Robot ground contact force estimation method and system based on high-order finite time observer

Abstract

The present disclosure relates to the technical field of foot-type robot control, and provides a robot touch ground force estimation method and system based on a high-order finite time observer, which comprises: establishing a foot-type robot leg dynamics model; selecting joint position, velocity and joint acceleration driving term as state variables, and establishing a foot-type robot leg state space model; according to the established state space model, using step-by-step iteration of the estimated value of the state variable, combining the power form of the nonlinear sign function and the set gain parameter matrix, and establishing a high-order finite time observer through a recursive mode to obtain the estimation of the foot-type robot foot end touch ground force. The high-order finite time observer is established in combination with the nonlinear sign function and the designed parameter matrix, so that the touch ground force estimation error can converge within a certain boundary in a limited time, and the problems of low touch ground force estimation accuracy and slow change touch ground force estimation of the traditional extended state observer are effectively solved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of foot-type robot control, in particular, to a robot touch force estimation method and system based on a high-order finite time observer. BACKGROUND

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute the prior art.

[0003] Foot-type robots adopt bionic design, and due to their discrete landing points, they exhibit superior adaptability in complex terrain environments. With the development of technology, foot-type robots are widely used in people's production and life, such as industrial inspection, disaster rescue, military operations, scientific research and education, etc. Effective and friendly physical contact between the foot end of the foot-type robot and the ground is a key factor to achieve efficient and stable movement. Such contact not only affects the walking efficiency and stability of the foot-type robot, but also relates to its ability to adapt to various terrain conditions and perform various tasks. The essence of achieving safe and friendly contact is to effectively control the touch force of the robot foot end, which requires accurate measurement of the foot end touch force.

[0004] The inventors found in their research that there are generally two ways to measure the touch force of the foot end of the foot-type robot at present: the first is to directly detect through the installation of a contact force sensor at the foot end; the second is to indirectly estimate through body sensors (such as joint torque sensors, displacement sensors, body posture sensors, etc.). The method of installing a force sensor at the foot end increases the difficulty of leg structure design and increases the weight of the foot end, which is not conducive to the motion control of the foot-type robot. The method of estimating using body sensors avoids the installation of foot end contact force sensors, mainly including using whole body dynamics and leg dynamics, using whole body dynamics estimation, which has a complex model and a huge amount of calculation, and requires high hardware requirements for actual deployment; the traditional estimation method using body sensors combined with leg dynamics is to estimate through an extended state observer, which has limited estimation performance and can only effectively estimate the contact force that changes slowly. For the touch force of the foot-type robot, which changes rapidly, the observation accuracy is insufficient. SUMMARY

[0005] To solve the above problems, the present disclosure provides a robot touch force estimation method and system based on a high-order finite time observer, which improves the estimation accuracy and efficiency of the touch force of the foot-type robot.

[0006] To achieve the above purpose, the present disclosure adopts the following technical solutions:

[0007] One or more embodiments provide a robot touch force estimation method based on a high-order finite time observer, comprising the following steps:

[0008] establish a leg dynamics model of a leg of a legged robot;

[0009] select joint position, velocity and joint acceleration driving term as state variable, and establish a state space model of the leg of the legged robot;

[0010] According to the established state space model, the estimation value of the state variable is iterated step by step, the power form of the nonlinear symbolic function and the set gain parameter matrix are combined, and a high-order finite time observer is established through recursion to obtain the estimation of the foot end ground force of the legged robot.

[0011] One or more embodiments provide a robot ground force estimation system based on a high-order finite time observer, comprising:

[0012] The dynamics model construction module is configured to establish a leg dynamics model of a leg of a legged robot;

[0013] The state space model construction module is configured to select joint position, velocity and joint acceleration driving term as state variable, and establish a state space model of the leg of the legged robot;

[0014] The estimation module is configured to establish a high-order finite time observer through recursion according to the established state space model, the estimation value of the state variable is iterated step by step, the power form of the nonlinear symbolic function and the set gain parameter matrix are combined, and the estimation of the foot end ground force of the legged robot is obtained.

