Track optimization and motion generation method, controller and system of four-wheel-foot robot

Abstract

The invention belongs to the field of four-wheel-foot robots, and provides a track optimization and motion generation method, controller and system for a four-wheel-foot robot. The method comprises the steps that a trajectory optimization objective function is constructed based on a motion dynamics model of the four-wheel-foot robot, and the whole-body optimal trajectory of the four-wheel-foot robot in a motion period is solved under the condition that set constraint conditions of the four-wheel-foot robot are met; introducing a virtual joint angle, enabling the world coordinate system to be always aligned with the contact coordinate system through rotation transformation, and constructing a single-wheel-leg kinematics model containing a wheel-ground contact point; and performing inverse kinematics calculation on the single-wheel-leg kinematics model containing the wheel-ground contact points to obtain position information of the wheel legs and the body, and performing motion control on the four-wheel-leg robot in combination with expected foot end contact force.

Description

Technical Field

[0001] The present invention belongs to the field of four-wheeled legged robots, and in particular relates to a trajectory optimization and motion generation method, a controller and a system for a four-wheeled legged robot. Background Art

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

[0003] Humans and animals perform complex locomotion behaviors that involve rapid interactions between the feet and the environment. Existing hybrid locomotion methods are based on handcrafted heuristics for online gait generation to decide when to step and when to drive (roll). Most gait patterns or motion primitives designed for legged robots are inspired by nature, but not by observations of biological organisms for wheeled-legged robots. At the same time, such handcrafted locomotion patterns cannot adapt to rough terrains and lack robustness to unperceived obstacles.

[0004] When approaching physical limits or overcoming challenging obstacles, the whole body coordinated movement of the wheeled robot is necessary, and the next movement needs to be planned in advance. However, designing high-dimensional trajectories, such as full-body trajectories that include ground reaction forces, is usually very difficult. In most cases, manual design is almost impossible due to the complexity of the constraints. The prior art provides a two-step trajectory planning technology for a quadruped wheeled robot, which only optimizes the trajectory by constructing an objective function for the wheels, but cannot solve the motion planning problem within the task time range, that is, simultaneously optimize the high-dimensional variable space composed of the whole body trajectory, gait sequence and time, and generate multi-mode multi-gait motion while taking into account the smoothness and stability of the motion. Summary of the invention

[0005] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a trajectory optimization and motion generation method, controller and system of a four-wheeled legged robot, which is based on nonlinear optimization and wheel-legged inverse kinematics to generate an action trajectory that satisfies complex constraints, and can produce complex mixed motions of the wheel-legged robot that are close to whole-body coordination within the time range of a specific task, thereby realizing multi-mode and multi-gait motions of the wheel-legged robot.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a trajectory optimization and motion generation method for a four-wheeled legged robot.

[0008] A method for trajectory optimization and motion generation of a four-wheeled legged robot, comprising:

[0009] Based on the kinematics model of the four-wheeled legged robot, a trajectory optimization objective function is constructed, and the optimal trajectory of the whole body of the four-wheeled legged robot within the motion cycle is solved under the set constraints of the four-wheeled legged robot; wherein the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the trunk coordinate system and the velocity of the wheel-ground contact point in the world coordinate system;

[0010] Based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, a virtual joint angle is introduced, and the world coordinate system is always aligned with the contact coordinate system through rotation transformation, thus constructing a single-wheel leg kinematic model including the wheel-ground contact point.

[0011] The inverse kinematics of the single-wheel leg kinematic model including the wheel-ground contact point is solved to obtain the position information of the wheel leg and the body, and then combined with the expected foot end contact force, the motion of the four-wheeled leg robot is controlled.

[0012] As an embodiment, the single-wheel leg kinematic model including the wheel-ground contact point represents the wheel speed ω and the end effector speed v E The relationship between them can be expressed as:

[0013]

[0014] Where J(q)∈R 3×3 is the Jacobian matrix of the leg; q = [q1, q2, q3] represents the leg joint angle vector; q1, q2, q3 represent the first leg joint angle, the second leg joint angle and the third leg joint angle respectively; represents the time derivative of the leg joint angle vector; r 4 is the distance between the wheel-ground contact point and the wheel center; is the rotation matrix of the wheel center coordinate system relative to the body coordinate system.

