Wheel type biped robot whole machine coupling control system and method

Abstract

The application provides a wheeled biped robot whole machine coupling control system and method, and belongs to the technical field of robot control, the system comprises a wheeled biped robot body, a sensor module, an embedded controller and a control software module; the wheeled biped robot body comprises a fuselage, left and right parallel five-link leg mechanisms, hub motors and joint motors, the left and right parallel five-link leg mechanisms are connected on both sides of the fuselage, and the joint motors and the hub motors are arranged in the left and right parallel five-link leg mechanisms; the sensor module comprises encoders and inertial measurement units (IMUs) for measuring the motion state and attitude information of the robot; the embedded controller executes a control algorithm; and the control software module comprises a state estimation module, a balance control module, an attitude control module and an altitude control module.The wheeled biped robot whole machine coupling control system and method solve the problem of insufficient joint motor and hub motor collaborative control in the prior art.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a whole-machine coupling control system and method for a wheeled bipedal robot. Background Technology

[0002] In the field of robot control technology, the design and control of mobile robots has always been a research hotspot. Currently, mobile robots are mainly divided into two categories: wheeled robots and legged robots. Wheeled robots, with their advantages of high motion efficiency and simple control, perform well in flat terrain; however, their adaptability decreases significantly when faced with complex and varied terrain. Conversely, although legged robots have strong terrain-crossing capabilities and can walk stably in complex environments, their high energy consumption and complex control algorithms limit their widespread application.

[0003] To combine the advantages of both, wheeled bipedal robots (WBRs) have gradually gained attention as an innovative solution. However, existing control technologies for wheeled bipedal robots are mostly based on an inverted pendulum model, decoupling the wheels from the superstructure or leg joints, and relying on hub motors to independently complete balance and motion control. This control method struggles to accurately describe the system's dynamic characteristics when the robot structure becomes more complex, the center of gravity changes with posture, or parallel leg mechanisms are present. This results in insufficient participation of the joint motors in balance control, insufficient system robustness, and susceptibility to drifting under posture changes or external disturbances. Furthermore, the hub motors experience high loads and energy consumption under acceleration and impact conditions. In addition, some methods based on model predictive control (MPC) or inverse dynamics are computationally complex and difficult to implement in high-frequency real-time operation on embedded controllers. Summary of the Invention

[0004] The purpose of this invention is to provide a whole-machine coupling control system and method for a wheeled bipedal robot, which solves the problem of insufficient coordinated control between joint motors and hub motors in the prior art.

[0005] To achieve the above objectives, the present invention provides a wheeled bipedal robot whole-machine coupling control system, the system including a wheeled bipedal robot body, a sensor module, an embedded controller and a control software module;

[0006] The wheeled bipedal robot body includes a body, a left and right parallel five-bar linkage mechanism, a wheel hub motor and a joint motor. The left and right parallel five-bar linkage mechanism is connected to both sides of the body, and the left and right parallel five-bar linkage mechanism is equipped with a joint motor and a wheel hub motor.

[0007] The sensor module includes an encoder and an inertial measurement unit (IMU) to measure the robot's motion state and attitude information;

[0008] The embedded controller executes the control algorithm;

[0009] The control software module includes a state estimation module, a balance control module, an attitude control module, and a height control module.

[0010] Preferably, the five-link leg mechanism includes a first joint, a second joint, a third joint, a fourth joint, and a fifth joint. A torso link is provided between the first and second joints, a thigh link is provided between the first and third joints, a lower leg link is provided between the third and fifth joints, and links are provided between the second and fourth joints and between the fourth and fifth joints. A hub is provided on the output shaft of the fifth joint, and the hub is connected to the output shaft of a hub motor. Joint motors are connected to the first and second joints, and a universal joint is also provided above the five-link leg mechanism.

[0011] Preferably, the control software module also includes a whole-machine dynamics model establishment unit, a parallel mechanism kinematic mapping unit, a linearization and state-space modeling unit, an LQR controller design unit, and an online control and execution unit. The online control and execution unit introduces yaw and roll cascade control loops. The yaw control inner loop output is the torque command applied to the hub motor, and the roll control inner loop output is used to update the equilibrium point.