[0015] An electronic device includes a memory and a processor, and computer instructions stored on the memory and running on the processor, when the computer instructions are run by the processor, the steps in the robot ground force estimation method based on the high-order finite time observer are completed.

[0016] A computer readable storage medium for storing computer instructions, when the computer instructions are executed by a processor, the steps in the robot ground force estimation method based on the high-order finite time observer are completed.

[0017] Compared with the prior art, the beneficial effects of the present disclosure are:

[0018] The present disclosure is based on the dynamics model of the leg of the legged robot, the state space equation is constructed by selecting the state variable, the high-order finite time observer is designed to observe the ground force of the legged robot, the high-order finite time observer is established through recursion by combining the nonlinear symbolic function and the designed parameter matrix, the estimation error of the ground force can be converged to a certain boundary within a limited time, and the problems of low estimation accuracy of the traditional extended state observer and only slow changing ground force can be effectively solved.

[0019] Advantages of the present disclosure and additional aspects will be more fully understood in view of the following detailed description, from the specific examples. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which form a part of the present disclosure, are intended to provide further understanding of the present disclosure and are incorporated herein in

[0021] Figure 1 is a flow chart of the ground contact force estimation method of embodiment 1 of the present disclosure;

[0022] Figure 2 is a foot robot leg simulation model diagram of the simulation example of embodiment 1 of the present disclosure;

[0023] Figure 3 is a comparison chart of the ground contact force estimation performance of the high-order finite time observer and the extended state observer of the present embodiment when the ground contact force is step type in the simulation example of embodiment 1 of the present disclosure;

[0024] Figure 4 is a comparison chart of the ground contact force estimation performance of the high-order finite time observer and the extended state observer of the present embodiment when the ground contact force is ramp type in the simulation example of embodiment 1 of the present disclosure;

[0025] Figure 5 is a comparison chart of the ground contact force estimation performance of the high-order finite time observer and the extended state observer of the present embodiment when the ground contact force is free fall in the simulation example of embodiment 1 of the present disclosure;

[0026] Wherein: 1, fuselage, 2, hip, 3, thigh, 4, shank, 5, ground, 6, support. DETAILED DESCRIPTION

[0027] The present disclosure will be further described below in conjunction with the drawings and examples.

[0028] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the present disclosure. 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 the present disclosure belongs.

[0029] It is to be noted that the terms used herein are merely for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprise" and / or "include" when used in this specification, specify the presence of stated features, steps, operations, devices, components and / or combinations thereof, but do not preclude the presence or addition of one or more other features, steps, operations, devices, components and / or combinations thereof. It should be noted that the various embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict, which will be described in detail below in conjunction with the drawings.

[0030] Embodiment 1

[0031] In the technical solutions disclosed in one or more embodiments, as shown in the accompanying drawings, a robot ground contact force estimation method based on a high-order finite time observer includes the following steps: Figures 1 to 5

[0032] Step 1, establishing a leg dynamics model of a legged robot;

[0033] Step 2, selecting joint position, velocity and joint acceleration driving term as state variables, and establishing a leg state space model of the legged robot;

[0034] Step 3, according to the established state space model, using step-by-step iteration of the estimated value of the state variable, combining the power form of the nonlinear sign function and the designed gain parameter matrix, and establishing a high-order finite time observer through a recursive manner to obtain the estimation of the foot end ground contact force of the legged robot.

[0035] In this embodiment, based on the leg dynamics model of the legged robot, the state space equation is constructed by selecting the state variable, the high-order finite time observer is designed to observe the ground contact force of the legged robot, the high-order finite time observer is established through a recursive manner by combining the nonlinear sign function and the designed parameter matrix, and the ground contact force estimation error can be converged within a certain boundary in a finite time, effectively solving the problems of low estimation accuracy of the traditional extended state observer for the ground contact force and the inability to estimate the slowly changing ground contact force.

