A pose control system and control method for an intelligent breaking robot

Abstract

The application discloses a kind of pose control systems and control method for intelligent breaking robot;Including arrangement in robot hydraulic module, pose signal acquisition module and pose control module;Hydraulic module is responsible for driving breaking robot action;Pose signal acquisition module is used to collect the rotation angle of each joint of breaking robot, the displacement of hydraulic cylinder and the angle and rotational speed of each motor output shaft;Pose control module includes pose calculation module, control instruction module;Pose calculation module is used to solve the information collected, pose monitoring;Control instruction module is used for pose adjustment control;The method fuses virtual trajectory with actual hydraulic system pose feedback, error calculation is carried out through pose control module, and control instruction is generated in real time in pose control module, improve the ability of breaking hammer mechanical arm pose autonomous adjustment;Realize the adaptive optimization and dynamic compensation of mechanical arm trajectory, to improve the end pose control precision of mechanical arm and breaking efficiency.

Description

Technical Field

[0001] This invention relates to the field of control technology for metal mining engineering machinery, and in particular to a posture control system and control method for an intelligent crushing robot. Background Technology

[0002] Intelligent crushing robots are playing an increasingly crucial role in mining, tunnel excavation, and building demolition. They typically consist of a rotating base, a boom, a forearm, a breaker hammer, and a hydraulic system. The boom and forearm extend via hydraulic cylinders, respectively. The boom is mounted on the rotating base, which houses a rotary motor that drives the base to rotate omnidirectionally in the horizontal plane. The posture control hydraulic system is a key component for realizing the robot's intelligence. By adding electro-hydraulic proportional control valves and posture sensing devices, a closed-loop feedback mechanism is established.

[0003] Currently, research on posture control technology for crushing robots in China is insufficient. Existing robotic arm crushing systems typically rely on manual control or preset trajectories for operation, and their posture control accuracy is insufficient to meet the requirements of complex unstructured environments. Furthermore, existing robotic arm posture control methods often depend on single sensors or fixed control algorithms, lacking comprehensive consideration of multi-sensor information fusion and real-time error compensation. This leads to problems such as target deviation or trajectory jitter during crushing operations, resulting in low crushing efficiency.

[0004] Based on the above problems, there is an urgent need to provide an intelligent crushing robot posture control system and method that can integrate virtual trajectory prediction and multi-sensor feedback, and combine hydraulic system characteristics for closed-loop control. This system can autonomously control the crushing robot based on posture errors during crushing operations, thereby improving the operation accuracy and efficiency of the crushing robot. Summary of the Invention

[0005] To address the problems of low posture control accuracy in existing crushing robots, and the tendency for the robotic arm to deviate from its target or jitter its trajectory during crushing operations, this application provides a posture control system and method for an intelligent crushing robot.

[0006] Firstly, this application provides a posture control system for an intelligent crushing robot, employing the following technical solution:

[0007] A pose control system for an intelligent crushing robot includes a hydraulic module, a pose signal acquisition module, and a pose control module arranged in the robot.

[0008] The hydraulic module includes a main circuit supplied by a high-pressure pump and three parallel branch circuits; one branch circuit consists of a three-position four-way electro-hydraulic valve 1 connected in series with the boom hydraulic cylinder, another branch circuit consists of a three-position four-way electro-hydraulic valve 2 connected in series with the forearm hydraulic cylinder, and the third branch circuit consists of a three-position four-way electro-hydraulic valve 3 connected in series with the rotary motor.

[0009] The pose signal acquisition module is used to acquire the rotation angle of each joint of the crushing robot, the displacement of the hydraulic cylinder, and the angle and speed of each motor output shaft.

[0010] The pose control module includes a pose calculation module and a control command module. The pose calculation module is used to solve the angle information collected by the pose signal acquisition module and to monitor the pose of the crushing robot. The control command module is used to perform pose adjustment control based on the current pose monitoring information.

[0011] Optionally, a booster oil tank is provided on the main circuit, and a hydraulic control check valve is provided between the booster oil tank and the main circuit; three branch circuits are connected in parallel with overflow valves, and the overflow valves are connected in series with the booster oil tank.