[0015] As an implementation method, the expression f(y(t)) of the trajectory optimization objective function is:

[0016]

[0017] Among them, y(t) is the optimization variable, which includes the four-wheeled robot in the motion period T TO The whole body trajectory on x torso (t)∈R 12 Contains the torso's heading θ∈R in Euler angles in the world coordinate system 3 and the centroid position p∈R 3 , the angular velocity of the center of mass in the torso coordinate system ω∈R 3 With linear velocity v∈R 3 , x leg,i(t) contains the position of the wheel-ground contact point in the world coordinate system speed and contact force Q is the semi-positive Hessian matrix, the state vector error n x =[1,0,0] T Choose a vector for the x direction, λ c ∈(0,1) is the end contact state; the superscript T represents the transpose of the vector.

[0018] As an implementation method, the constraint conditions set for the quadruped robot include: initial and end state constraints, kinematic constraints, dynamic constraints and wheel-ground contact constraints.

[0019] As an implementation method, the kinematic constraints are determined by the allowed joint angle range and ensure that there is no interference between the front and rear wheels;

[0020]

[0021] Among them, K i represents the predefined working space of the end effector, and R represents the radius of the wheel.

[0022] As an implementation method, the dynamic constraint is characterized by a single rigid body dynamic model:

[0023]

[0024] in, Represents the angular velocity of the torso in Euler angles in the world coordinate system; Represents the center of mass velocity of the torso in the world coordinate system; represents the angular acceleration of the center of mass in the torso coordinate system; represents the linear acceleration of the center of mass in the torso coordinate system; represents the rotation matrix, which maps the components of the vector from the torso coordinate system B to the world coordinate system W; T(θ) is the transformation matrix that maps the angular velocity of the torso to the Euler angular velocity in the world coordinate system; I W is the inertia tensor of the torso; m is the total mass of the robot; g(θ) is the gravitational acceleration in the torso coordinate system.

[0025] As an implementation manner, the wheel-ground contact constraint includes:

[0026]

[0027] (x normal ×n)·n=0

[0028]

[0029] in, is the elevation map h terrain (x,y) Local surface normal in world coordinate system W; friction cone constraint limits ground reaction force Stay at the friction coefficient μ C The Coulomb friction cone C(n,μ C ), the trunk direction x normal Approximately the direction in which the wheels roll.

[0030] As an implementation mode, the initial and end state constraints are artificially set in advance.

[0031] A second aspect of the present invention provides a controller.

[0032] A controller comprising:

[0033] The whole body optimal trajectory solving module is used to construct a trajectory optimization objective function based on the motion dynamics model of the four-wheeled legged robot, and solve the whole body optimal trajectory of the four-wheeled legged robot within the motion cycle under the constraints set by the four-wheeled legged robot; wherein the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the torso coordinate system and the velocity of the wheel-ground contact point in the world coordinate system;

[0034] The kinematic model building module is used to introduce a virtual joint angle based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, and to align the world coordinate system with the contact coordinate system through rotation transformation, so as to build a single-wheel leg kinematic model including the wheel-ground contact point;

[0035] The inverse kinematics solution and control module is used to perform inverse kinematics solution on the single-wheel leg kinematic model including the wheel-ground contact point, obtain the position information of the wheel leg and the body, and then combine the expected foot end contact force to control the motion of the four-wheeled leg robot.

[0036] A third aspect of the present invention provides a four-wheeled legged robot system.

[0037] A four-wheeled legged robot system comprises a four-wheeled legged robot and a controller, wherein the controller is used for controlling the movement of the four-wheeled legged robot.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention provides a trajectory optimization and motion generation method for a four-wheeled leg robot. Based on the optimal whole-body trajectory of the four-wheeled leg robot during the motion cycle and the single-wheel leg kinematic model including the wheel-ground contact point, combined with inverse kinematics solution, a complex hybrid motion of the wheeled leg robot is generated that enables the robot to approach whole-body coordination within the time range of a specific task, thereby realizing the multi-mode and multi-gait motion of the four-wheeled leg robot. At the same time, it is also convenient for researchers to manually design various actions and produce wheeled leg robot data sets, providing a direction for future research on navigation and mobility of wheeled leg robots.