[0036] In step 1, the Lagrange dynamics equation is used to establish the leg dynamics model of the legged robot as follows:

[0037]

[0038] wherein q∈Rn×1、 and are joint position, velocity and acceleration, respectively; M(q)∈Rn×n、 ​and G(q)∈Rn are the inertia matrix, Coriolis force vector and gravity vector respectively; f represents the friction force vector; J(q)∈Rn×n is the Jacobian matrix calculated by derivation of the robot foot position relative to the joint angle vector q; Fc∈Rn is the ground contact force vector, that is, the force vector generated by the contact between the foot and the ground environment; τ is the joint torque.

[0039] Step 2: Select joint positions, velocities, and joint acceleration driving terms as state variables to establish a state space model of the leg of the legged robot;

[0040] The joint acceleration driving term x3 is specifically: the direct force of the contact force vector on the robot acceleration after the transformation of the Jacobian matrix and the inertia matrix.

[0041] This embodiment maps the force of the joint coordinate system to the foot end contact point through the Jacobian matrix J(q), which can accurately express the force transmission relationship.

[0042] Specifically, select the state variable x1=q, x3=M(x1)-1J(q)TFc.

[0043] The state space equation of the constructed single-leg system can be expressed as:

[0044]

[0045] in,

[0046] F(x1,x2)=-M(x1)-1(C(x1,x2)+g(x1)+f) (3)

[0047] In step 3, based on the state space model of the legged robot in step 2, a high-order finite-time observer is designed. The specific form is as follows:

[0048]

[0049] In formula (4), is the estimated ground contact force. z1, z2, z3, ..., zn+2 and They are x1, x2, x3... and the estimated value of Fc. L1,L2,…,Ln+2∈Rn×n is the designed gain parameter matrix, which is used to adjust the convergence speed and accuracy of the observer. i The numerator and denominator of are both odd numbers, mi=1+(i-1)σ; parameters σ∈(-1 / (n+2),0), i=1,2,…,n+3.

[0050] In the recursive formula of the high-order finite-time observer constructed in this embodiment, the estimated value refers to the estimated value of the state variable generated by the observer, i.e., the variables z1, z2, z3, …, zn+2; the observer estimates the state variables x1, x2, x3… wherein the estimated value zi of the current layer is calculated by using the estimated value zi+1 of the previous layer, the error term (x1-z1), and the weight matrix;

[0051] For example:

[0052] In the formula, is the estimated current state x1, which is corrected by using the estimated value z2 of the previous layer and the error term (x1-z1).

[0053]

[0054] Here, is the estimated state x2, which depends on the estimated value z3 of the previous layer;

[0055] The observer depends on such step-by-step iteration calculation, and the estimated value zi of each layer will gradually converge to the actual value xi, so as to finally realize the estimation of the target state variable; the error term (x1-z1) in the observer is used to gradually correct the estimated value of each layer, so that the final estimated value zi converges to the real state xi.

[0056] Further, the estimation error of the high-order finite-time observer can be reduced by increasing the gain parameter Li, i=1, 2, …, n+2; or / and, the estimation error can be reduced by adjusting the fractional power mi of the sign function;

[0057] In the high-order finite-time observer, the error dynamic equation contains the power form sgnm of the sign function n+3 , and the response characteristics of the observer can be adjusted by the fractional power mi;

[0058] mi determines the weight of error correction of different orders. By designing the decreasing rule of mi, the low-order error (for example, e1) has a higher priority in the error correction process, so as to improve the overall stability and convergence speed of the observer;

[0059] By gradually decreasing the m i exponent, the error dynamic system can stably converge in a finite time, which meets the design goal of the high-order finite-time observer.

