Method and system for estimating and compensating mass center of variable-configuration quadruped robot with arm

Abstract

The invention provides a centroid estimation and compensation method and system for a variable-configuration quadruped robot with an arm, and relates to the technical field of quadruped robot motion planning, and the method comprises the steps: obtaining the state feedback information of the variable-configuration quadruped robot in the motion process; on the basis of the state feedback information, foot end contact force of the supporting legs is calculated through a whole-body dynamic model of the variable-configuration quadruped robot and a disturbance force observer based on generalized momentum; according to the foot end contact force of the supporting leg, adopting a centroid estimation algorithm of an online recursive strategy to estimate a centroid deviation value after unknown disturbance; according to the estimated centroid deviation value, a motion trail planner and a centroid dynamical model are adjusted, centroid deviation compensation is conducted through a whole body controller of the null space based on the adjusted result, the optimal joint torque is obtained, and real-time servo control is conducted; according to the invention, stable control of the quadruped robot under disturbance of internal force or external force such as double-arm movement, leg configuration change and unknown load is realized.

Description

Technical Field

[0001] The present disclosure relates to the technical field of quadruped robot motion planning, and in particular to a method and system for estimating and compensating the center of mass of a variable configuration quadruped robot with arms. Background Art

[0002] Among legged robots, quadruped robots have attracted the attention of many scientific researchers and research institutions due to their advantages such as simple structure, strong stability and strong adaptability to rugged terrain. In particular, variable-configuration quadruped robots with two arms are favored because they have multiple configurations such as mammalian configuration, reptilian configuration and two-arm mobile operation configuration, and they show strong terrain adaptability, the ability to cross irregular obstacles and the ability to move with two arms.

[0003] The model-based control method relies on an accurate dynamic model. However, the dynamic inertia parameters of the robot are affected by internal force factors such as the installation position of the manipulator arm, the movement position of the two arms, the configuration of the robot legs, or the error in reading the dynamic parameters of the CAD software. At the same time, the robot is also affected by external factors such as the sudden application of unknown external loads or the clamping of heavy objects by the two arms. These internal or external forces will cause the robot's center of mass to deviate from the robot's geometric center, thereby increasing the controller's attitude angle tracking error, causing the robot's torso to tilt or even become unstable and fall. Therefore, it is necessary to accurately estimate the center of mass offset to correct the dynamic model parameters.

[0004] However, most of the existing center of mass estimation and compensation methods require the robot to remain stationary during the center of mass estimation process or can only identify the center of mass during static gait motion. There are very few methods that can achieve center of mass estimation during dynamic gait motion. At the same time, for quadruped robots with robotic arms, when performing unknown target gripping, most of the current methods assume that the mass of the gripping target is known to establish a dynamic model. However, in actual operation, the mass of the gripping target cannot be obtained in real time, and the center of mass offset caused by unknown loads is ignored. The above methods will reduce the robot's operating accuracy or motion efficiency. For quadruped robots with arms, it is impossible to estimate the center of mass offset in real time during dynamic motion, which ultimately affects the stability of the robot system. Summary of the invention

[0005] In order to solve the above problems, the present disclosure proposes a method and system for estimating and compensating the center of mass of a variable-configuration quadruped robot with arms, so as to achieve smooth control of the dynamic motion of the quadruped robot under internal or external force disturbances such as movement of both arms, change of leg configuration and unknown load.

[0006] According to some embodiments, the present disclosure adopts the following technical solutions: A method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms, comprising: Obtain state feedback information of the variable configuration quadruped robot during movement; Based on the state feedback information, the foot contact force of the supporting leg is calculated through the whole-body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum. According to the contact force of the foot end of the supporting leg, the center of mass estimation algorithm with online recursive strategy is used to estimate the center of mass offset value after unknown disturbance. According to the estimated center of mass offset value, the motion trajectory planner and the center of mass dynamics model are adjusted. Based on the adjusted results, the center of mass deviation is compensated by the whole body controller in the zero space to obtain the optimal joint torque and perform real-time servo control. Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by superimposing and sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane.