[0040] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 It is a schematic diagram of the structure of a four-wheeled legged robot;

[0043] Figure 2 The single-wheel leg kinematic model including the wheel-ground contact point according to an embodiment of the present invention;

[0044] Figure 3 It is a motion diagram for wheeled movement patterns while planning the trunk posture angle;

[0045] Figure 4 It is the leg joint angle curve that simultaneously plans the trunk posture angle in the wheeled movement mode;

[0046] Figure 5 It is the wheel speed curve of each wheel movement mode while planning the trunk posture angle;

[0047] Figure 6 This is the wheel-foot complex motion diagram in the trot gait;

[0048] Figure 7 is the leg joint angle curve of the wheel-foot complex motion in the trot gait;

[0049] Figure 8 It is the wheel speed curve of the wheel-foot complex motion in the trot gait;

[0050] Fig. 9 It is the wheel-foot complex motion diagram in bound gait;

[0051] Fig.10 is the leg joint angle curve of the wheel-foot complex motion in bound gait;

[0052] Fig.11 It is the wheel speed curve of the wheel-foot complex motion in bound gait;

[0053] Fig.12 is a flow chart of a trajectory optimization and motion generation method of a four-wheeled legged robot according to an embodiment of the present invention;

[0054] Fig.13 Schematic diagram of the controller structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0056] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. 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 invention belongs.

[0057] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. 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.

[0058] Legged robots are promising mobile platforms that can walk on unstructured terrains such as steep slopes, steps and gaps by freely choosing the support surface. However, it is still very challenging for robots to have both high mobility and off-road capabilities in complex terrain in the wild - the robot must be fast and efficient on flat terrain, while being able to overcome terrain such as mud, gravel, steps, etc. Traditional wheeled robots cannot effectively overcome these obstacle terrains, and legged robots alone are not enough to achieve the necessary speed and efficiency. Wheeled-legged robots provide a comprehensive solution to meet these requirements. They can use drive wheels to reduce transportation costs, while using legged movement to overcome challenging obstacles. The present invention is aimed at a four-wheeled, footed robot that is equipped with a drive wheel at the end of a four-legged robot.

[0059] In view of the problems that wheeled robots have no bionic simulation objects, that is, lack of such bio-inspired motion primitives, and at the same time, wheeled robots have many degrees of freedom in their joints and complex motion behaviors, the present invention proposes a trajectory optimization and motion generation method, controller and system for four-wheeled robots. Based on nonlinear optimization and wheel-foot inverse kinematics, motion trajectories that meet complex constraints are generated, which provides a promising direction for future research on navigation and mobility of wheeled robots.

[0060] Embodiment 1

[0061] Fig.12 FIG. 1 is a flow chart of the trajectory optimization and motion generation method of the four-wheeled legged robot according to an embodiment of the present invention. Fig.12 As shown, this embodiment provides a trajectory optimization and motion generation method for a four-wheeled legged robot, which includes:

[0062] S1201: Based on the motion dynamics model of the four-wheeled legged robot, a trajectory optimization objective function is constructed, and under the set constraints of the four-wheeled legged robot, the optimal trajectory of the whole body of the four-wheeled legged robot within the motion cycle is solved; wherein, the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the torso coordinate system and the velocity of the wheel-ground contact point in the world coordinate system.

[0063] The four-wheeled legged robot structure of this embodiment is as follows: Figure 1 As shown, each leg is equipped with three joint motors and one hub motor.

[0064] The trajectory optimization problem is expressed as a nonlinear optimization problem including the optimization target f(y(t)), the equality constraint c(y(t)) and the inequality constraint h(y(t)). In this embodiment, the robot trunk velocity and the terminal velocity are added to the optimization problem to find the optimal trajectory that satisfies the kinematic and dynamic constraints, that is:

[0065]

[0066] In the specific implementation process, the speed of the trunk is composed of the leg joint movement speed and the wheel speed. For the consideration of stability and energy utilization efficiency, the wheel speed is expected to be as large as possible. Therefore, the expression f(y(t)) of the trajectory optimization objective function is designed as follows:

[0067]

[0068] x torso (t) = [θ T p T ω T v T ] T

[0069]

[0070] Among them, y(t) is the optimization variable, which includes the four-wheeled robot in the motion period T TO The whole body trajectory on x torso (t)∈R 12 Contains the torso's heading θ∈R in Euler angles in the world coordinate system 3 and the centroid position p∈R3 , the angular velocity of the center of mass in the torso coordinate system ω∈R 3 With linear velocity v∈R 3 , x leg,i (t) contains the position of the wheel-ground contact point in the world coordinate system speed and contact force Q is the semi-positive Hessian matrix, the state vector error n x =[1,0,0] T Choose a vector for the x direction, λ c ∈(0,1) is the end contact state; the superscript T represents the transpose of the vector.