[0060] The observer proposed in the embodiment can not only reduce the estimation error of the high-order finite-time observer by increasing the gain Li,i=1,2,…,n+2, but also can adjust the fractional power mi to reduce the estimation error, so that the performance of the proposed observer is better than that of the ESO. Moreover, the high-order finite-time observer is more universal. When the powers of all fractional exponential power functions m2,m3,m4, etc. are equal to 1, and z4=z5=…=z n+2 =0, the high-order finite-time observer will become the ESO.

[0061] Further, the stability of the high-order finite-time observer is verified by the Lyapunov stability theory; it is proved that the high-order finite-time observer designed in step 3 can realize the finite-time stability of the ground contact force observation, and the observation error can converge to a certain boundary in a finite time.

[0062] Specifically, the stability process of the established high-order finite-time observer is proved by the Lyapunov stability theory, including the following steps:

[0063] Step 41, define the error variable as the difference between the actual value x i of the state variable and the estimated value z i , convert the formula of the high-order finite-time observer into a recursive form as the error recursive formula;

[0064] Let Solving (2) and (4) together can obtain:

[0065]

[0066] In this step, the error variable is expressed in a recursive form by the dynamic model and the observer formula, which is convenient for analyzing the evolution law of the error.

[0067] Step 42, standardize and scale the error variable through the parameter matrix in the high-order finite-time observer;

[0068] Redefine to obtain:

[0069]

[0070] This step standardizes the error variable by the inverse of the parameter matrix, which simplifies the subsequent analysis and derivation. The redefined normalized error variable is to simplify the analysis of the error dynamic system, so that the subsequent stability proof is more intuitive and easy to handle.

[0071] Step 43, take the product of the parameter matrix in the high-order finite-time observer and the inverse of the parameter matrix as the proportional factor K, to simplify the error recursive formula;

[0072] Let K1=L1, We can get:

[0073]

[0074] The simplification of the above error recursive formula reduces the complexity of the formula, making it easier to analyze and calculate, especially in subsequent stability proof (such as the construction of Lyapunov function) more easily handled;

[0075] Step 44, assuming the ground contact force F c (t) exists, the state variable x3 is continuous and differentiable, and its high-order derivative is bounded, that is As a constraint condition.

[0076] This step sets the constraint condition for verification: assuming that the contact force has continuous and differentiable high-order derivatives, which provides a mathematical basis for the subsequent observer design; By limiting the boundedness of the derivative, it is ensured that the solution of the error variable and the dynamic equation is feasible in finite time. This assumption is a prerequisite for proving the stability of the system through Lyapunov function in the future.

[0077] Step 45, define the high-order error coupling term H representing the coupling relationship between error variables i,j (e j+1 ,e j ), based on the high-order coupling term to express the error variables in recursive relationship;

[0078] Let Then Then formula (7) can be rewritten as:

[0079]

[0080] In this step, the high-order error coupling term is used to describe the dynamic relationship between error variables, which further optimizes the expression of the recursive formula;

[0081] Step 46, according to the defined error variables and coupling relationship, construct Lyapunov function, and judge the stability of the observer through simplified calculation;

[0082] The constructed Lyapunov function formula is:

[0083]

[0084] V(e) as a kind of "energy" index, is used to measure the performance of error variables in dynamic system, when the error tends to zero, the energy function V(e) also decreases to zero;

[0085] By derivation, V(e) is transformed into the recursive relationship of error variables and its derivative, the derivative of V(e) can be written as:

[0086]

[0087] Split each item, analyze the contribution of error variables separately, where:

[0088]

[0089] The error variable is assumed to have a nonlinear characteristic, and the error term is constrained by inequality, considering that the following inequality is true:

[0090]

[0091] Where b>0. Then:

[0092]

[0093] Given the lemma: When ζ,η,h>0, the following inequality is true:

[0094]

[0095] Refer to existing mathematical inequalities to constrain the growth characteristics of error variables and further refine the dynamic characteristics.