[0012] A method for overall coupling control of a wheeled bipedal robot includes the following steps:

[0013] S1. Establish the overall dynamic model: The robot is equivalent to a system with a wheeled mobile base and a variable-length inverted pendulum. Displacement, tilt angle, body attitude angle and height are selected as generalized coordinates. The overall dynamic equation is established based on the Lagrange method.

[0014] S2. Parallel Mechanism Dynamics Mapping: Based on the geometric constraints of the parallel five-bar linkage, the fuselage attitude angle and height are mapped to the joint space to achieve a two-way mapping between the joint angle and the overall state of the machine.

[0015] S3. Linearization to state-space modeling: Given motion speed, attitude and height reference values, the nonlinear dynamic model is linearized to obtain the state-space expression;

[0016] S4, LQR controller design: Construct a state vector containing motion speed, tilt angle, attitude angle and altitude, and obtain the full state feedback gain matrix by setting the weight matrix and solving the algebraic Riccati equation;

[0017] S5. Online Control and Execution: Based on feedback from the sensor module, the control quantity is calculated and coordinated to the hub motor and joint motor to achieve synchronous control of balance, motion and attitude height. Yaw and roll cascade control links are introduced to enhance anti-disturbance capability.

[0018] Preferably, the overall dynamic equation expression is as follows:

[0019] ;

[0020] in, For time, For the system's generalized coordinates, For generalized forces acting on the corresponding degrees of freedom, For the first A generalized coordinate;

[0021] ;

[0022] in, and These represent the system's kinetic energy and potential energy, respectively.

[0023] Preferably, the dynamic mapping expression for the parallel mechanism is:

[0024] ;

[0025] in, Let be the displacement along the X direction. The angle of the first joint, For the angle of the second joint, For the angle of the third joint, The angle of the fourth joint, The angle of inclination. For attitude angle, Let OE be the equivalent length of the inverted pendulum. Let be the Jacobian matrix mapped to generalized coordinates.

[0026] Preferably, the state-space expression is:

[0027] ;

[0028] ;

[0029] ;

[0030] in, The velocity is along the X direction. The tilt angular velocity, For attitude angular velocity, Let OE be the rate of change of the length of the equivalent inverted pendulum as a function of time. The horizontal thrust acting on the trolley The motor torque applied to joint A, The torque applied to joint B is the motor torque.

[0031] Preferably, the expression for the full-state feedback gain matrix is:

[0032] ;

[0033] in, For full-state feedback gain, To achieve equilibrium, The various control variables that need to be applied to achieve equilibrium.

[0034] Therefore, the present invention employs the above-mentioned overall coupling control system and method for a wheeled bipedal robot, and the technical effects are as follows:

[0035] 1. This invention introduces whole-machine coupled dynamics modeling, enabling joint motors and hub motors to work together and participate in the robot's balance and motion control, significantly improving the robot's motion coordination and stability.

[0036] 2. This invention effectively reduces the peak torque of the hub motor by using whole-machine coupled dynamics modeling and LQR control method, which enhances the system's ability to resist disturbances under complex terrain and external impacts, enabling the robot to maintain stable operation in more unstable environments.

[0037] 3. By optimizing the control algorithm, this invention significantly reduces the peak torque of the hub motor, thereby reducing the energy consumption of the motor and improving the overall energy efficiency of the robot.

[0038] 4. This invention allows for independent setting of attitude angle and height, enabling the robot to adapt to various task conditions, improving the robot's flexibility and practicality. The system's good robustness to complex terrain and external impacts further enhances the robot's environmental adaptability. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall structure of Ghost, a wheeled bipedal robot according to the present invention.

[0040] Figure 2 This is a schematic diagram of a template model of a wheeled bipedal robot in a two-dimensional plane according to the present invention; (a) is the generalized coordinates of the model; (b) is the kinematics of the parallel five-bar linkage.

[0041] Figure 3 This is a block diagram of a cascaded control system for a wheeled bipedal robot in multiple domains according to the present invention.