[0071] Specifically, the constraints set for the quadruped robot include: initial and end state constraints, kinematic constraints, dynamic constraints and wheel-ground contact constraints.

[0072] (1) Initial and final state constraints:

[0073] The initial and final states are determined by equality constraints as follows:

[0074]

[0075] The initial state With the end state Pre-set by humans.

[0076] (2) Kinematic constraints

[0077] The kinematic constraints are determined by the allowed joint angle range and ensure that there is no interference between the front and rear wheels;

[0078]

[0079] Among them, K i represents the predefined working space of the end effector, and R represents the radius of the wheel.

[0080] It should be noted here that the predefined working space of the end effector may be a rectangular space, a spherical space or other artificially set three-dimensional space of irregular shape.

[0081] (3) Dynamic constraints

[0082] The dynamics are based on a kinematic model of a quadrupedal robot, which defines a single rigid body dynamics model (SRBD) as well as the kinematics of each leg while treating the wheels as mobile ground contacts with locked rotation angles. The SRBD model assumes that the momentum of the limb joints is negligible relative to the momentum of the trunk.

[0083] The dynamic constraints are characterized by a single rigid body dynamics model:

[0084]

[0085] in, Represents the angular velocity of the torso in Euler angles in the world coordinate system; Represents the center of mass velocity of the torso in the world coordinate system; represents the angular acceleration of the center of mass in the torso coordinate system; represents the linear acceleration of the center of mass in the torso coordinate system; represents the rotation matrix, which maps the components of the vector from the torso coordinate system B to the world coordinate system W; T(θ) is the transformation matrix that maps the angular velocity of the torso to the Euler angular velocity in the world coordinate system; I W is the inertia tensor of the torso; m is the total mass of the robot; g(θ) is the gravitational acceleration in the torso coordinate system.

[0086] (4) Wheel-ground contact constraints

[0087] When the wheel is in the ground contact phase, compared with the foot contact, the wheel can move along the rolling direction of the wheel, so the unilateral constraint, friction cone constraint and rolling constraint are expressed as follows:

[0088]

[0089] (x normal ×n)·n=0

[0090]

[0091] in, is the elevation map h terrain (x,y) Local surface normal in world coordinate system W; friction cone constraint limits ground reaction force Stay at the friction coefficient μ C The Coulomb friction cone C(n,μ C ), the trunk direction x normal Approximately the direction in which the wheels roll.

[0092] S1202: Based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, a virtual joint angle is introduced. Through rotation transformation, the world coordinate system is always aligned with the contact coordinate system, and a single-wheel leg kinematic model including the wheel-ground contact point is constructed.

[0093] Since the optimized trajectory y * (t) does not contain key joint information, and uses wheel-foot inverse kinematics to calculate the leg joint position trajectory q j *(t) = [q1(t)q2(t)q3(t)], the corresponding velocity trajectory And the velocity trajectory of the wheel joint ω * (t)∈R 3 Since the wheel-ground contact point changes dynamically when the robot rolls, the traditional Denavit–Hartenberg kinematic modeling method cannot be directly used to establish the kinematic model of a wheeled robot with a wheel-ground contact point.

[0094] In order to solve the above problem, a virtual joint angle q4 is introduced. Through rotation transformation, the coordinate system 4 is always aligned with the contact coordinate system C, so that the kinematic model of a single wheel leg including the wheel-ground contact point is established, as shown in Figure 2 Leg joint information It can be directly obtained through the classic leg inverse kinematics, and the wheel joint velocity ω(t)∈R 3 Then it is associated with q4 and is affected by the other joint angles q2 and q3 as well as the roll angle α and pitch angle β of the trunk. Since the change of α relative to β is negligible, α is approximated to zero, which makes q4 = -β-q2-q3. Under the assumption of no wheel slip, the wheel speed ω and the end effector speed v are established. E The relationship between.