[0007] According to some embodiments, the present disclosure adopts the following technical solutions: A mass center estimation and compensation system for a variable configuration quadruped robot with arms, comprising: The state acquisition module is configured to: obtain state feedback information of the variable configuration quadruped robot during movement; The contact force calculation module is configured to: calculate the foot end contact force of the supporting leg based on the state feedback information through the whole body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum; The offset estimation module is configured to: estimate the center of mass offset value after being subjected to unknown disturbances by using a center of mass estimation algorithm of an online recursive strategy according to the foot end contact force of the supporting leg; The offset compensation module is configured to: adjust the motion trajectory planner and the center of mass dynamics model according to the estimated center of mass offset value, and based on the adjusted result, compensate the center of mass deviation through the whole body controller of the null space to obtain the optimal joint torque and perform real-time servo control; Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by superimposing and sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product comprises a computer program, wherein when the computer program is executed by a processor, the method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms is implemented.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, a method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms is implemented.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms.

[0011] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention proposes a disturbance force observer based on generalized momentum, calculates the generalized momentum of the robot according to the state estimation information such as the joint torque and joint velocity of the robot during the movement, and uses the robot's whole-body dynamics model and the disturbance force observer to calculate the accurate foot-end contact force of the robot, thereby achieving accurate feedback of the robot's foot-end interaction force and contact state.

[0012] (2) The present invention proposes a center of mass estimation algorithm based on an online recursive strategy, combines the estimated accurate foot-end contact force with the difference in contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, and adopts a recursive iterative strategy to estimate the offset of the center of mass in the horizontal plane, thereby realizing accurate center of mass deviation estimation of a variable configuration quadruped robot with a robotic arm under various disturbance factors.

[0013] (3) The present invention proposes a center of mass compensator based on whole-body control. According to the estimated center of mass offset of the robot system, combined with a whole-body controller and a motion trajectory planner based on task priority, the center of mass offset of the robot caused by internal force factors such as modeling errors, arm movement or leg configuration, and external force factors such as unknown loads is compensated, thereby achieving smooth motion control and mobile operation of a quadruped robot with a robotic arm. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0015] Figure 1 is a structural diagram of a variable configuration quadruped dual-arm robot of Example 1; Figure 2 is a motion control block diagram of a variable configuration quadruped dual-arm robot of Example 1; Figure 3 It is a schematic diagram for analyzing the overturning torque and the plantar force of the foot end of the variable configuration quadruped dual-arm robot of Example 1. DETAILED DESCRIPTION

[0016] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0017] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0018] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0019] Example 1 In one embodiment of the present disclosure, a method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms is provided to realize the stable dynamic motion control of the quadruped robot under internal or external force disturbances such as double-arm movement, leg configuration change and unknown load. Figure 1 Taking the quadruped robot with a variable dual-arm configuration as an example, the specific implementation scheme is described in detail.

[0020] Figure 1 In the figure, the directions of the world coordinate system {W}, the trunk coordinate system {B}, and the hip joint coordinate system {H} are consistent with those in the figure. i represents the position of the foot of the ith supporting leg relative to the trunk coordinate system, r com Represents the offset value of the robot's center of mass relative to the torso coordinate system, l b 、w b 、h b Represent the length, width and height of the robot torso respectively.

[0021] In order to solve the Figure 1The quadruped robot shown in the figure has a mass center offset CoM caused by the installation position, the movement of the robotic arm or the unknown load during the movement. It is necessary to design an estimation and compensation scheme for the mass center offset to achieve the robot posture stability and accurate end-to-end following during the movement. To this end, this embodiment proposes a mass center offset estimation and compensation method that can be applied to a quadruped robot with a robotic arm. The specific process is as follows: Figure 2 As shown in the figure, the disturbance force observer based on generalized momentum can realize accurate estimation of the contact force at the foot end during movement; the center of mass estimation algorithm of the online recursive strategy can realize accurate estimation of the center of mass offset of the variable configuration quadruped robot with two arms; the center of mass compensator based on whole-body control can realize compensation and smooth control of the center of mass offset of the variable configuration quadruped robot with two arms.