[0042] Figure 4 This is a schematic diagram of the overall control architecture of a wheeled bipedal robot according to the present invention.

[0043] Figure Labels

[0044] 1. First joint; 2. Second joint; 3. Third joint; 4. Fourth joint; 5. Fifth joint; 6. Thigh link; 7. Torso link; 8. Lower leg link; 9. Wheel hub; 10. Joint motor; 11. Wheel hub motor; 12. Universal joint. Detailed Implementation

[0045] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0046] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0047] Example 1

[0048] like Figure 1 As shown, the present invention provides a whole-machine coupling control system for a wheeled bipedal robot, the system including a wheeled bipedal robot body, a sensor module, an embedded controller and a control software module;

[0049] The wheeled bipedal robot body includes a body, a left and right parallel five-bar linkage mechanism, a hub motor 11 and a joint motor 10. The left and right parallel five-bar linkage mechanism is connected to both sides of the body, and the joint motor 10 and the hub motor 11 are installed in the left and right parallel five-bar linkage mechanism.

[0050] The sensor module includes an encoder and an inertial measurement unit (IMU) to measure the robot's motion state and attitude information;

[0051] The embedded controller executes the control algorithm;

[0052] The control software module includes a state estimation module, a balance control module, an attitude control module, and a height control module.

[0053] The wheeled bipedal robot consists of a chassis and a gimbal system with two degrees of freedom. The chassis has two independent leg structures mounted on each side of the body, each leg employing a parallel five-bar linkage design. Each leg structure is equipped with two articulated motors 10 for adjusting and maintaining the robot's attitude angle and height, and hub motors 11 at their ends for driving the robot's movement and achieving dynamic balance control. The five-link leg mechanism includes a first joint 1, a second joint 2, a third joint 3, a fourth joint 4, and a fifth joint 5. A torso link 7 is provided between the first joint 1 and the second joint 2. A thigh link 6 is provided between the first joint 1 and the third joint 3. A lower leg link 8 is provided between the third joint 3 and the fifth joint 5. Links are provided between the second joint 2 and the fourth joint 4, and between the fourth joint 4 and the fifth joint 5. A hub 9 is provided on the output shaft of the fifth joint 5. The hub 9 is connected to the output shaft of the hub motor 11. Joint motors 10 are connected to the first joint 1 and the second joint 2. A universal joint 12 is also provided above the five-link leg mechanism.

[0054] The first joint 1 and the second joint 2 are equipped with joint motors 10, mainly used for attitude adjustment and height control of the robot body links; the third joint 3 and the fourth joint 4 are passive joints and are not equipped with drive devices. The wheel drive device is mounted at the fifth joint 5 via a hub motor 11, used to drive the robot's movement and maintain its dynamic balance. The gimbal structure is set as an additional mechanism, and it is assumed in this paper to remain fixed relative to the robot body links.

[0055] The control software module also includes a whole-machine dynamics model establishment unit, a parallel mechanism kinematics mapping unit, a linearization and state-space modeling unit, an LQR controller design unit, and an online control and execution unit. The online control and execution unit introduces yaw and roll cascade control loops. The yaw control inner loop output is the torque command applied to the hub motor 11, and the roll control inner loop output is used to update the balance point.

[0056] To simplify the modeling process, this paper makes the following assumptions: the gimbal structure remains fixed relative to the chassis; the rotational inertia of all links except the torso link AB is ignored; the wheels always satisfy the pure rolling condition with no slippage; furthermore, it is assumed that the robot's center of mass O is located at the midpoint of the torso link AB. These assumptions may introduce some errors in the dynamic modeling, but these can be compensated for by introducing additional PID feedback control.

[0057] S1. Establish the overall dynamic model: The robot is equivalent to a system with a wheeled mobile base and a variable-length inverted pendulum. Displacement, tilt angle, body attitude angle and height are selected as generalized coordinates. The overall dynamic equation is established based on the Lagrange method.

