A posture control method and system for a tandem quadruped wheel-legged robot

Abstract

The present application discloses a posture control method and system for a serial quadruped wheel-leg robot, which relates to the field of robot control technology. The method includes: according to the pre-established wheel-leg robot synthetic center of mass kinematic analysis model, according to the expected posture information, the expected joint angle is calculated; then, on the basis of generating joint feedback torque based on the actual joint angle and the expected joint angle, according to the observed error disturbance, the joint feedforward torque is calculated; and according to the actual joint angle, the actual wheel information and the expected wheel information, the wheel control torque is calculated; finally, according to the joint control torque and the wheel control torque, the posture change of the wheel-leg robot and the wheel travel are controlled. The above-mentioned posture control method proposed in the present application effectively ensures the stability of the quadruped robot in the driving state through the coordinated action of the wheels and the leg joints, and improves the robot's anti-interference ability and control ability of the motion posture during driving.

Description

Technical Field

[0001] The present application relates to the field of robot control technology, and in particular to a posture control method and system for a serial quadruped wheel-legged robot. Background Art

[0002] The control of robot posture is a key issue in robot research, and many studies are currently being conducted in this field. Klemm et al. proposed a posture control strategy based on a linear quadratic regulator for the two-wheeled legged robot Ascento, which maintains the balance of the robot through the coordinated actuation of the wheels and leg joints. Lei Tao's team regarded the legs of the parallel six-wheeled legged robot as a suspension system. To address the problems of foot suspension and posture tilt, they established a single-leg impedance model and considered the environmental contact model as a stiffness damping model, and proposed a body stability control framework that integrates the foot force controller, posture controller, and center of gravity height controller.

[0003] Currently, there is little research on the posture adjustment of four-wheeled legged robots. The robotics team of the Swiss Federal Institute of Technology mainly uses reinforcement learning to achieve difficult movements of robots in various postures, but this method has a high learning cost and is difficult to generalize. Carlo et al. designed a model predictive controller for a quadruped robot with a similar structure to the four-wheeled legged robot. The controller uses the single rigid body dynamics that ignores the mass of the legs and feet as the reference model, and has a good real-time posture adjustment effect. However, the mass below the body of the wheeled leg robot accounts for a certain proportion. If the mass of the legs and wheels is ignored, the control effect will be affected.

[0004] In summary, the existing posture control methods of wheel-legged robots are mainly based on two-wheeled legged robots, and there is little research on four-wheeled legged robots. In addition, the corresponding posture control methods are only applicable to unique configurations and have limitations. In addition, the current posture adjustment methods have high learning costs and are difficult to generalize. Summary of the invention

[0005] The purpose of this application is to provide a posture control method and system for a serial quadruped wheeled robot, which can effectively ensure the stability of the quadruped robot in the driving state, and improve the robot's anti-interference ability and control ability of the motion posture during driving.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a posture control method for a tandem quadruped wheeled robot, comprising the following steps:

[0008] According to the current posture information of the wheel-legged robot, the posture command and the travel command obtained externally, the expected posture information and the expected wheel information of the wheel-legged robot are generated.

[0009] The pre-established wheel-leg robot synthetic center of mass kinematic analysis model is used to calculate the expected joint angles according to the expected posture information.

[0010] The actual joint angles of the wheel-legged robot are obtained, and the joint feedback torque is calculated based on the actual joint angles and the expected joint angles.

[0011] Using the pre-established synthetic center of mass dynamics analysis model of the wheel-legged robot, the joint feedforward torque and wheel-end force are calculated according to the expected posture information, the measured actual posture information and the observed error disturbance. The error disturbance is the error value determined by the extended state observer according to the actual joint angle and wheel-end force. The wheel-end force is the ground force acting on the wheel end.

[0012] The wheel control torque is calculated based on the actual joint angle, actual wheel information and expected wheel information using the pre-established wheel kinematic model.

[0013] Based on the joint feedback torque and the joint feedforward torque, the joint control torque is obtained, and according to the joint control torque and the wheel control torque, the posture change of the wheel-legged robot and the movement of the wheels are controlled.

[0014] Optionally, the posture control method for the tandem quadruped wheel-legged robot further comprises the following steps:

[0015] Based on the structural characteristics of the serial quadruped wheel-legged robot, a kinematic analysis model of the serial quadruped wheel-legged robot is established.

[0016] Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the contact plane is fitted using the least squares method to obtain the terrain slope.

[0017] Based on the terrain slope and the structural characteristics of four sets of wheel-legged robots in series, a synthetic center of mass kinematic analysis model of the wheel-legged robot is established with comprehensive leg mass and modeling error.

[0018] Optionally, based on the kinematic analysis model of the serial quadruped wheel-legged robot, the contact plane is fitted using the least squares method to obtain the terrain slope, which specifically includes the following steps:

[0019] Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the coordinates of each wheel end in the world coordinate system are obtained.