[0022] Step S1: Obtaining state feedback information of the variable configuration quadruped robot during movement.

[0023] The state feedback information here includes the trunk position and speed information of the variable configuration quadruped robot, which is obtained as follows: First, the joint torque is collected through the robot's proprioceptors and joint angle .

[0024] Then, based on the joint angles, the trunk position p is calculated using forward kinematics com , speed information v .

[0025] Step S2: Based on the state feedback information, the foot end contact force of the supporting leg is calculated through the whole body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum, specifically: (1) Based on the whole-body dynamics model and state feedback information, the disturbance torque exerted by the ground during the robot’s motion is calculated.

[0026] Specifically, based on the robot state feedback information, a full-body dynamics model of a variable-configuration quadruped robot with a robotic arm is established:

[0027] in, is the inertia matrix, is the Coriolis force and the centrifugal force, is the gravity term, represents the robot joint torque, is the corresponding leg joint motor torque, is the joint motor torque corresponding to the robotic arm, is the torque of the trunk. Represents the robot's torso pose, leg joint angles, and arm joint angles. and are the corresponding angular velocity and angular acceleration respectively. represents the Jacobian matrix of the corresponding torso, leg and arm. f represents the contact force between the foot and the ground.

[0028] According to the robot's whole-body dynamics model and state feedback information, the disturbance torque of the ground during the robot's movement is calculated. for:

[0029] in, is the selection matrix for the floating base, and are the angular velocity and angular acceleration corresponding to the trunk posture and the leg-arm joints, respectively.

[0030] (2) Using the disturbance force observer, the disturbance torque is discretized.

[0031] Disturbance torque Discretize it and express it as follows:

[0032] Where z is a z-domain variable and γ is the cutoff frequency of the filter. The associated filter, and .

[0033] Data sampling time is 500 Hz, assuming that the Coriolis and centrifugal forces The decomposition satisfies , so the above formula can be decomposed into:

[0034] in, Represents the generalized momentum, which is substituted into the disturbance force observer to obtain the discretized disturbance torque:

[0035] in, represents the gain related to the cutoff frequency, is the sampling time of the data.

[0036] (3) Based on the discretized disturbance torque, the contact force at the foot end of the supporting leg is calculated through the relationship between the Jacobian matrix of the robot leg and the joint torque.

[0037] According to the calculated disturbance torque , through the Jacobian matrix of the robot leg and joint torque The relationship between the foot end and the contact force is obtained. for:

[0038] Step S3: Based on the foot contact force of the supporting leg, an online recursive strategy is used to estimate the center of mass offset after being subjected to an unknown disturbance.

[0039] Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane. The specific steps are: (1) Assume that during the diagonal gait movement, the robot's trunk center of mass (CoM) is located at the robot's geometric center and the robot system maintains balance, such as Figure 3 As shown, the foot-end support leg contact force and the trunk y-direction torque are established The relationship equation is:

[0040] Among them, and It represents the foot positions of the front and rear supporting legs relative to the trunk coordinate system {B} during trotting. and Represents the contact force applied to the front and rear supporting legs during trotting.

[0041] (2) Assume that the robot's center of mass shift causes gravity mg to generate a tipping torque along the y-axis: ,in is the distance in the x direction from the center of the torso of the robot's mass center in the torso coordinate system {B}.

[0042] This additional torque will disrupt the robot's equilibrium, so additional contact force needs to be applied to the robot. and , so that the robot returns to a balanced state, which can be expressed as:

[0043] Therefore, combined with the above and , the front supporting leg contact force can be obtained Contact force with rear support leg The relationship between them is:

[0044] in, , They represent the foot positions of the front and rear supporting legs relative to the hip joint coordinate system {H}, is the length of the torso.

[0045] Based on the stability of the diagonal legs and the stability of the trunk standing height, we can get:

[0046] When the velocity changes in the x-direction are small, the longitudinal contact force remains stable. and The change in can be ignored, and we can get:

[0047] Therefore, combined with the above formula, by analyzing the contact force of the front outrigger Contact force with rear outrigger The difference between direction.