[0058] S2. Parallel Mechanism Dynamics Mapping: Based on the geometric constraints of the parallel five-bar linkage, the fuselage attitude angle and height are mapped to the joint space to achieve a two-way mapping between the joint angle and the overall state of the machine.

[0059] S3. Linearization to state-space modeling: Given motion speed, attitude and height reference values, the nonlinear dynamic model is linearized to obtain the state-space expression;

[0060] S4, LQR controller design: Construct a state vector containing motion speed, tilt angle, attitude angle and altitude, and obtain the full state feedback gain matrix by setting the weight matrix and solving the algebraic Riccati equation;

[0061] S5. Online Control and Execution: Based on feedback from the sensor module, the control quantity is calculated and coordinated to the hub motor and joint motor to achieve synchronous control of balance, motion and attitude height. Yaw and roll cascade control links are introduced to enhance anti-disturbance capability.

[0062] The system dynamics model is based on the Lagrange equations, and its expression is:

[0063] ;

[0064] in, For time, For the system's generalized coordinates, For generalized forces acting on the corresponding degrees of freedom, For the first A generalized coordinate;

[0065] The Lagrange function is defined as

[0066] ;

[0067] in, and These represent the system's kinetic energy and potential energy, respectively.

[0068] like Figure 2 As shown, the robot Ghost can be simplified in a two-dimensional plane as a system with four degrees of freedom, represented by four independent generalized coordinates: This represents the displacement of the trolley along the X direction; The tilt angle of the equivalent inverted pendulum (represented by the dashed line); The attitude angle of link AB; and The length of the pendulum is also referred to as the fuselage height in this paper. Although the choice of generalized coordinates is not unique, we chose these variables to simplify the derivation of the system's Lagrangian function to the greatest extent possible.

[0069] ;

[0070] like Figure 2 As shown in (a), under the assumption of pure rolling constraints, the wheel is transformed into a wheel driven by thrust. The driven car. The position of the car along the X direction. This is the first generalized coordinate system. The dashed line segment OE represents the robot's equivalent inverted pendulum, and its length... Variable. Length of the equivalent inverted pendulum. and the tilt angle relative to the direction of gravity These are all generalized coordinates. The last generalized coordinate is the attitude angle. It represents the angle between the connecting rod AB and the horizontal line.

[0071] like Figure 2 As shown in (b), kinematics can be obtained by projecting all links onto and The direction is used to deduce. and It is obtained directly from the encoder of the joint motor, and and Then in the given and Solve under the given conditions.

[0072] The kinetic and potential energy of the system can be expressed as:

[0073] ;

[0074] ;

[0075] in, Indicates the mass of the equivalent car. and These represent the mass and moment of inertia of the robot's connecting link, which includes the robot's body and gimbal structure. The position of the vehicle along the X-axis is the first generalized coordinate. The dashed segment OE represents the robot's equivalent inverted pendulum. Its length is variable. The length of the equivalent inverted pendulum and its angle of inclination relative to the direction of gravity are also factors. These are all generalized coordinates. The last generalized coordinate is the attitude angle. It represents the angle between the connecting rod AB and the horizontal line.

[0076] Generalized force Defined as:

[0077] ;

[0078] in, This represents the external force acting on the system. This represents the displacement corresponding to the external force. This refers to the quantity of external forces. In this system, external forces include: ;in, The horizontal thrust acting on the equivalent trolley, and This refers to the joint torque generated by the motor in the leg joint.

[0079] Assuming there is no slippage between the wheel and the ground, the horizontal thrust With the output torque of the hub motor The following relationship exists between them:

[0080] ;

[0081] in, The radius is the wheel radius.

[0082] like Figure 2 As shown in (b), the link is projected onto and Direction:

[0083] ;

[0084] ;

[0085] in, , , These are the lengths of the torso link, thigh link, and lower leg link, respectively. , , , These are the angles of the first joint, the second joint, the third joint, and the fourth joint, respectively.