[0020] Select any three wheel end coordinates for permutation and combination, and use the least squares method to fit the plane to obtain several plane normal vectors.

[0021] The average value of several plane normal vectors obtained in the previous step is taken as the normal vector of the contact plane.

[0022] The terrain slope is determined according to the normal vector of the contact plane and the mapping relationship between the normal vector of the contact plane and the terrain slope.

[0023] Optionally, the actual wheel information includes an actual wheel speed and an actual wheel angle; using a pre-established wheel kinematics model, according to the actual joint angle, the actual wheel information and the expected wheel information, the wheel control torque is calculated, specifically comprising the following steps:

[0024] The expected wheel speed and the expected wheel angle are determined based on the expected wheel information using the pre-established wheel kinematic model.

[0025] The wheel travel torque is determined according to the acquired actual wheel speed, actual wheel rotation angle, expected wheel speed and expected wheel rotation angle.

[0026] According to the actual joint angle, the position vector of the wheel end relative to the hip joint and the velocity vector of the wheel end relative to the hip joint are determined.

[0027] The wheel adjustment torque is determined according to the position vector of the wheel end relative to the hip joint and the velocity vector of the wheel end relative to the hip joint.

[0028] The wheel control torque is determined based on the wheel travel torque and the wheel adjustment torque.

[0029] In a second aspect, the present application provides a posture control system for a tandem quadruped wheeled robot, comprising:

[0030] The center of mass and wheel-end trajectory planning module is used to generate the expected posture information and expected wheel information of the wheel-legged robot according to the current posture information of the wheel-legged robot, the posture change command and the travel command obtained externally.

[0031] The posture-joint angle calculation module is used to use the pre-established wheel-leg robot synthetic center of mass kinematic analysis model to calculate the expected joint angles according to the expected posture information.

[0032] The joint control module is used to obtain the actual joint angle of the wheel-leg robot and calculate the joint feedback torque based on the actual joint angle and the expected joint angle.

[0033] The extended state observation module is used to observe and obtain the current posture information and error disturbance of the wheel-legged robot based on the actual joint angle and wheel-end force using the extended state observer; the wheel-end force is the ground force acting on the wheel end.

[0034] The MPC posture control module is used to utilize the pre-established wheel-leg robot synthetic center of mass dynamics analysis model to solve the joint feedforward torque and wheel-end force according to the expected posture information, the measured actual posture information and the observed error disturbance.

[0035] The wheel control module is used to utilize a pre-established wheel kinematics model to calculate the wheel control torque according to actual joint angles, actual wheel information and expected wheel information.

[0036] The robot control and feedback module is used to: obtain the joint control torque based on the joint feedback torque and the joint feedforward torque; control the posture change and wheel movement of the wheel-legged robot according to the joint control torque and the wheel control torque; and send the actual joint angle, actual wheel information and actual posture information of the wheel-legged robot to the corresponding module.

[0037] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0038] The present application provides a posture control method and system for a serial quadruped wheel-leg robot, the method comprising: generating the desired posture information and the desired wheel information of the wheel-leg robot according to the current posture information of the wheel-leg robot, the posture command and the travel command obtained externally; using the pre-established wheel-leg robot synthetic center of mass kinematic analysis model, according to the desired posture information, solving the desired joint angle, generating the joint feedback torque based on the actual joint angle and the expected joint angle, and solving the joint feedforward torque according to the observed error disturbance; using the pre-established wheel kinematic model, according to the actual joint angle, the actual wheel information and the expected wheel information, solving the wheel control torque; based on the joint feedback torque and the joint feedforward torque, obtaining the joint control torque, and controlling the posture change and wheel travel of the wheel-leg robot according to the joint control torque and the wheel control torque. The posture control method for a serial quad-wheel-leg robot proposed in the present application effectively ensures the stability of the quad-wheel-leg robot in the driving state through the coordinated action of the wheels and the leg joints, and improves the robot's anti-interference ability and the control ability of the motion posture during the driving process. In the synthetic center of mass kinematic analysis model of the wheel-legged robot pre-established in this application, the leg mass and modeling error are taken into account, and an extended state observer is designed based on the idea of ​​self-disturbance rejection control to observe and compensate for modeling error disturbances, and a wheel-legged robot motion control framework with an unbiased model predictive posture controller as the core is established, which effectively improves the wheel-legged robot's ability to overcome obstacles and resist interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1A flowchart of a posture control method for a tandem quadruped wheeled robot provided in one embodiment of the present application.

[0041] Figure 2 A flowchart of step S5 in a posture control method for a tandem quadruped wheeled robot provided in one embodiment of the present application.

[0042] Figure 3 A schematic diagram of a robot leg structure in a posture control method for a tandem quadruped wheeled robot provided in another embodiment of the present application.