[0012] The pose signal acquisition module includes tilt sensor one, displacement sensor one, tilt sensor two, displacement sensor two, rotary encoder, and lidar. Tilt sensor one and tilt sensor two are respectively installed at the joints of the upper arm and forearm to monitor the rotation angle of each joint. Displacement sensor one and displacement sensor two are installed inside the hydraulic cylinders of the upper arm and forearm to measure the displacement of the hydraulic cylinders. The rotary encoder is installed on the output shaft of the rotary motor to measure the angle of the rotary motor output shaft. The lidar is installed on the top of the crushing robot's working platform to acquire point cloud data of the crushing working area.

[0013] Secondly, this application provides a pose control method for an intelligent crushing robot, employing the following technical solution:

[0014] A pose control method for an intelligent crushing robot, employing a pose control hydraulic system for an intelligent crushing robot as described in the first aspect, includes the following:

[0015] Collect and process relevant pose data of the intelligent crushing robot;

[0016] Calculation of the actual pose of the end effector based on the DH parameter method;

[0017] Virtual trajectories are generated based on Dynamic Motion Element (DMP).

[0018] The virtual trajectory is compared with the actual pose, and the pose error is calculated.

[0019] The error is converted into joint space adjustment amount, and the robot trajectory is optimized by adjusting the adjustment.

[0020] Optionally, the relevant pose data includes: the attitude angle signals of the upper arm and lower arm of the crushing robot, the position signals of the hydraulic cylinders of the upper arm and lower arm, the rotation pose signal of the rotary motor, and the point cloud data of the crushing operation area.

[0021] Optionally, the forward kinematics matrix of the robotic arm is constructed based on the DH parameter method, and the end effector pose is calculated. The specific details are as follows:

[0022] DH transformation matrix:

[0023] ;

[0024] In the formula, For joint angle, This is the link offset. The length of the link. Linkage torsion angle; Substituting the slewing angle, boom angle, and forearm angle into the above formula, we obtain the following:

[0025] End-effector pose matrix:

[0026] ;

[0027] End position coordinates:

[0028] ;

[0029] Euler angle attitude of the end effector:

[0030] Roll angle: ;

[0031] Pitch angle: ;

[0032] Yaw angle: .

[0033] Optionally, the virtual trajectory generated based on Dynamic Motion Element (DMP) is as follows:

[0034] Determine the coordinates of the break point based on point cloud data. And the initial conditions for DMP output: Constructing the DMP dynamic system:

[0035] ;

[0036] In the formula, This indicates the current end-effector pose of the crushing robot. This is represented as the initial pose; It is an inertial parameter, and is a scaling direction; Let be the spatial vector of the joint. and They represent The first and second derivatives, For the first Each joint angular velocity at time t, For the first Each joint Angular acceleration at time t; The target state is the state that the DMP dynamics system will eventually converge to. and For system parameters; The trajectory shape learner is a non-linear function;

[0037] By changing the target state and nonlinear The items are used to adjust the endpoint and shape of the trajectory;

[0038] nonlinear functions The desired trajectory is approximated by a combination of Gaussian functions;

[0039] , ;

[0040] In the formula, The number of Gaussian functions; It is the first One basis function; As weight; Center of Gaussian function The width of the Gaussian function;

[0041] Finally, the process is iterated to generate joint trajectories;

[0042] At each time step Update joint angular acceleration:

[0043] ;

[0044] Update joint angular velocity:

[0045] ;

[0046] Update joint angles:

[0047] ;

[0048] Output joint space virtual trajectory:

[0049] ;

[0050] In the formula, For the first Each joint Angular acceleration at time t; These are inertial parameters; For stiffness gain; For the first The target position of each joint; For damping gain; For force feedback gain; For the first External forces acting on each joint; For the first Each joint angular velocity at time t; For the first Each joint The position at that moment; is the spatial vector of the joint.