[0086] The calculation method is as follows:

[0087] ;

[0088] Therefore, the Jacobian matrix It can be represented as:

[0089] ;

[0090] ;

[0091] The dynamic equations of the system can be explicitly written in the following form:

[0092] ;

[0093] ;

[0094] ;

[0095] The robot's control is constructed as an LQR problem, and the dynamics function can therefore be transformed into the following state-space form:

[0096] ;

[0097] ;

[0098] ;

[0099] variable The following form represents the equilibrium point:

[0100] ;

[0101] ;

[0102] in, , and The constraints imposed by the parallel five-bar linkage of the legs must be satisfied. and It can be uniquely determined by the following equation:

[0103] ;

[0104] After linearization, the state-space expression can be represented in the following form:

[0105] ;

[0106] in, , , , .

[0107] For a given quadratic cost function, the symmetric and positive (semi-)definite matrix , By solving the algebraic Riccati equation, the control variables are... It can be given by the following formula:

[0108] ;

[0109] matrix and It depends on a specific equilibrium point. To achieve online control, the state feedback gain... It is obtained by interpolation from pre-calculated values.

[0110] The aforementioned controller is designed for a two-dimensional plane and does not include control over yaw and roll. To improve disturbance response performance, a cascaded control method is used for yaw and roll, such as... Figure 3 As shown in the diagram, both the master and slave controllers in this method are proportional controllers to simplify the design. The outer loop generates an angular velocity reference based on a given reference and the yaw or roll angle, which serves as the setpoint for the inner loop. The inner loop then outputs a control quantity based on the error between the reference and the actual yaw or roll angular velocity, which is applied to the robot. In yaw control, the inner loop output is a torque command applied to the wheel motors for directional control; in roll control, the inner loop output is used to update the balance point, thereby adjusting the ground support force of the legs and controlling the roll angle of the robot.

[0111] like Figure 3 As shown, the structure includes internal feedback and external feedback. and These are the system's reference input and output, respectively. The main controller features external feedback. Provide setpoints for internal feedback, while the internal feedback slave controller The output acts as a control variable on the system. Both the master and slave controllers could be PID controllers, but a proportional controller was chosen in this work to simplify the design. Cascaded control is used for yaw and roll because when the system... Disturbed At this time, its performance is superior to that of single PID control. The external feedback reference is provided by remote control, and the feedback from the external feedback and internal feedback is the yaw or roll angle and angular velocity of the aircraft, respectively.

[0112] The overall control architecture of the robot is as follows Figure 4 As shown. The controller updates the feedback gain and equilibrium point through interpolation based on the desired speed, attitude angle, and body height. Torque commands for all motors are generated through state feedback and cascaded controllers. The robot also collects data through proprioceptive measurements to estimate leg states that cannot be directly measured, such as... Figure 2 (b) , The proposed controller runs on an STM32F407 microcontroller (MCU) at a control frequency of 1kHz.

[0113] The user sets the robot's desired speed, posture, and height via the remote controller (blue). Based on the user's commands, the controller first interpolates the feedback gain from values ​​pre-calculated using the LQR method. And update the balance point. Then, the state feedback controller applies torque commands to the joint and wheel motors to control the robot's motion, balance, and attitude (yellow and gray). The output of the roll cascade controller is incorporated into the balance point update, indirectly affecting the torque of the joint motors through state feedback, while the output of the yaw cascade controller directly acts on the wheel motors to achieve steering (green). State estimation is performed using only proprioceptive measurements from the motor encoders and inertial measurement units (IMUs) to provide feedback to the controller (red). The proposed overall controller operates at a control frequency of 1 kHz.

