145] Step 5: Establish a linear MPC problem, predict and optimize the derivative of the center of mass motion acceleration and the position of the foothold, and generate the center of mass motion trajectory and foothold position of the human and exoskeleton system:

[0146] In order to ensure that the human, the exoskeleton system and the safety bracket can maintain coordinated motion, it is necessary to set the adaptive coordinated control target weight. On this basis, with the help of linearized dynamics formula (8) and formula (9) and kinematic constraint formula (10), the original model predictive control (MPC) problem can be transformed into a convex problem (convex problems have a unique optimal solution). Specifically:

[0147] The acceleration derivatives of the center of mass of the human and the exoskeleton system in the x and z directions and the position of the foothold are selected as decision variables. In this embodiment, formula 8) and formula (9) representing the dynamic constraint model and formula (10) representing the kinematic constraint model are regarded as inequality constraints in the linear MPC problem. At the same time, for the objective functions d1, d2, d3 and d4 of the linear MPC problem, corresponding weights α1, α2, α3 and α4 are assigned to them. The control period is set to 0.01s, and the prediction time step N of the MPC problem is 15. The specific linear MPC framework can be expressed as:

[0148]

[0149] The linear MPC framework in Eq. (21) is converted to the standard form of the quadratic programming solver through the discrete-time state-space equations in Eqs. (11)-(15). The conversion from the linear MPC framework to the standard quadratic programming solver form is expressed as:

[0150]

[0151] in, and Contains the entire dynamic constraints and kinematic constraints; decision variables Contains the acceleration derivatives in the x and z directions and the position of the footstep for N time steps.

[0152] The Hessian matrix Q input to the quadratic programming solver is expressed as The original Hessian matrix It is expressed as:

[0153]

[0154] in, represents the zero matrix, and the components of the original Hessian matrix Q0 are expressed as:

[0155]

[0156] R 13 = -α3(P pu -λP au ) T , (25)

[0157]

[0158] R 31 = -α3(P pu -λP au ), (27)

[0159] R 33 =α3I, (28)

[0160] in, is the identity matrix.

[0161] The gradient vector of the original objective function in the MPC framework is expressed as:

[0162] f=[f1,f2,f3] T , (29)

[0163] Among them, the components of the gradient vector f are expressed as:

[0164]

[0165]

[0166]

[0167] in, is a vector of all 1s, and

[0168] The matrix Q is processed to ensure that the Hessian matrix in the quadratic programming problem is positive semidefinite. The first step of each prediction result is obtained using the optimal decision variables. The first step of each prediction result can be used to obtain the state of the MPC calculation at the next moment through formula (11).

[0169] The moment when the person and the exoskeleton system reach the highest point is selected from the standard quadratic programming solver as the moment of gait switching, and the moment when the foot reaches the lowest point is selected as the moment of foot contact. In addition, the PVT (Position-Velocity-Time) curve interpolation method is used to plan the motion trajectory of the foot in each swing phase. The joint motion trajectory of the person and the exoskeleton system during the movement is obtained through pseudo-inverse matrix and inverse kinematics. Joint control is achieved by controlling the position of the joints, which is achieved by setting PID parameters.

[0170] At this point, the coordinated motion trajectory of the human and the exoskeleton system and the position of the foothold can be generated. Therefore, the present invention can generate the coordinated motion of the human-exoskeleton-safety support under the action of artificial external force.

Claims

1. A method for cooperative control of an exoskeleton and a safety support with adaptive external thrust, applicable to a human-exoskeleton-safety support system, characterized in that: The method comprises the following steps: Step 1: construct a spring damping model of the human-exoskeleton-safety bracket system; based on the spring damping model, simplify the dynamic model of the human and exoskeleton system into a spring inverted pendulum model with horizontal external force in the XOZ plane; Step 2: Set the dynamic constraints and kinematic constraints of the spring inverted pendulum model to complete the linearization of the kinematic model of the human and the exoskeleton system; Step 3: Based on the assumption that the derivative of acceleration is constant within a set time, a discrete time linear model is established in the sagittal plane to represent the dynamic constraints and kinematic constraints obtained in step 2; Step 4: construct an adaptive collaborative control task target of the MPC function of the human and the exoskeleton system according to the expected speed, center of mass height, center of pressure, and derivative of acceleration; Step 5. Use the discrete-time linear model obtained in step 3 to convert the MPC function to be solved into a linear MPC problem model. Input the current center of mass position of the person and the exoskeleton system and the current foothold position into the linear MPC problem, predict and optimize the acceleration derivative of the center of mass motion of the person and the exoskeleton system at the next moment and the foothold position, generate the center of mass motion trajectory and foothold position of the person and the exoskeleton system, and realize the coordinated control of the person-exoskeleton-safety support system.

2. A method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust according to claim 1, characterized in that: The implementation method of step 1 comprises the following steps: 1.

1. The spring damping model of the human-exoskeleton-safety bracket system is established as follows: Where F represents the force generated by the spring damping system, k k represents the stiffness coefficient of the spring-damper system, k c represents the damping coefficient of the spring damping system, x represents the position of the person and the exoskeleton system in the x direction, represents the speed of the person and the exoskeleton system in the x direction, x d represents the position of the safety bracket in the x direction, represents the speed of the safety bracket in the x direction; 1.

2. Based on the spring damping model, the dynamic model of the human and exoskeleton system is simplified to a spring inverted pendulum model with horizontal external force in the sagittal plane: in, represents the constraint range of the pressure center, x represents the position of the center of mass of the human and the exoskeleton system in the x direction, z represents the position of the center of mass of the human and the exoskeleton system in the z direction, represents the acceleration of the center of mass of the human and the exoskeleton system in the z direction, g represents the acceleration due to gravity, and m represents the mass of the human and the exoskeleton system. represents the acceleration of the center of mass of the human and the exoskeleton system in the x direction, P cop Indicates the location of the center of pressure of the human and the exoskeleton system.