[0043] Figure 4 A schematic diagram of an equivalent form of leg motion in a posture control method for a serial quadruped wheeled-legged robot provided in another embodiment of the present application.

[0044] Figure 5 A schematic diagram of the definition of a wheel-legged robot coordinate system in a posture control method for a serial quadruped wheel-legged robot provided in another embodiment of the present application.

[0045] Figure 6 A schematic diagram of a wheel control strategy in a posture control method for a tandem quadruped wheel-legged robot provided in another embodiment of the present application.

[0046] Figure 7 A schematic diagram of the functional modules of a posture control system for a tandem quadruped wheeled robot provided in one embodiment of the present application.

[0047] Figure 8 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0050] In an exemplary embodiment, Figure 1 As shown, a posture control method for a serial quadruped wheeled robot is provided, comprising the following steps S1 to S6:

[0051] S1. Generate expected posture information and expected wheel information of the wheel-legged robot according to the current posture information of the wheel-legged robot, the posture command and the travel command obtained externally.

[0052] S2. Using the pre-established wheel-leg robot synthetic center of mass kinematics analysis model, according to the expected posture information, the expected joint angle is calculated. The wheel-leg robot synthetic center of mass kinematics analysis model used is expressed by the following formula:

[0053] .

[0054] in, d is the error disturbance, t For time, p is the position vector of the synthetic center of mass of the wheeled-legged robot relative to the origin of the body coordinate system in the world coordinate system, is the velocity of the synthetic center of mass of the wheel-legged robot relative to the origin of the body coordinate system in the world coordinate system, Θ is the posture vector of the wheel-legged robot, w is the angular velocity of the wheel-legged robot, 0 3×3 is a three-row, three-column zero matrix, I 3×3 is the identity matrix with three rows and three columns, , γ is the pitch angle, g represents the gravitational acceleration vector, m is the total mass of the robot, S To select the matrix, I W is the moment of inertia of the synthetic center of mass of the fuselage in the world coordinate system, is the rotation matrix related to the ground slope, r 1 , r 2 , r 3 and r 4 are the position vectors of the wheel ends of the four wheel legs relative to the combined mass center of the wheel-leg robot, [ ]× is the vector cross product symbol, f 1 , f 2 , f 3 and f 4 are the wheel end forces of each wheel leg respectively.

[0055] S3. Obtain the actual joint angle of the wheel-leg robot, and calculate the joint feedback torque based on the actual joint angle and the expected joint angle. The joint feedback torque is calculated according to the following formula:

[0056] .

[0057] in, For the i The joint feedback torque of the wheel leg, and are the proportional gain coefficient matrix and differential gain coefficient matrix of the joint controller, respectively. ref is the expected value, the superscript act is the actual value, and the point on the symbol represents the derivative, that is, the change in angle.

[0058] S4. Using the pre-established wheel-leg robot synthetic center of mass dynamics analysis model, the joint feedforward torque and wheel-end force are calculated based on the expected posture information, the measured actual posture information and the observed error disturbance. The error disturbance is the error value determined by the extended state observer based on the actual joint angle and wheel-end force; the wheel-end force is the ground force acting on the wheel end. The extended state observer is represented by the following formula:

[0059] .

[0060] .

[0061] Among them, the superscript represents an estimated value, , , are the gain matrices of the extended state observer, x is the state vector, u is the control vector, Synthesize the coefficients of the control vectors in the center-of-mass kinematic analysis model for a simplified wheel-legged robot.

[0062] The joint feedforward torque is obtained by solving the following formula:

[0063] .

[0064] in, For the i The joint feedforward torque of the wheel leg, is the Jacobian matrix, For the i The wheel end force of the wheel leg, For the i The position vector of the wheel end of the wheel leg relative to the hip joint, For the i The hip angle of the wheel leg, For the i The knee angle of the wheel leg.

[0065] S5. Using the pre-established wheel kinematics model, the wheel control torque is calculated based on the actual joint angle, the actual wheel information and the expected wheel information. The actual wheel information includes the actual wheel speed and the actual wheel angle; Figure 2 In the flowchart shown, step S5 specifically includes the following steps:

[0066] S51. Determine an expected wheel speed and an expected wheel angle based on expected wheel information using a pre-established wheel kinematics model.

[0067] S52: Determine the wheel travel torque according to the acquired actual wheel speed, actual wheel rotation angle, expected wheel speed and expected wheel rotation angle. Determine the wheel travel torque according to the following formula:

[0068] .

[0069] in, For the i The wheel torque of each wheel, and They are the proportional gain coefficient matrix and differential gain coefficient matrix of the wheel PI controller respectively.

[0070] S53. Determine the position vector of the wheel end relative to the hip joint and the velocity vector of the wheel end relative to the hip joint based on the actual joint angle.

[0071] S54, determining the wheel adjustment torque according to the position vector of the wheel end relative to the hip joint and the speed vector of the wheel end relative to the hip joint. The wheel adjustment torque is determined according to the following formula:

[0072] .