[0051] Optional, pose error ; for The actuator end-effector pose at a given moment. for The coordinates of the breakpoint at that moment;

[0052] The error is converted into joint space adjustment amount using the Jacobian matrix. : ;

[0053] In the formula is the generalized inverse of the Jacobian matrix;

[0054] Optimized pose adjustment amount : .

[0055] Optionally, when comparing the virtual trajectory with the actual pose, Kalman filtering is used to fuse the input data and output the end effector pose signal. The pose error is formed by comparing it with the virtual trajectory. Trajectory optimization is performed; specifically, the following steps are included:

[0056] S1. Define the state vector of the end-effector pose as follows: Then in State vector at time step It can be represented as:

[0057] ;

[0058] S2. Establish the basic mathematical equations for Kalman filtering:

[0059] State transition equation: ;

[0060] Observation equation: ;

[0061] In the formula, Here is the state transition matrix. To control the input matrix, The joint angular velocity, For process noise, for The sensor measurement vector at any given time. For the observation matrix, For measuring noise;

[0062] S3, Prediction Step: Make predictions based on the state of the previous time step;

[0063] ;

[0064] ;

[0065] In the formula, Indicates based on Vector estimates at time t The state vector generated after predicting the value at time 1. Indicates the predicted state at the next moment. Represents the process noise covariance;

[0066] S4. Update step: Update the status based on the observation data;

[0067] ;

[0068] ;

[0069] ;

[0070] In the formula: Indicates Kalman gain, This represents the updated state estimate; This represents the updated error covariance matrix. Represents the measurement noise covariance matrix. Represents the identity matrix.

[0071] Optionally, during the final attitude adjustment, based on the calculated pose error, different control commands are generated for the three different three-position four-way electro-hydraulic valves in the hydraulic module. This maps the end-position attitude error to the control of the hydraulic cylinder and rotary motor, achieving virtual-real fusion adaptive trajectory optimization control. The specific details are as follows:

[0072] The control command module is based on robust adaptive control principles and adjusts the position and orientation according to the optimized pose adjustment. The control voltage of the electro-hydraulic valve can be calculated using the following expression:

[0073] proportional valve voltage : ;

[0074] Robust item : ;

[0075] Online update of unknown parameters : ;

[0076] In the formula, For adaptive gain; , For compensation terms of unknown system parameters estimated online; Robust gain; Saturation function; Error limits; Adaptive learning rate; These are characteristic quantities related to system input, such as hydraulic cylinder displacement and speed.

[0077] In summary, this application includes the following beneficial technical effects:

[0078] This invention achieves high-precision real-time estimation of the robotic arm's end-effector pose by fusing multi-source pose information from tilt sensors, displacement sensors, and rotary encoders, combined with a virtual-real fusion algorithm based on Kalman filtering and adaptive weighting strategies. Furthermore, robust adaptive control converts the pose error into real-time control voltage for the electro-hydraulic proportional valve, effectively solving problems such as hydraulic system lag, low trajectory tracking accuracy, poor anti-interference capability, and insufficient adaptability to complex working conditions inherent in traditional crushing robots. This enables adaptive optimization and dynamic compensation of the robotic arm's trajectory, thereby improving the accuracy of end-effector pose control and crushing efficiency. Attached Figure Description

[0079] Figure 1 This is a schematic diagram of the arrangement of the crusher and tilt sensor in this application;

[0080] Figure 2 This is a schematic diagram of the hydraulic module in the control system of this application;

[0081] Figure 3 This is the overall flowchart of the control method of this application.

[0082] Explanation of reference numerals in the attached figures:

[0083] 1. Motor controller; 2. High-speed motor; 3. High-speed high-pressure pump; 4. Relief valve; 5. Three-position four-way proportional valve I; 6. Boom hydraulic cylinder; 7. Three-position four-way proportional valve II; 8. Arm hydraulic cylinder; 9. Three-position four-way proportional valve III; 10. Rotary motor; 11. Tilt sensor I; 12. Displacement sensor I; 13. Tilt sensor II; 14. Displacement sensor II; 15. Rotary encoder; 16. LiDAR; 17. Pose calculation module; 18. Command control module; 19. Pressure sensor; 20. Booster tank; 21. Check valve; 22. Filter. Detailed Implementation

[0084] The following is in conjunction with the appendix Figure 1-3 This application will be described in further detail.