[0096] According to the above lemma (formula 14), we have:

[0097]

[0098]

[0099] Solving formula (15) and formula (16) together, combining the constraint condition to simplify the formula, and scaling formula (13) to:

[0100]

[0101] Where,

[0102] Formula (11) can be converted to:

[0103]

[0104] Where,

[0105] According to the simplified results of formula (15) to formula (18), the Lyapunov derivative is expressed as a linear combination of a group of error variables, and the simplified Lyapunov derivative is obtained. Formula (10) can be converted to:

[0106]

[0107] wherein, the designed p = y - y1 > 0, and Θ > 0 is a constant.

[0108] From equation (19), the error of the estimated ground contact force will converge to a bounded region Ω in finite time

[0109]

[0110] That is, the designed high-order finite-time observer (equation 4) is finite-time stable.

[0111] Further, the estimated order is selected according to the ground contact force model. The ground contact force model includes constant type, slope type, parabolic type, and higher order type, etc.

[0112] wherein, by selecting n = 1, 2, 3, the estimation of constant, slope and parabolic type ground contact forces is realized respectively.

[0113] Similarly, a higher order n can be selected to ensure the estimation accuracy of a more general order ground contact force, but it will increase the amount of calculation. In actual engineering, the low order case can be considered first, and if the accuracy does not meet the requirements, the estimation accuracy can be improved by appropriately increasing the order.

[0114] The method proposed in this embodiment can estimate the foot end ground contact force using the torque sensor, displacement sensor, etc. of the leg of the legged robot without installing a foot end contact force sensor, which can avoid increasing the complexity of the leg structure and the weight of the leg of the legged robot. Compared with the traditional estimation method based on the extended state observer (ESO), the high-order finite-time observer proposed in this embodiment can more accurately estimate the time-varying high-order ground contact force and realize finite-time stability; while the ESO can only estimate the slowly changing force and has insufficient estimation accuracy for the time-varying ground contact force, and only realizes asymptotic stability.

[0115] In order to illustrate the effect of the method described above in this embodiment, a simulation experiment was carried out, and the ground contact force estimation method proposed in this embodiment was verified for a three-degree-of-freedom single leg system of a legged robot. As shown in Figure 2 , the leg simulation model of the legged robot constructed by a software is composed of a body 1, a hip 2, a thigh 3, a shank 4, a ground 5 and a support 6, and the simulation model parameters are shown in Table 1.

[0116] Table 1

[0117] Parameter Value and unit Hip length 0.1m Thigh length 0.4m Calf length 0.4m Body mass 8 kg Hip mass 2 kg Thigh mass 5 kg Calf mass 2 kg

[0118] Since the robot leg in the simulation model is three degrees of freedom, the dimension of the system parameter matrix in the formula (1) is 3x3, and the dimension of the vector is 3x1, that is, the establishment of the leg dynamics model of the legged robot is as follows:

[0119]

[0120] Where q∈R 3×1 、 and are joint position, velocity and acceleration, respectively; M(q)∈R 3×3 、 and G(q)∈R 3 are the inertia matrix, the Coriolis force vector and the gravity vector, respectively; f represents the friction force vector; J(q)∈R 3×3 is the Jacobian matrix calculated by the derivative of the robot foot position with respect to the joint angle vector q; F c ∈R 3 is the contact ground force vector, that is, the force vector generated by the contact between the foot and the ground environment; τ∈R 3 is the joint torque.

[0121] In order to verify the effectiveness and superiority of the proposed observer for the estimation of the touch ground force of the legged robot, the following two comparison schemes are designed, and the simulation comparison verification is carried out in the following three scenes.

[0122] Scheme 1: (denoted as C1) This scheme adopts the touch ground force estimation scheme based on the high-order finite time observer proposed in this embodiment, which adopts a 4th order high-order finite time observer, that is, n=2, and other parameters of the observer are set as L1=22, L2=22, L3=30, L4=440.