[0073] in, For the i The wheel adjustment torque of each wheel, and are the proportional gain coefficient matrix and differential gain coefficient matrix of the wheel PD controller respectively, express Unit vector of direction, superscript ref is the expected value, the superscript act is the actual value, and the point on the symbol represents the derivative, that is, the change in the position vector.

[0074] S55. Determine the wheel control torque according to the wheel travel torque and the wheel adjustment torque. The wheel control torque is obtained according to the following formula:

[0075] .

[0076] in, For the iThe wheel control torque of each wheel, and are the weights of the joint feedforward torque and the joint feedback torque respectively.

[0077] S6. Based on the joint feedback torque and the joint feedforward torque, the joint control torque is obtained. The joint control torque is obtained according to the following formula:

[0078] .

[0079] in, For the i The joint control torque of the wheel leg, and are the weights of the joint feedforward torque and the joint feedback torque respectively.

[0080] S7. Control the posture change and wheel movement of the wheel-legged robot according to the joint control torque and the wheel control torque.

[0081] In another exemplary embodiment of the present application, before performing step S1, the following steps are also included:

[0082] A1. Based on the structural characteristics of the serial quadruped wheel-legged robot, a kinematic analysis model of the serial quadruped wheel-legged robot is established.

[0083] A2. Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the contact plane is fitted using the least squares method to obtain the terrain slope. Step A2 specifically includes the following steps:

[0084] A21. Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the coordinates of each wheel end in the world coordinate system are obtained.

[0085] A22. Select any three wheel end coordinates for arrangement and combination, and use the least squares method to fit the plane to obtain several plane normal vectors.

[0086] A23. Take the average value of several plane normal vectors obtained in the previous step as the normal vector of the contact plane.

[0087] A24. Determine the terrain slope based on the normal vector of the contact plane and the mapping relationship between the normal vector of the contact plane and the terrain slope. Specifically, the mapping relationship between the normal vector of the contact plane and the terrain slope is expressed by the following formula:

[0088] .

[0089] in, is the longitudinal slope, is the side slope, , n is the normal vector of the contact plane.

[0090] A3. Based on the terrain slope and the structural characteristics of the four-group wheel-leg robot in series, a synthetic center of mass kinematic analysis model of the wheel-leg robot is established with comprehensive leg mass and modeling error.

[0091] In the following detailed embodiment, the process of establishing a wheel-legged robot synthetic center of mass kinematic analysis model that integrates leg mass and modeling error and establishing an extended state observer will be described in detail, including the following contents:

[0092] Based on the structural characteristics of the serial four-wheeled legged robot, its kinematic analysis is first carried out. Figure 3 The robot leg structure is shown, including: , , , Respectively represent connecting rod , , , Length, It is a connecting rod The distance between the two upper hinge points, and , , the four hinge points and the connecting rod form a parallelogram. Hip joint motor drives the connecting rod Rotation, that is, driving the thigh to rotate, the knee joint motor drives the connecting rod Rotation, through the parallelogram connecting rod transmission, that is, driving the connecting rod Rotation and connecting rod Rotation angle and connecting rod The rotation angle is the same.

[0093] Therefore, the leg kinematic relationship can be equivalent to Figure 4 shown. Indicates the length of the thigh, Indicates the length of the calf. , , They are The hip, knee and wheel angles of the legs, Indicates The legs are fixed to the body of the coordinate system, where They represent the left front leg, right front leg, left hind leg and right hind leg respectively.

[0094] According to the kinematic relationship, we can get The position vector of the thigh mass center relative to the hip joint in the coordinate system , the position vector of the calf mass center relative to the hip joint and the position vector of the wheel end relative to the hip joint It can be expressed by the following formula (1):

[0095] (1).

[0096] Figure 5 This is a schematic diagram of the coordinate system of the wheel-legged robot. represents the world coordinate system, is the body coordinate system of the wheel-legged robot, is the center of mass of the robot. The point in the figure coincides with the origin of the body coordinate system, which is a special case. To maintain generality, the position vector of the center of mass in the world coordinate system relative to the origin of the body coordinate system is expressed as , , are the length and width of the hip joint from the center of mass, respectively.

[0097] Since the robot's legs do not have an outward swing joint, the yaw angle in the posture freedom degree is not considered, only the roll angle is considered and pitch angle , then the rotation matrix of the fuselage coordinate system relative to the world coordinate system is as shown in equation (2).

[0098] (2).

[0099] The position vector of the thigh mass center relative to the hip joint in the world coordinate system , the position vector of the calf mass center relative to the hip joint and the position vector of the wheel end relative to the hip joint It can be expressed as formula (3):

[0100] (3).

[0101] Combined with the structural parameters, the position vector of the wheel end relative to the center of mass in the world coordinate system is obtained as shown in the following formula (4):

[0102] (4).