[0085] This application discloses a pose control system and control method for an intelligent crushing robot; the control system includes a hydraulic module, a pose signal acquisition module and a pose control module arranged in the robot;

[0086] The pose signal acquisition module is used to acquire the rotation angle of each joint of the crushing robot, the displacement of the hydraulic cylinder, and the angle and speed of each motor output shaft.

[0087] refer to Figure 1 The pose signal acquisition module includes tilt sensor 11, displacement sensor 12, tilt sensor 2 13, displacement sensor 2 14, rotary encoder 15, speed sensor, and lidar 16. Tilt sensor 11 and tilt sensor 2 13 are installed at the joints of the upper arm and lower arm, respectively, to monitor the rotation angle of each joint. Displacement sensor 12 and displacement sensor 2 14 are installed inside the hydraulic cylinders of the upper arm and lower arm to measure the displacement of the hydraulic cylinders. Rotary encoder 15 is installed on the output shaft of rotary motor 10 to measure the angle of the output shaft of rotary motor 10. Lidar 16 is installed on the top of the crushing robot's working platform to acquire point cloud data of the crushing working area.

[0088] refer to Figure 1 The hydraulic module includes a main circuit powered by a high-pressure pump and three parallel branch circuits. The high-pressure pump is driven by a high-speed motor 2, which is connected to a power supply and a motor controller 1. One branch circuit consists of a three-position four-way electro-hydraulic valve 1 connected in series with the boom hydraulic cylinder 6; another branch circuit consists of a three-position four-way electro-hydraulic valve 2 connected in series with the forearm hydraulic cylinder 8; and a third branch circuit consists of a three-position four-way electro-hydraulic valve 3 connected in series with the rotary motor 10. Pressure sensors 19 are installed at the outlet and inlet of each three-position four-way electro-hydraulic valve in each branch circuit. A booster tank 20 is also connected in series on the main circuit, and an overflow valve 4 is connected in parallel to the three branch circuits. A filter is installed before the booster tank 20.

[0089] The pose control module includes a pose calculation module 17 and a command control module 18. The pose calculation module 17 is used to solve the angle information collected by the pose signal acquisition module and to monitor the pose of the crushing robot. The command control module 18 is used to perform pose adjustment control based on the current pose monitoring information.

[0090] This application also provides a pose control method for an intelligent crushing robot, the method including the following:

[0091] 1. Collect and calculate relevant pose data of the intelligent crushing robot, including: the attitude angle signals of the upper arm and lower arm of the crushing robot, the position signals of the hydraulic cylinders of the upper arm and lower arm, the rotation pose signal of the rotary motor, and the point cloud data of the crushing operation area.

[0092] 2. The forward kinematics matrix of the robotic arm is constructed based on the DH parameter method, and the end effector pose is calculated. The specific details are as follows:

[0093] DH transformation matrix:

[0094] ;

[0095] In the formula, For joint angle, This is the link offset. The length of the link. Linkage torsion angle; Substitute the rotation angle, boom angle, and forearm angle into the above formula;

[0096] End-effector pose matrix:

[0097] ;

[0098] End position coordinates:

[0099] ;

[0100] Euler angle attitude of the end effector:

[0101] Roll angle: ;

[0102] Pitch angle: ;

[0103] Yaw angle: ;

[0104] 3. Generating virtual trajectories based on Dynamic Motion Element (DMP);

[0105] Determine the coordinates of the break point based on point cloud data. And the initial conditions for DMP output: Constructing the DMP dynamic system:

[0106] ;

[0107] In the formula, This indicates the current end-effector pose of the crushing robot. This is represented as the initial pose; It is an inertial parameter, and is a scaling direction; Let be the spatial vector of the joint. and They represent The first and second derivatives, For the first Each joint angular velocity at time t, For the first Each joint Angular acceleration at time t; The target state is the state that the DMP dynamics system will eventually converge to. and For system parameters; The trajectory shape learner is a non-linear function;