[0048] (3) Similarly, according to the overturning torque of the robot in the x direction , and the contact force of the left supporting leg is obtained Contact force with right supporting leg The relationship between them is:

[0049] Similar to the above process, based on the synchronization of the diagonal legs and ignoring and The change of can be obtained as follows:

[0050] According to the above formula, by analyzing the contact force of the left supporting leg Contact force with right supporting leg The difference between direction.

[0051] (4) CoM direction deviation in x and y, using piecewise functions It is expressed as follows:

[0052]

[0053] In the formula, , is the contact force difference threshold of the front and rear supporting legs, , ,same, , represents the contact force threshold of the left and right supporting legs, where , ;if , then the CoM is in front of the center of the trunk if , then the CoM is to the left of the center of the trunk; the front supporting leg , rear support leg , Left supporting leg , right supporting leg The contact forces must all exceed the minimum contact force threshold Only then can you confirm that the foot has touched the ground.

[0054] In order to improve the accuracy of CoM estimation, a recursive iterative strategy is used to estimate the offset, and a sliding window is introduced to reduce the lag. Superposition as variable N k , the calculation formula is:

[0055] Among them, k x , k y Indicates the sampling time.

[0056] The deviation is identified using a recursive strategy and multiple corrections are used to determine the center of mass offset. Therefore, the position of the CoM on the horizontal plane is estimated to be:

[0057] in, , is the scale factor in the x and y directions, , are the differential coefficients in the x and y directions respectively; the displacement of the center of mass in the z direction is relatively small and basically does not produce overturning torque, so it is assumed that the displacement of the robot's center of mass in the z direction is 0, and the center of mass displacement caused by internal and external force disturbances in the horizontal plane is calculated to be .

[0058] Step S4: According to the estimated center of mass offset value, the motion trajectory planner and the center of mass dynamics model are adjusted. Based on the adjusted results, the center of mass deviation is compensated by the whole body controller in the null space to obtain the optimal joint torque, and real-time servo control is performed to achieve the stability and smooth operation of the robot system. The specific process is as follows: (1) The center of mass offset value is compensated into the motion trajectory planner, and the center of mass deviation is compensated for the foot trajectory in the support phase and the swing phase.

[0059] Specifically, according to the robot's expected speed input by the user and the estimated centroid offset Plan the foot trajectory of the robot in the support phase and the swing phase to achieve compensation for the center of mass deviation of the foot trajectory. Adjust the planned foot trajectory to offset the center of mass. Introduced into the motion planner for compensation, expressed as:

[0060] in,

[0061]

[0062] in, represents the expected foot trajectory of the jth leg foot. is the position of the torso center in the world coordinate system {W} at the kth time step, represents the rotation matrix of the robot torso relative to the world coordinate system {W}, is the position of the hip joint of the jth swing leg relative to the trunk coordinate system {B}. Therefore, is the hip joint position of the jth leg relative to the world coordinate system {W}. and represents the feedback velocity and the desired velocity of the robot trunk, represents the desired angular velocity of the trunk, represents the duration of the support phase, Indicates the standing height of the robot's torso.

[0063] (2) Based on the state feedback of the motion trajectory planner, the robot's dynamic model is compensated, and the optimal contact force at the foot end during the motion process is solved according to the robot's state space equation.

[0064] Specifically, the centroid offset value Compensate to the trunk center of mass dynamics model to achieve offset compensation of the dynamics model and ensure the stability of the robot during movement. It can be expressed as:

[0065] Among them, m is the body mass of the robot, and c is the number of limbs in contact with the ground. , , They represent the robot’s torso position, limb contact force, and gravitational acceleration, respectively. is the position of the ith contact point, is the offset of the robot's CoM, both relative to the origin of the torso coordinate system. represents the inertia tensor of the robot in the world coordinate system, and ω represents the angular velocity of the torso in the world coordinate system.