[0103] in, is a symbol matrix, which can be expressed as formula (5):

[0104] (5).

[0105] In order to obtain a more accurate dynamic model, the terrain slope angle factor needs to be considered to generalize the dynamic equation. After the coordinates of each wheel end in the world coordinate system are obtained by equation (4), three wheel end coordinates are randomly selected for permutation and combination, and the least squares method is used to fit the plane to obtain the plane normal vector of the group. Finally, the average value of all normal vectors is calculated as the normal vector of the contact plane. The slope of the terrain includes the longitudinal slope and side slope Since the yaw angle in the fuselage attitude is not considered, the relationship between the contact plane normal vector and the slope of the terrain is as shown in formula (6).

[0106] (6).

[0107] Therefore, the slope of the plane that the current wheel-legged robot is in contact with can be indirectly estimated based on the angle feedback from the joint encoder and the posture measured by the IMU (Inertial measurement unit), thereby obtaining the rotation matrix related to the ground slope. , its form is similar to formula (2).

[0108] According to the mass point theory, the position vector of the synthetic center of mass of the wheeled leg robot in the world coordinate system relative to the origin of the body coordinate system can be obtained: As shown in formula (7):

[0109] (7).

[0110] in, The quality of the fuselage. , Indicates The thigh mass, calf mass and wheel mass of the leg, Then represents the corresponding position vector. Combining formula (4), we can get The expression of the wheel end of the leg relative to the composite center of mass is shown in formula (8):

[0111] (8).

[0112] According to the definition of Euler angle, without considering the yaw angle, the angular velocity of the wheeled leg robot in the world coordinate system can be expressed as follows (9):

[0113] (9).

[0114] Ignoring the moment of inertia of the leg links and wheels, the moment of inertia of the composite center of mass of the fuselage in the world coordinate system is It can be obtained by formula (10). Among them, is the moment of inertia of the center of mass in the fuselage coordinate system.

[0115] (10).

[0116] Select the posture vector as Based on the Newton-Euler method, through equations 8) to (10), the synthetic center of mass dynamic state space equation can be obtained as shown in equation (11).

[0117] (11).

[0118] in, represents the time derivative, Represents the gravity acceleration vector, which acts only in the Z direction; is the total weight of the robot, symbol represents the vector cross product, is the ground reaction force of each leg, , , , is a bounded perturbation vector about the modeling error, It is the selection matrix used to filter the roll angle and pitch angle degrees of freedom. represents a zero matrix with three rows and three columns, and the others are similar; represents the identity matrix with three rows and three columns; is the position vector of the wheel end of the four legs relative to the synthetic center of mass, obtained by expression (8); It is the first matrix on the right side of the equal sign in equation (9).

[0119] For the convenience of subsequent expression, the above equation can be simplified into the form shown in equation (12).

[0120] (12).

[0121] in, , , , , Contains gravitational acceleration and disturbance terms.

[0122] , , , .

[0123] The main movement form of the tandem four-wheeled legged robot is to drive the robot through the wheels and adjust the posture through the leg joints, so that the robot can smoothly pass through various terrains. In the synthetic center of mass dynamics model, the present invention takes into account the error disturbance caused by the unmodeled characteristics and considers that the disturbance is bounded. The state quantity, control quantity and disturbance quantity can be defined as a new vector , h For disturbance d The derivative of , then equation (11) can be expanded into the form of equation (13).

[0124] (13).

[0125] in, , according to the matrix It can be judged that the dynamic system is observable, so the Romberg observer is constructed as shown in the following formula (14), with the superscript represents an estimated value, , and are the gain matrices of the observer, where the β with two subscripts represents the diagonal elements in the corresponding gain matrix, such as β 1,1 Represents the matrix β 1 The first row and first column of the element, β 2,10 Represents the matrix β 2 The element in the tenth row and tenth column of .

[0126] (14).

[0127] Using the inertial measurement unit and joint encoder information, combined with equation (6), the feedback state quantity can be obtained: From the relationship between system stability and zero-pole configuration, it can be seen that by setting a suitable gain matrix to make the observer converge, a more accurate unmodeled error disturbance can be obtained.

[0128] In the following embodiments, how to perform posture control based on the synthetic center of mass dynamics model constructed above will be described in detail, including the following contents:

[0129] The forward Euler method is used to discretize the synthetic center of mass dynamics model considering disturbances, and the changes of disturbances within a control cycle are ignored, and equation (15) is obtained.

[0130] (15).

[0131] in, is the step length, I is the identity matrix, .

[0132] Define the number of prediction steps as , the number of control steps is , and satisfy Based on formula (15), we can get In the moment Step control sequence And the state output sequence As shown in the following formula (16).

[0133] (16).

[0134] The matrix subscript indicates the number of vectors in the matrix, for example Only indicates middle For the future of the system The output of the step prediction can be calculated by the following equation (17): is the coefficient matrix constructed according to formula (16).