[0123] Scheme 2: (denoted as C2) This scheme adopts the touch ground force estimation scheme based on the extended state observer (ESO), and the parameters of the ESO are set as k1=8, k2=50, k3=300.

[0124] Scene 1: Given the step type touch ground force estimation. The touch ground force of three degrees of freedom is set to [20,-30,-30] T N, which is applied to the foot end of the single leg of the robot from 5s to 20s.

[0125] Scene 2: Given the slope type touch ground force estimation. The touch ground force signals of three degrees of freedom are all set to 0.1Hz triangular wave signals, the amplitude of the touch ground force signal in the x direction is 70N, the amplitude of the touch ground force signal in the y direction is 40N, and the amplitude of the touch ground force signal in the z direction is-50N.

[0126] Scenario 3: Free fall ground contact force estimation. Let the legged robot free fall from a height of 0.4m, after a period of falling, the robot foot end collides with the ground, and the ground contact force estimation effect is compared.

[0127] The simulation results of the legged robot ground contact force estimation are shown in Figures 3 to 5 The vertical coordinate in the figure is the ground contact force, and the horizontal coordinate is the time. Figure 3 The estimation results of the given step type foot end ground contact force in scenario 1 are given, Figure 4 The estimation results of the given slope type foot end ground contact force in scenario 2 are given, Figure 3 And Figure 4 In the figure, the vertical coordinate of each small graph from top to bottom is the component force F cx (N) of the foot end ground contact force in the x direction, the component force F cy (N) in the y direction, and the component force F cz (N) in the z direction. Figure 5 The estimation results of the foot end ground contact force when the robot free falls in scenario 3 are given. It can be found that in the three scenarios, the ground contact force estimation method of the high-order finite-time observer proposed in this embodiment has high coincidence with the given force, and the estimation time is before scheme 2, so the estimation method of this embodiment is better than the traditional ground contact force estimation method based on ESO in the three scenarios.

[0128] The above simulation results verify the effectiveness of the ground contact force estimation method of the legged robot based on the high-order finite-time observer proposed in this embodiment.

[0129] Embodiment 2

[0130] Based on embodiment 1, the robot ground contact force estimation system based on the high-order finite-time observer provided in this embodiment includes:

[0131] A dynamics model construction module configured to establish a legged robot leg dynamics model;

[0132] A state space model construction module configured to select joint position, velocity and joint acceleration driving term as state variable, and establish a legged robot leg state space model;

[0133] An estimation module configured to establish a high-order finite-time observer by recursive method according to the established state space model, using step-by-step iteration of the estimated value of the state variable, combining the power form of the nonlinear sign function and the set gain parameter matrix, and obtaining the estimation of the foot end ground contact force of the legged robot.

[0134] It should be noted that each module in this embodiment corresponds to each step in embodiment 1, and the specific implementation process is the same, which will not be repeated here.

[0135] Embodiment 3

[0136] Based on embodiment 1, this embodiment provides an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, when the computer instructions are run by the processor, the steps in the robot ground contact force estimation method based on high-order finite time observer described in embodiment 1 are completed.

[0137] Embodiment 4

[0138] Based on embodiment 1, this embodiment provides a computer readable storage medium for storing computer instructions, when the computer instructions are executed by the processor, the steps in the robot ground contact force estimation method based on high-order finite time observer described in embodiment 1 are completed.

[0139] The above only describes the preferred embodiments of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art can make various modifications and changes to the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

[0140] Although the specific embodiments of the present disclosure are described above in combination with the accompanying drawings, the present disclosure is not limited to the scope of the present disclosure. Those skilled in the art should understand that various modifications or changes made on the basis of the technical solutions of the present disclosure without creative labor are still within the protection scope of the present disclosure.