[0108] By changing the target state and nonlinear The term can be used to adjust the endpoint and shape of the trajectory; if it is also necessary to change the velocity of the trajectory to obtain trajectories with different convergence velocities, the term can be scaled. accomplish;

[0109] nonlinear functions The desired trajectory is approximated by a combination of Gaussian functions;

[0110] , ;

[0111] In the formula, The number of Gaussian functions; It is the first One basis function; As weight; and These are the center and width of the Gaussian function, respectively;

[0112] Finally, the process is iterated to generate joint trajectories;

[0113] At each time step Update joint acceleration:

[0114] ;

[0115] Update joint speed:

[0116] ;

[0117] Update joint angles:

[0118] ;

[0119] Output joint space virtual trajectory:

[0120] ;

[0121] In the formula, For the first Each joint Angular acceleration at time t; These are inertial parameters; For stiffness gain; For the first The target position of each joint; For damping gain; For force feedback gain; For the first External forces acting on each joint; For the first Each joint angular velocity at time t; For the first Each joint The position at that moment; The spatial vector of the joint;

[0122] 4. Compare the virtual trajectory with the actual pose and calculate the pose error;

[0123] Position error ; for The actuator end-effector pose at a given moment. for The coordinates of the breakpoint at that moment;

[0124] The error is converted into joint space adjustment amount using the Jacobian matrix. : ;

[0125] In the formula is the generalized inverse of the Jacobian matrix;

[0126] Optimized pose adjustment amount : .

[0127] In this process, Kalman filtering technology is needed to fuse the input data to improve the accuracy of pose information. The specific details are as follows:

[0128] S1. Define the state vector of the end-effector pose as follows: Then in State vector at time step It can be represented as:

[0129] ;

[0130] S2. Establish the basic mathematical equations for Kalman filtering:

[0131] State transition equation: ;

[0132] Observation equation: ;

[0133] In the formula, Here is the state transition matrix. To control the input matrix, The joint angular velocity, For process noise, for The sensor measurement vector at any given time. For the observation matrix, For measuring noise;

[0134] S3, Prediction Step: Make predictions based on the state of the previous time step;

[0135] ;

[0136] ;

[0137] In the formula, Indicates based on Vector estimates at time t The state vector generated after predicting the value at time 1. Indicates the predicted state at the next moment. Represents the process noise covariance;

[0138] S4. Update step: Update the status based on the observation data;

[0139] ;

[0140] ;

[0141] ;

[0142] In the formula: Indicates Kalman gain, This represents the updated error covariance matrix. Represents the measurement noise covariance matrix. Represents the identity matrix;

[0143] 5. Convert the error into joint space adjustment amount and optimize the robot trajectory accordingly.

[0144] During the final attitude adjustment, based on the calculated pose error, different control commands are generated for the three different three-position four-way electro-hydraulic valves in the hydraulic module. This maps the end-position pose error to the control of the hydraulic cylinder and rotary motor, achieving virtual-real fusion adaptive trajectory optimization control. The specific details are as follows:

[0145] The control command module can be based on robust adaptive control laws and adjust the position and orientation according to the optimized pose adjustment amount. The control voltage of the electro-hydraulic valve can be calculated using the following expression:

[0146] proportional valve voltage : ;

[0147] Robust item : ;

[0148] Online update of unknown parameters : ;

[0149] In the formula, For adaptive gain; , For compensation terms of unknown system parameters estimated online; Robust gain; Saturation function; Error limits; Adaptive learning rate; These are characteristic quantities related to system input, such as hydraulic cylinder displacement and speed.