[0135] (17).

[0136] The expected reference sequence can be expressed as formula (18).

[0137] (18).

[0138] In the formula, It contains the expected value of the state quantity, which can be indirectly calculated by setting the expected center of mass posture. Considering the control requirements, three cost functions are established in this embodiment as shown in the following formula (19).

[0139] (19).

[0140] In the formula, Represents the weight The two-norm form of is a diagonally positive definite weight matrix, and the total cost function is The first cost function is to make the actual attitude state as close as possible to the desired state, the second cost is to make the rate of change of the control quantity as small as possible to ensure that the system does not move violently, and the last cost function emphasizes that the energy consumption is as small as possible.

[0141] For the constraints of the wheel-leg robot posture control problem, only the ground force on the wheel end is considered to be always in the force friction cone, then the wheel end force constraint of each leg can be expressed as formula (20):

[0142] (20).

[0143] After constructing the model predictive control problem, the control sequence is obtained by solving it through the quadratic programming method, and the first vector is used as the controller output .

[0144] The mapping relationship between the ground force on the wheel end and the joint torque is expressed by the Jacobian matrix Realization, Joint feedforward torque of the legs It can be obtained by the following formula (21). is the output of the model predictive controller. The wheel end force of the leg.

[0145] (twenty one).

[0146] The feedforward torque can maintain the balance of posture, but it cannot limit the angle of the leg joints. Therefore, a joint angle closed loop needs to be added to the control strategy to avoid the phenomenon that the four legs are in inconsistent angle states due to the use of only feedforward torque for posture control.

[0147] According to the kinematic relationship, the desired posture trajectory is solved into the desired joint angle and angular velocity , combined with the actual angle information fed back by the motor encoder The feedback torque output by the joint controller can be obtained As shown in formula (22).

[0148] (twenty two).

[0149] in, is the proportional gain coefficient matrix and differential gain coefficient matrix of the joint controller, which is a diagonally positive definite gain matrix, and the gain is set to a small value to avoid overshoot of leg movement. Give corresponding weights to the feedforward torque and feedback torque , and the final control torque signal input to the joint motor on each leg is shown in (23).

[0150] (twenty three).

[0151] Since there is no closed loop control of the robot posture in the process of generating feedback torque, the main role of posture adjustment is the feedforward torque, so the weight matrix must be set to meet Diagonal elements greater than Diagonal elements.

[0152] The friction cone constraint in the model predictive controller and the feedback torque of the joint controller cannot completely limit the movement of the wheel end during the posture adjustment process. Considering that the robot has wheel-driven degrees of freedom, a wheel control strategy is proposed for traveling and assisting in adjusting the leg state. The block diagram of the control strategy is shown in Figure 6 According to the speed difference steering kinematic reference model, the remote control forward or steering command is resolved into the desired wheel speed. and wheel angle , the actual wheel speed and wheel angle The torque that drives the wheel-legged robot to move is obtained from the encoder of the wheel motor It can be obtained by the following formula (24). are the proportional gain and integral gain coefficients of the wheel PI controller.

[0153] (twenty four).

[0154] By substituting the angle of the joint motor encoder feedback into equation (3), the position vector of the wheel end relative to the hip joint in the current state can be calculated: , and obtain the velocity vector according to the kinematic relationship , express The unit vector of the direction, Direction Expected wheel end position According to experience, the position just below the hip joint and slightly inside is usually selected. As shown in the following formula (25), are the proportional gain coefficient and the differential gain coefficient of the wheel PD controller.

[0155] (25).

[0156] Final wheel control torque It can be expressed as shown in the following formula (26):

[0157] (26).

[0158] in, is the weight of the stress moment.

[0159] The above-mentioned embodiments proposed in the present application effectively ensure the stability of the four-wheeled legged robot in the driving state through the coordinated action of the wheels and leg joints, and improve the robot's anti-interference ability and control ability of the motion posture during driving; in the pre-established wheel-legged robot synthetic center of mass kinematic analysis model, the leg mass and modeling error are taken into account, and an extended state observer is designed based on the self-disturbance rejection control idea to observe and compensate for the modeling error disturbance, and establish a wheel-legged robot motion control framework with an unbiased model prediction posture controller as the core, which effectively improves the wheel-legged robot's ability to overcome obstacles and resist interference.

[0160] Based on the same inventive concept, the embodiment of the present application also provides a posture control system for a tandem quadruped wheeled leg robot for implementing the posture control method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more posture control system embodiments provided below can refer to the limitations on the posture control method above, and will not be repeated here.

[0161] In an exemplary embodiment, Figure 7 As shown, a posture control system for a serial quadruped wheeled robot is provided, comprising:

[0162] The center of mass and wheel-end trajectory planning module is used to generate the expected posture information and expected wheel information of the wheel-legged robot according to the current posture information of the wheel-legged robot, the posture change command and the travel command obtained externally.