[0066] Construct the state space equation of the robot system, solve the above optimization problem as a quadratic programming, and solve the optimal contact force of the foot during the movement:

[0067] in, is the calculated feedforward optimal foot-end contact force, is the compensating contact force calculated based on the torso posture deviation. Q and W are positive definite weight matrices, Assign weights to secondary objectives (such as regularization to make the solution bounded), C is the friction cone inequality constraint matrix approximated as a pyramid, , are the upper and lower bounds of the friction cone constraint, respectively.

[0068] (3) Using a whole-body controller based on zero space, the expected acceleration and angular velocity that satisfy the robot's trunk center of mass task, trunk posture task, swing leg trajectory task, and multi-limb operation task are solved, and combined with the optimal contact force at the foot end, a convex optimization function is constructed to calculate the optimal interaction force of the limbs, which is finally converted into joint torque to ensure the accurate execution of multiple tasks and realize the center of mass offset compensation and stable whole-body control of the robot system after various disturbances.

[0069] Specifically, the expected acceleration that satisfies multiple tasks is solved by performing task iteration through null space projection. , expected angular velocity , and the optimal feedforward contact force calculated from the state equation constructed based on the center-of-mass dynamics Combined with the above, a convex optimization function that satisfies multiple constraints such as the whole-body dynamics model and friction cone constraints is constructed to solve the minimum relaxation variables of the floating base acceleration and contact force under multiple constraints. and , as shown in the following formula:

[0070] in, and are the relaxation variables for the optimal feedforward force of the floating base acceleration and center of mass dynamics, respectively. Feedforward forces calculated for an optimization problem formulated based on the robot state-space equations. and are the weight matrices respectively. is the selection matrix for the floating base, is the linearized friction cone constraint matrix. Represents the Jacobian matrices corresponding to the torso, legs, and arms.

[0071] This convex optimization function uses the QP optimizer QuadProg++ to calculate the optimal interaction force of the limbs and converts it into joint torque for servo control.

[0072] Example 2 In one embodiment of the present disclosure, a mass center estimation and compensation system for a quadruped robot with arms is provided, comprising: The state acquisition module is configured to: obtain state feedback information of the variable configuration quadruped robot during movement; The contact force calculation module is configured to: calculate the foot end contact force of the supporting leg based on the state feedback information through the whole body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum; The offset estimation module is configured to: estimate the center of mass offset value after being subjected to unknown disturbances by using a center of mass estimation algorithm of an online recursive strategy according to the foot end contact force of the supporting leg; The offset compensation module is configured to: adjust the motion trajectory planner and the center of mass dynamics model according to the estimated center of mass offset value, and based on the adjusted result, compensate the center of mass deviation through the whole body controller of the null space to obtain the optimal joint torque and perform real-time servo control; Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by superimposing and sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane.

[0073] Example 3 In one embodiment of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method for estimating and compensating the center of mass of a deformable quadruped robot with arms is implemented.

[0074] Example 4 In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the center of mass estimation and compensation method of a deformable quadruped robot with arms is implemented.

[0075] Example 5 In one embodiment of the present disclosure, an electronic device is provided, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the method for estimating and compensating the center of mass of a deformable quadruped robot with arms.

[0076] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0078] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A method for estimating and compensating the center of mass of a variable configuration quadruped robot with arms, characterized in that: include: Obtain state feedback information of the variable configuration quadruped robot during movement; Based on the state feedback information, the foot contact force of the supporting leg is calculated through the whole-body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum. According to the contact force of the foot end of the supporting leg, the center of mass estimation algorithm with online recursive strategy is used to estimate the center of mass offset value after unknown disturbance. According to the estimated center of mass offset value, the motion trajectory planner and the center of mass dynamics model are adjusted. Based on the adjusted results, the center of mass deviation is compensated by the whole body controller in the zero space to obtain the optimal joint torque and perform real-time servo control. Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by superimposing and sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane.

2. The method for estimating and compensating the center of mass of a quadruped robot with arms according to claim 1, characterized in that: The state feedback information includes the trunk position and speed information of the variable configuration quadruped robot.