3. A method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust according to claim 2, characterized in that: The implementation method of step 2 comprises the following steps: By introducing the component η, the motion of the dynamic model of the human and the exoskeleton system in the x-direction and the z-direction is decoupled to realize the constraints of the spring inverted pendulum model in the x-direction and the z-direction; The dynamic constraints between the human and the exoskeleton system in the x-direction are expressed as: Where x represents the position of the center of mass of the person and the exoskeleton system in the x direction; η represents the component that only includes the movement of the person and the exoskeleton system in the z direction, The constraint range of η is: Where η represents the upper limit of the movement of the human and the exoskeleton system in the z direction, It represents the lower bound of the motion of the human and the exoskeleton system in the z direction; P cop Indicates the position of the heel. Indicates the position of the toes; The dynamic constraints between the human and the exoskeleton system in the z direction are expressed as: in, represents the acceleration of the center of mass of the human and the exoskeleton system in the z direction, and g represents the acceleration due to gravity; The kinematic constraints are implemented using multiple segmented line segments, and the dynamic constraints are expressed as: A c (c-p)≤b Among them, A c represents the matrix of linear kinematic constraints, b represents the vector of linear kinematic constraints, c represents the center of mass position of the human and the exoskeleton system; p represents the position of the contact point between the human and the exoskeleton system and the ground.

4. A method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust according to claim 3, characterized in that: The implementation method of step 3 comprises the following steps: Based on the fact that the derivative of acceleration per unit time is constant, the motion state of the human and the exoskeleton system is spatially discretized to obtain a discrete-time linear model for representing the dynamic and kinematic constraints of the human and the exoskeleton system. The discrete-time linear model of the human and the exoskeleton system in the x direction is as follows: The discrete-time state at the k+1th time step: Where t represents the discrete time period, x k represents the position of the person and the exoskeleton system in the x-direction at the kth time step, represents the speed of the human and the exoskeleton system in the x-direction at the kth time step, represents the acceleration of the human and the exoskeleton system in the x-direction at the kth time step; The state matrix (position, velocity, acceleration) of N time steps: The matrix of acceleration derivatives for N time steps: The result obtained by N-step iterative calculation of the discrete time state of the k+1th time step is: Among them, A represents the state transfer matrix, B represents the control input vector, represents the control input from the kth moment to the k+N-1th moment, Indicates the current status; Discrete-time state coefficient matrix P ps , P pu , P vs , P vu , P as and P au , are calculated from the matrix of acceleration derivatives of N time steps of N iterations; The discrete-time linear model of the human and the exoskeleton system in the z direction is expressed in the same form as the discrete-time linear model of the human and the exoskeleton system in the x direction.

5. The method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust according to claim 4, characterized in that: The implementation method of step 4 comprises the following steps: According to the expected speed, center of mass height, center of pressure, and derivative of acceleration, the adaptive collaborative control task objective of the MPC function of the human and exoskeleton system is constructed as follows: d3=||zh ref || 2 in, is the desired walking speed, h ref is the target of the center of mass height control task in the support state, is the acceleration derivative of the center of mass of the human and the exoskeleton system in the x direction, It is the acceleration derivative of the center of mass of the human and the exoskeleton system in the z direction; d1 represents the speed control task target, d2 represents the pressure center control task target, and d3 represents the center of mass height control task target.

6. A method for cooperatively controlling an exoskeleton and a safety support with adaptive external thrust according to claim 5, characterized in that: The implementation method of step 5 comprises the following steps: The acceleration derivatives of the center of mass of the human and the exoskeleton system in the x-direction and the z-direction, and the position of the foothold are selected as decision variables; The dynamic constraints and kinematic constraints obtained in step 2 are regarded as inequality constraints in the linear MPC problem; the objective functions d1, d2, d3 and d4 of the linear MPC problem are assigned corresponding weights α1, α2, α3 and α4; the control cycle and the prediction time step N of the MPC problem are set according to the application scenario; The linear MPC framework is obtained as: The discrete-time linear model obtained in step 3 is used to convert the MPC framework into the standard form of the quadratic programming solver. The conversion from the linear MPC framework to the standard quadratic programming solver form is expressed as: s.t.A qp u k ≤b qp in, and All dynamic and kinematic constraints are included; decision variables It includes the acceleration derivatives in the x-direction and the z-direction of N time steps, as well as the position of the foothold; Q represents the Hessian matrix of the quadratic programming solver, is the original Hessian matrix, expressed as: in, represents the zero matrix, and the components of the original Hessian matrix Q0 are expressed as: R 13 =-α3(P pu -λP au ) T R 31 =-α3(P pu -λP au ) R 33 =α3I in, is the identity matrix; The gradient vector of the original objective function in the MPC framework is expressed as: <h2 style=";text-align:left;direction:ltr">f = [f1,f2,f3]<h2 style=";text-align:left;direction:ltr"> T The components of the gradient vector f are expressed as: is a vector of all 1s, and The center of mass position of the current person and exoskeleton system and the current foothold position are input into the standard form of the quadratic programming solver for optimization and solution, and the acceleration derivative of the center of mass motion of the person and exoskeleton system and the foothold position at the next moment are predicted and optimized; on this basis, the center of mass motion trajectory and foothold position of the person and exoskeleton system are generated to achieve coordinated control of the person-exoskeleton-safety support system.

Images