[0163] The posture-joint angle calculation module is used to use the pre-established wheel-leg robot synthetic center of mass kinematic analysis model to calculate the expected joint angles according to the expected posture information.

[0164] The joint control module is used to obtain the actual joint angle of the wheel-leg robot and calculate the joint feedback torque based on the actual joint angle and the expected joint angle.

[0165] The extended state observation module is used to observe and obtain the current posture information and error disturbance of the wheel-legged robot based on the actual joint angle and wheel-end force using the extended state observer; the wheel-end force is the ground force acting on the wheel end.

[0166] The MPC posture control module is used to utilize the pre-established wheel-leg robot synthetic center of mass dynamics analysis model to solve the joint feedforward torque and wheel-end force according to the expected posture information, the measured actual posture information and the observed error disturbance.

[0167] The wheel control module is used to utilize a pre-established wheel kinematics model to calculate the wheel control torque according to actual joint angles, actual wheel information and expected wheel information.

[0168] The robot control and feedback module is used to: obtain the joint control torque based on the joint feedback torque and the joint feedforward torque; control the posture change and wheel movement of the wheel-legged robot according to the joint control torque and the wheel control torque; and send the actual joint angle, actual wheel information and actual posture information of the wheel-legged robot to the corresponding module.

[0169] In further embodiments, some modules may also have subunits for implementing their functions. Figure 7 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different functions. Figure 7 One or at least two components of the system shown.

[0170] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 8As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is used to implement a posture control method for a serial quadruped wheeled leg robot.

[0171] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0172] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0173] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0174] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0175] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0176] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0177] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0178] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0179] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A posture control method for a tandem quadruped wheeled robot, characterized in that: include: Generate the desired posture information and desired wheel information of the wheel-legged robot according to the current posture information of the wheel-legged robot, the posture command and the travel command obtained externally; Using a pre-established wheel-leg robot synthetic center of mass kinematics analysis model, according to the desired posture information, the desired joint angles are calculated; Acquire the actual joint angle of the wheel-leg robot, and calculate the joint feedback torque based on the actual joint angle and the expected joint angle; Using the pre-established wheel-leg robot synthetic center of mass dynamics analysis model, the joint feedforward torque and wheel-end force are calculated according to the desired posture information, the measured actual posture information and the observed error disturbance; the error disturbance is an error value determined according to the actual joint angle and wheel-end force using an extended state observer; The wheel end force is the ground force acting on the wheel end; Using a pre-established wheel kinematics model, according to the actual joint angle, the actual wheel information and the expected wheel information, the wheel control torque is calculated; Based on the joint feedback torque and the joint feedforward torque, a joint control torque is obtained, and according to the joint control torque and the wheel control torque, a posture change of the wheel-legged robot and a wheel travel are controlled; The posture control method also includes: Based on the structural characteristics of the serial quadruped wheel-legged robot, a kinematic analysis model of the serial quadruped wheel-legged robot is established; Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the contact plane is fitted using the least square method to obtain the terrain slope. Based on the terrain slope and the structural characteristics of the series-type four-group wheel-leg robot, a wheel-leg robot synthetic center of mass kinematic analysis model is established that integrates leg mass and modeling error; Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the contact plane is fitted using the least squares method to obtain the terrain slope, including: Based on the kinematic analysis model of the serial quadruped wheel-legged robot, the coordinates of each wheel end in the world coordinate system are obtained; Select any three wheel end coordinates for permutation and combination, and use the least square method to fit the plane to obtain several plane normal vectors; The average value of several plane normal vectors obtained in the previous step is used as the normal vector of the contact plane; Determine the terrain slope according to the normal vector of the contact plane and the mapping relationship between the normal vector of the contact plane and the terrain slope; The mapping relationship between the normal vector of the contact plane and the terrain slope is expressed by the following formula: ； in, is the longitudinal slope, is the side slope, , n is the normal vector of the contact plane; The wheel-leg robot synthetic center of mass kinematics analysis model is expressed by the following formula: ； in, d is the error disturbance, t For time, p is the position vector of the synthetic center of mass of the wheeled-legged robot relative to the origin of the body coordinate system in the world coordinate system, is the velocity of the synthetic center of mass of the wheel-legged robot relative to the origin of the body coordinate system in the world coordinate system, Θ is the posture vector of the wheel-legged robot, w is the angular velocity of the wheel-legged robot, 0 3×3 is a three-row, three-column zero matrix, I 3×3 is the identity matrix with three rows and three columns, , γ is the pitch angle, g represents the gravitational acceleration vector, m is the total mass of the robot, S To select the matrix, I W is the moment of inertia of the synthetic center of mass of the fuselage in the world coordinate system, is the rotation matrix related to the ground slope, r 1, r 2, r 3 and r 4 are the position vectors of the wheel ends of the four wheel legs relative to the combined mass center of the wheel-leg robot, [ ]× is the vector cross product symbol, f 1. f 2. f 3 and f 4 are the wheel end forces of each wheel leg.