[0095] The single-wheel leg kinematic model including the wheel-ground contact point represents the wheel speed ω and the end effector speed v E The relationship between them is expressed as:

[0096]

[0097] Where J(q)∈R 3×3 is the Jacobian matrix of the leg; q = [q1, q2, q3] represents the leg joint angle vector; q1, q2, q3 represent the first leg joint angle, the second leg joint angle and the third leg joint angle respectively; represents the time derivative of the leg joint angle vector; r 4 is the distance between the wheel-ground contact point and the wheel center; is the rotation matrix of the wheel center coordinate system relative to the body coordinate system.

[0098] S1203: Inverse kinematics solution is performed on the single-wheel leg kinematic model including the wheel-ground contact point to obtain position information of the wheel leg and the body, and then combined with the expected foot-end contact force, motion control of the four-wheeled leg robot is performed.

[0099] Figure 3-Figure 5 The kinematic diagram of the wheeled motion mode, the curves of the leg joint angles and the speed of each wheel are given. Figure 6-Figure 8The wheel-foot complex motion diagram, leg joint angle curve and wheel speed curves in trot gait are given; Figure 9-11 The wheel-foot complex motion diagram, leg joint angle curve and wheel speed curves in bound gait are given.

[0100] Embodiment 2

[0101] Fig.13 Schematic diagram of the controller structure of an embodiment of the present invention. Fig.13 As shown, a controller is provided, which includes:

[0102] The whole body optimal trajectory solving module 1301 is used to construct a trajectory optimization objective function based on the motion dynamics model of the four-wheeled legged robot, and solve the whole body optimal trajectory of the four-wheeled legged robot within the motion cycle under the constraints set by the four-wheeled legged robot; wherein the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the torso coordinate system and the velocity of the wheel-ground contact point in the world coordinate system;

[0103] A kinematic model building module 1302 is used to introduce a virtual joint angle based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, and to make the world coordinate system always aligned with the contact coordinate system through rotation transformation, so as to build a single-wheel leg kinematic model including the wheel-ground contact point;

[0104] The inverse kinematics solution and control module 1303 is used to perform inverse kinematics solution on the single-wheel leg kinematic model including the wheel-ground contact point, obtain the position information of the wheel leg and the body, and then control the motion of the four-wheeled leg robot in combination with the expected foot-end contact force.

[0105] It should be noted here that the various modules in this embodiment correspond one-to-one to the various steps in the above-mentioned embodiment 1, and the specific implementation process is the same, which will not be described in detail here.

[0106] Embodiment 3

[0107] The present embodiment provides a four-wheeled legged robot system, which includes a four-wheeled legged robot and a controller, wherein the controller is used to control the movement of the four-wheeled legged robot.

[0108] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A trajectory optimization and motion generation method for a quadruped robot, characterized in that: include: Based on the kinematics model of the four-wheeled legged robot, a trajectory optimization objective function is constructed, and the optimal trajectory of the whole body of the four-wheeled legged robot within the motion cycle is solved under the set constraints of the four-wheeled legged robot; wherein the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the trunk coordinate system and the velocity of the wheel-ground contact point in the world coordinate system; Based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, a virtual joint angle is introduced, and the world coordinate system is always aligned with the contact coordinate system through rotation transformation, thus constructing a single-wheel leg kinematic model including the wheel-ground contact point. The inverse kinematics of the single-wheel leg kinematic model including the wheel-ground contact point is solved to obtain the position information of the wheel leg and the body, and then combined with the expected foot end contact force, the motion of the four-wheeled leg robot is controlled.

2. The method for trajectory optimization and motion generation of a quadruped robot according to claim 1, characterized in that: The single-wheel leg kinematic model including the wheel-ground contact point represents the wheel speed ω and the end effector speed v E The relationship between them can be expressed as: Where J(q)∈R 3×3 is the Jacobian matrix of the leg; q = [q1, q2, q3] represents the leg joint angle vector; q1, q2, q3 represent the first leg joint angle, the second leg joint angle and the third leg joint angle respectively; represents the time derivative of the leg joint angle vector; r 4 is the distance between the wheel-ground contact point and the wheel center; is the rotation matrix of the wheel center coordinate system relative to the body coordinate system.