3. The method for estimating and compensating the center of mass of a quadruped robot with arms according to claim 1, characterized in that: The calculation of the foot end contact force of the supporting leg is specifically as follows: According to the whole-body dynamics model and state feedback information, the disturbance torque of the ground during the robot's movement is calculated; The disturbance torque is discretized using the disturbance force observer; Based on the discretized disturbance torque, the contact force at the foot end of the supporting leg is calculated through the relationship between the Jacobian matrix of the robot leg and the joint torque.

4. The method for estimating and compensating the center of mass of a quadruped robot with arms according to claim 1, characterized in that: The difference in the foot contact force between the front and rear and left and right supporting legs is used to estimate the direction of the center of mass deviation in the horizontal plane, specifically: Establish the relationship between the foot contact force of the supporting leg and the trunk torque, and calculate the overturning torque in the x-direction and y-direction; Based on the overturning torque in the x-direction and the y-direction, the difference between the contact force of the front supporting leg and the contact force of the rear supporting leg, and the difference between the contact force of the left supporting leg and the contact force of the right supporting leg are calculated respectively; Using the difference, estimate the direction of the centroid offset in the horizontal plane.

5. The method for estimating and compensating the center of mass of a quadruped robot with arms according to claim 1, characterized in that: The offset is gradually identified by sliding windows one by one, specifically: Use piecewise functions to represent the deviation of the center of mass offset in the x and y directions respectively; The directional deviation of each segment is used as a window. By superimposing and sliding each window one by one, the centroid offset direction is recalculated in turn to determine whether the centroid is offset. If the centroid is still offset, the next window is iteratively superimposed and recalculated. Otherwise, the current window superposition value is the final centroid offset value.

6. The method for estimating and compensating the center of mass of a quadruped robot with arms according to claim 1, characterized in that: The optimal joint torque is obtained as follows: The center of mass offset value is compensated into the motion trajectory planner, and the center of mass deviation is compensated for the foot end trajectory in the support phase and the swing phase; According to the state feedback of the motion trajectory planner, the dynamic model of the robot is compensated, and the optimal contact force of the foot during the motion process is solved according to the state space equation of the robot; A whole-body controller based on null space is used to solve the expected acceleration and angular velocity that satisfy the robot's trunk center of mass task, trunk posture task, swing leg trajectory task, and multi-limb working task. It is then combined with the optimal contact force at the foot end to construct a convex optimization function, calculate the optimal interaction force of the limbs, and finally convert it into joint torque.

7. A mass center estimation and compensation system for a quadruped robot with arms, characterized in that: include: The state acquisition module is configured to: obtain state feedback information of the variable configuration quadruped robot during movement; The contact force calculation module is configured to: calculate the foot end contact force of the supporting leg based on the state feedback information through the whole body dynamics model of the variable configuration quadruped robot and the disturbance force observer based on generalized momentum; The offset estimation module is configured to: estimate the center of mass offset value after being subjected to unknown disturbances by using a center of mass estimation algorithm of an online recursive strategy according to the foot end contact force of the supporting leg; The offset compensation module is configured to: adjust the motion trajectory planner and the center of mass dynamics model according to the estimated center of mass offset value, and based on the adjusted result, compensate the center of mass deviation through the whole body controller of the null space to obtain the optimal joint torque and perform real-time servo control; Among them, the center of mass estimation algorithm of the online recursive strategy uses the difference in foot contact force between the front and rear and left and right supporting legs to estimate the direction of the center of mass offset in the horizontal plane, introduces a sliding window in the direction of the center of mass offset, and gradually identifies the offset by superimposing and sliding windows one by one, and finally obtains the estimated value of the center of mass offset in the horizontal plane.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for estimating and compensating the center of mass of a deformable quadruped robot with arms as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, a method for estimating and compensating the center of mass of a deformable quadruped robot with arms as described in any one of claims 1 to 6 is implemented.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes a method for estimating and compensating the center of mass of a deformable quadruped robot with arms as described in any one of claims 1 to 6.

Images