2. The posture control method for a tandem quadruped wheeled robot according to claim 1, characterized in that: The extended state observer is expressed as follows: ； ； Among them, the superscript represents an estimated value, , , are the gain matrices of the extended state observer, x is the state vector, u is the control vector, Synthesize the coefficients of the control vectors in the center-of-mass kinematic analysis model for a simplified wheel-legged robot.

3. The posture control method for a tandem quadruped wheeled robot according to claim 2, characterized in that: The joint feedforward torque is obtained by solving the following formula: ； in, For the i The joint feedforward torque of the wheel leg, is the Jacobian matrix, For the i The wheel end force of the wheel leg, For the i The position vector of the wheel end of the wheel leg relative to the hip joint, For the i The hip angle of the wheel leg, For the i The knee angle of the wheel leg; The joint feedback torque is obtained by solving the following formula: ； in, For the i The joint feedback torque of the wheel leg, and are the proportional gain coefficient matrix and differential gain coefficient matrix of the joint controller, respectively. ref is the expected value, the superscript act is the actual value, and the point on the symbol represents the derivative, that is, the change in angle; The joint control torque is obtained according to the following formula: ； in, For the i The joint control torque of the wheel leg, and are the weights of the joint feedforward torque and the joint feedback torque, respectively.

4. The posture control method for a tandem quadruped wheeled robot according to claim 3, characterized in that: The actual wheel information includes the actual wheel speed and the actual wheel angle; the wheel control torque is calculated based on the actual joint angle, the actual wheel information and the expected wheel information using a pre-established wheel kinematics model, specifically including: Determining an expected wheel speed and an expected wheel angle based on the expected wheel information by using a pre-established wheel kinematics model; Determine the wheel travel torque according to the acquired actual wheel speed, the actual wheel rotation angle, the expected wheel speed and the expected wheel rotation angle; Determine the position vector of the wheel end relative to the hip joint and the velocity vector of the wheel end relative to the hip joint according to the actual joint angle; Determine the wheel adjustment torque according to the position vector of the wheel end relative to the hip joint and the speed vector of the wheel end relative to the hip joint; A wheel control torque is determined according to the wheel travel torque and the wheel adjustment torque.

5. The posture control method for a tandem quadruped wheeled robot according to claim 4, characterized in that: The wheel travel torque is determined by the following formula: ； in, For the i The wheel torque of each wheel, and They are the proportional gain coefficient matrix and differential gain coefficient matrix of the wheel PI controller respectively; The wheel adjustment torque is determined according to the following formula: ； in, For the i The wheel adjustment torque of each wheel, and are the proportional gain coefficient matrix and differential gain coefficient matrix of the wheel PD controller respectively, express Unit vector of direction, superscript ref is the expected value, the superscript act is the actual value, and the point on the symbol represents the derivative, that is, the change in the position vector; The wheel control torque is obtained according to the following formula: ； in, For the i The wheel control torque of each wheel, and are the weights of the joint feedforward torque and the joint feedback torque, respectively.

6. A posture control system for a tandem quadruped wheeled robot, characterized in that: Used to implement the posture control method for a serial quadruped wheel-legged robot as described in any one of claims 1 to 5, the posture control system for the serial quadruped wheel-legged robot comprises: The center of mass and wheel-end trajectory planning module is used to generate the desired posture information and desired wheel information of the wheel-legged robot according to the current posture information of the wheel-legged robot, the posture change command and the travel command obtained externally; A posture-joint angle calculation module is used to calculate the expected joint angle according to the expected posture information by using a pre-established wheel-leg robot synthetic center of mass kinematic analysis model; The joint control module is used to obtain the actual joint angle of the wheel-legged robot, and calculate the joint feedback torque based on the actual joint angle and the expected joint angle; The extended state observation module is used to observe and obtain the current posture information and error disturbance of the wheel-legged robot according to the actual joint angle and wheel-end force by using the extended state observer; the wheel-end force is the ground force acting on the wheel end; The MPC posture control module is used to use a pre-established wheel-leg robot synthetic center of mass dynamics analysis model to solve the joint feedforward torque and wheel-end force according to the desired posture information, the measured actual posture information and the observed error disturbance; A wheel control module, used to calculate the wheel control torque according to the actual joint angle, the actual wheel information and the expected wheel information by using a pre-established wheel kinematics model; The robot control and feedback module is used to: obtain the joint control torque based on the joint feedback torque and the joint feedforward torque; control the posture change and wheel movement of the wheel-legged robot according to the joint control torque and the wheel control torque; and send the actual joint angle, actual wheel information and actual posture information of the wheel-legged robot to the corresponding module.

Images