3. The trajectory optimization and motion generation method of a quadruped robot according to claim 1, characterized in that: The expression of the trajectory optimization objective function f(y(t)) is: x torso (t)=[θ T p T oh T v T ] T Among them, y(t) is the optimization variable, which includes the four-wheeled robot in the motion period T TO The whole body trajectory on x torso (t)∈R 12 Contains the torso's heading θ∈R in Euler angles in the world coordinate system 3 and the centroid position p∈R 3 , the angular velocity of the center of mass in the torso coordinate system ω∈R 3 With linear velocity v∈R 3 , x leg,i (t) contains the position of the wheel-ground contact point in the world coordinate system speed and contact force Q is the semi-positive Hessian matrix, the state vector error n x =[1,0,0] T Choose a vector for the x direction, λ c ∈(0,1) is the end contact state; the superscript T represents the transpose of the vector.

4. The method for trajectory optimization and motion generation of a quadruped robot as claimed in claim 3, characterized in that: The constraint conditions set for the quadruped robot include: initial and final state constraints, kinematic constraints, dynamic constraints and wheel-ground contact constraints.

5. The method for trajectory optimization and motion generation of a quadruped robot as claimed in claim 4, characterized in that: The kinematic constraints are determined by the allowed joint angle range and ensure that there is no interference between the front and rear wheels; r Ei ∈K i (θ,p,R) Among them, K i represents the predefined working space of the end effector, and R represents the radius of the wheel.

6. The method for trajectory optimization and motion generation of a quadruped robot as claimed in claim 4, characterized in that: The dynamic constraints are characterized by a single rigid body dynamics model: in, Represents the angular velocity of the torso in Euler angles in the world coordinate system; Represents the velocity of the center of mass of the torso in the world coordinate system; represents the angular acceleration of the center of mass in the torso coordinate system; represents the linear acceleration of the center of mass in the torso coordinate system; represents the rotation matrix, which maps the components of the vector from the torso coordinate system B to the world coordinate system W; T(θ) is the transformation matrix that maps the angular velocity of the torso to the Euler angular velocity in the world coordinate system; I W is the inertia tensor of the torso; m is the total mass of the robot; g(θ) is the gravitational acceleration in the torso coordinate system.

7. The method for trajectory optimization and motion generation of a quadruped robot as claimed in claim 4, characterized in that: The wheel-ground contact constraints include: (x normal ×n)·n=0 in, is the elevation map h terrain (x,y) Local surface normal in world coordinate system W; friction cone constraint limits ground reaction force Stay at the friction coefficient μ C The Coulomb friction cone C(n,μ C ), the trunk direction x normal Approximately the direction in which the wheels roll.

8. The method for trajectory optimization and motion generation of a quadruped robot as claimed in claim 4, characterized in that: The initial and final state constraints are artificially set in advance.

9. A controller, characterized in that: include: The whole body optimal trajectory solving module is used to construct a trajectory optimization objective function based on the motion dynamics model of the four-wheeled legged robot, and solve the whole body optimal trajectory of the four-wheeled legged robot within the motion cycle under the constraints set by the four-wheeled legged robot; wherein the trajectory optimization objective function is a state vector error related function, and the state vector error is the difference between the linear velocity of the center of mass of the torso coordinate system and the velocity of the wheel-ground contact point in the world coordinate system; The kinematic model building module is used to introduce a virtual joint angle based on the optimal whole-body trajectory of the quadruped robot during the motion cycle, and to align the world coordinate system with the contact coordinate system through rotation transformation, so as to build a single-wheel leg kinematic model including the wheel-ground contact point; The inverse kinematics solution and control module is used to perform inverse kinematics solution on the single-wheel leg kinematic model including the wheel-ground contact point, obtain the position information of the wheel leg and the body, and then combine the expected foot end contact force to control the motion of the four-wheeled leg robot.

10. A four-wheeled robot system, characterized in that: It comprises a four-wheeled legged robot and a controller as claimed in claim 9, wherein the controller is used to control the movement of the four-wheeled legged robot.

Images