[0114] Therefore, this invention adopts the above-mentioned whole-machine coupling control system and method for a wheeled bipedal robot. By establishing a whole-machine dynamic model including the parallel mechanism of wheels and legs and the body, a unified generalized coordinate system is introduced to construct a linearized state space model. The linear quadratic regulation (LQR) method is used to realize the cooperative control of multiple actuators. This not only solves the problem of insufficient cooperative control between joint motors and hub motors in the prior art, but also significantly reduces the peak torque of the hub motors, improves the robustness and stability of the system under complex terrain and external impacts, and ensures the high-frequency real-time operation of the control algorithm on a low-computing-power embedded controller, laying a solid foundation for the widespread application of wheeled bipedal robots.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for overall coupling control of a wheeled bipedal robot, comprising applying an overall coupling control system for a wheeled bipedal robot to the overall coupling control of the robot, characterized in that, The system includes a wheeled bipedal robot body, a sensor module, an embedded controller, and a control software module; The wheeled bipedal robot body includes a body, a left and right parallel five-bar linkage mechanism, a wheel hub motor and a joint motor. The left and right parallel five-bar linkage mechanism is connected to both sides of the body, and the left and right parallel five-bar linkage mechanism is equipped with a joint motor and a wheel hub motor. The sensor module includes an encoder and an inertial measurement unit (IMU) to measure the robot's motion state and attitude information; The embedded controller executes the control algorithm; The control software module includes a state estimation module, a balance control module, an attitude control module, and a height control module; The method includes the following steps: S1. Establish the overall dynamic model: The robot is equivalent to a system with a wheeled mobile base and a variable-length inverted pendulum. Displacement, tilt angle, body attitude angle and height are selected as generalized coordinates. The overall dynamic equation is established based on the Lagrange method. S2. Parallel Mechanism Dynamics Mapping: Based on the geometric constraints of the parallel five-bar linkage, the fuselage attitude angle and height are mapped to the joint space to achieve a two-way mapping between the joint angle and the overall state of the machine. S3. Linearization to state-space modeling: Given motion speed, attitude and height reference values, the nonlinear dynamic model is linearized to obtain the state-space expression; S4, LQR controller design: Construct a state vector containing motion speed, tilt angle, attitude angle and altitude, and obtain the full state feedback gain matrix by setting the weight matrix and solving the algebraic Riccati equation; S5. Online Control and Execution: Based on feedback from the sensor module, the control quantity is calculated and coordinated to the hub motor and joint motor to achieve synchronous control of balance, motion and attitude height. Yaw and roll cascade control links are introduced to enhance anti-disturbance capability.

2. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The five-link leg mechanism includes a first joint, a second joint, a third joint, a fourth joint, and a fifth joint. A torso link is provided between the first and second joints, a thigh link is provided between the first and third joints, a lower leg link is provided between the third and fifth joints, and links are provided between the second and fourth joints, as well as between the fourth and fifth joints. A hub is provided on the output shaft of the fifth joint, and the hub is connected to the output shaft of a hub motor. Joint motors are connected to the first and second joints. A universal joint is also provided on the top of the five-link leg mechanism.

3. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The control software module also includes a whole-machine dynamics model establishment unit, a parallel mechanism kinematic mapping unit, a linearization and state-space modeling unit, an LQR controller design unit, and an online control and execution unit. The online control and execution unit introduces yaw and roll cascaded control loops. The yaw control inner loop output is the torque command applied to the hub motor, and the roll control inner loop output is used to update the equilibrium point.

4. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The equation of motion for the whole machine is: ； in, For time, For the system's generalized coordinates, For generalized forces acting on the corresponding degrees of freedom, For the first A generalized coordinate; ； in, and These represent the system's kinetic energy and potential energy, respectively.

5. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The expression for mapping the dynamics of a parallel mechanism to a generalized coordinate system is: ； in, Let be the displacement along the X direction. The angle of the first joint, For the angle of the second joint, For the angle of the third joint, The angle of the fourth joint, The angle of inclination. For attitude angle, Let OE be the equivalent length of the inverted pendulum. Let be the Jacobian matrix mapped to generalized coordinates.

6. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The state-space expression of the dynamic equation is: ； ； ； in, The velocity is along the X direction. The tilt angular velocity, For attitude angular velocity, Let OE be the rate of change of the length of the equivalent inverted pendulum as a function of time. The horizontal thrust acting on the trolley The motor torque applied to joint A, The torque applied to joint B is the motor torque.

7. The whole-machine coupling control method for a wheeled bipedal robot according to claim 1, characterized in that, The expression for the full-state feedback gain matrix is: ； in, For full-state feedback gain, To achieve equilibrium, The various control variables that need to be applied to achieve equilibrium.

Images