Claims

1. A robot ground contact force estimation method based on a high-order finite-time observer, characterized by, The method comprises the following steps: establishing a leg dynamics model of a legged robot; selecting joint position, velocity and joint acceleration driving term as state variables to establish a state space model of the leg of the legged robot; according to the established state space model, using step-by-step iteration of the estimated value of the state variable, combining the power form of the nonlinear symbolic function and the set gain parameter matrix, and establishing a high-order finite-time observer in a recursive manner to obtain the estimation of the foot end ground contact force of the legged robot; The high-order finite-time observer is specifically as follows: wherein is the estimated ground force; , , , , and are the estimated values of , , … and respectively; is a set gain parameter matrix, is a fractional power of a sign function; wherein , , are the inertia matrix, the Coriolis force vector and the gravitational force vector, respectively, in the states and , denotes the friction force vector.

2. The robot ground contact force estimation method based on a high-order finite time observer of claim 1, wherein: The leg dynamics model of the legged robot is established by using Lagrange dynamics equation.

3. The robot ground contact force estimation method based on a high-order finite time observer of claim 1, wherein, The joint acceleration driving term is: the direct force of the contact ground force vector on the robot acceleration after the transformation of the Jacobian matrix and the inertia matrix.

4. The robot ground contact force estimation method based on a high-order finite time observer of claim 1, wherein: In the recursive formula of the high-order finite-time observer, the estimation value of the current layer is calculated by using the estimation value of the last layer , the error term and the weight matrix.

5. The high-order finite-time observer based robot ground contact force estimation method of claim 1, wherein: By increasing the gain parameter to reduce the estimation error of the high-order finite-time observer; or / and, by adjusting the fractional power of the sign function to reduce the estimation error.

6. The high-order finite-time observer based robot ground contact force estimation method of claim 1, wherein, The stability of the high-order finite-time observer is verified by Lyapunov stability theory, and the process is as follows: The error variable is defined as the difference between the actual value and the estimated value of the state variable, and the formula of the high-order finite-time observer is converted into a recursive form as the error recursive formula; The error variable is standardized and scaled by the parameter matrix in the high-order finite-time observer; The product of the parameter matrix in the high-order finite-time observer and the inverse of the parameter matrix is taken as the proportional factor K to simplify the error recursive formula; Assuming ground contact force of the existence of the second derivative, the state variable is continuously differentiable and its higher-order derivatives are bounded as a constraint; The high-order error coupling term representing the coupling relationship between the error variables is defined, and the error variables are expressed in a recursive relationship based on the high-order coupling term; According to the defined error variable and coupling relationship, a Lyapunov function is constructed to judge the stability of the observer by simplifying the calculation.

7. A robot ground contact force estimation system based on a higher order finite time observer, characterized by, It comprises: The dynamics model construction module is configured to establish a leg dynamics model of a legged robot; The state space model construction module is configured to select joint position, velocity and joint acceleration driving term as state variables to establish a state space model of the leg of the legged robot; The estimation module is configured to, according to the established state space model, use step-by-step iteration of the estimated value of the state variable, combine the power form of the nonlinear symbolic function and the set gain parameter matrix, and establish a high-order finite-time observer in a recursive manner to obtain the estimation of the foot end ground contact force of the legged robot; The high-order finite-time observer is specifically as follows: wherein is the estimated ground force; , , , , and are the estimated values of , , … and ; is a set gain parameter matrix, is the fractional power of the sign function; wherein , , are the inertia matrix, the Coriolis force vector and the gravitational force vector, respectively, in the states and , denotes the friction force vector.

8. An electronic device, comprising: It comprises a memory and a processor, and computer instructions stored in the memory and running on the processor, when the computer instructions are run by the processor, the steps in the robot ground contact force estimation method based on the high-order finite-time observer in any one of claims 1-6 are completed.

9. A computer-readable storage medium, characterized in that, A computer instruction storage device, when the computer instruction is executed by the processor, the steps in the robot ground contact force estimation method based on the high-order finite-time observer in any one of claims 1-6 are completed.

Images