[0150] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A pose control system for an intelligent crushing robot, characterized in that: This includes a hydraulic module, a pose signal acquisition module, and a pose control module installed in the robot; The hydraulic module includes a main circuit supplied by a high-pressure pump and three parallel branch circuits; one branch circuit consists of a three-position four-way electro-hydraulic valve 1 connected in series with the boom hydraulic cylinder, another branch circuit consists of a three-position four-way electro-hydraulic valve 2 connected in series with the forearm hydraulic cylinder, and the third branch circuit consists of a three-position four-way electro-hydraulic valve 3 connected in series with the rotary motor. The pose signal acquisition module is used to acquire the rotation angle of each joint of the crushing robot, the displacement of the hydraulic cylinder, and the angle and speed of each motor output shaft. The pose control module includes a pose calculation module and a control command module. The pose calculation module is used to solve the angle information collected by the pose signal acquisition module and to monitor the pose of the crushing robot. The control command module is used to perform pose adjustment control based on the current pose monitoring information.

2. The pose control system for an intelligent crushing robot according to claim 1, characterized in that: The main circuit is equipped with a booster oil tank, and a hydraulic control check valve is installed between the booster oil tank and the main circuit; three branch circuits are connected in parallel with overflow valves, and the overflow valves are connected in series with the booster oil tank. The pose signal acquisition module includes tilt sensor one, displacement sensor one, tilt sensor two, displacement sensor two, rotary encoder, and lidar. Tilt sensor one and tilt sensor two are respectively installed at the joints of the upper arm and forearm to monitor the rotation angle of each joint. Displacement sensor one and displacement sensor two are installed inside the hydraulic cylinders of the upper arm and forearm to measure the displacement of the hydraulic cylinders. The rotary encoder is installed on the output shaft of the rotary motor to measure the angle of the rotary motor output shaft. The lidar is installed on the top of the crushing robot's working platform to acquire point cloud data of the crushing working area.

3. A pose control method for an intelligent crushing robot, employing the pose control system for an intelligent crushing robot as described in claim 2, characterized in that: Includes the following: Collect and process relevant pose data of the intelligent crushing robot; Calculate the actual pose of the end effector based on the DH parameter method; Virtual trajectories are generated based on Dynamic Motion Element (DMP). The virtual trajectory is compared with the actual pose, and the pose error is calculated. The error is converted into joint space adjustment amount, and the robot trajectory is optimized by adjusting the adjustment.

4. The pose control method for an intelligent crushing robot according to claim 3, characterized in that: The relevant pose data includes: the attitude angle signals of the upper and lower arms of the crushing robot, the position signals of the hydraulic cylinders of the upper and lower arms, the rotation pose signal of the rotary motor, and the point cloud data of the crushing operation area.

5. The pose control method for an intelligent crushing robot according to claim 4, characterized in that: The forward kinematics matrix of the robotic arm is constructed based on the DH parameter method, and the end effector pose is calculated in detail as follows: DH transformation matrix: ； In the formula, For joint angle, This is the link offset. The length of the link. Linkage torsion angle; Substituting the slewing angle, boom angle, and forearm angle into the above formula, we obtain the following: End-effector pose matrix: ； End position coordinates: ； Euler angle attitude of the end effector: Roll angle ; Pitch angle ; Yaw angle .

6. The pose control method for an intelligent crushing robot according to claim 5, characterized in that: The specific details of generating virtual trajectories based on Dynamic Motion Element (DMP) are as follows: Determine the coordinates of the break point based on point cloud data. And the initial conditions for DMP output: Constructing the DMP dynamic system: ； In the formula, This indicates the current end-effector pose of the crushing robot. This is represented as the initial pose; It is an inertial parameter, and is a scaling direction; Let be the spatial vector of the joint. and They represent The first and second derivatives, For the first Each joint angular velocity at time t, For the first Each joint Angular acceleration at time t; The target state is the state that the DMP dynamics system will eventually converge to. and For system parameters; The trajectory shape learner is a non-linear function; By changing the target state and nonlinear The items are used to adjust the endpoint and shape of the trajectory; nonlinear functions The desired trajectory is approximated by a combination of Gaussian functions; ， ； In the formula, The number of Gaussian functions; It is the first One basis function; As weight; Center of Gaussian function The width of the Gaussian function; Finally, the process is iterated to generate joint trajectories; At each time step Update joint angular acceleration: ； Update joint angular velocity: ； Update joint angles: ； Output joint space virtual trajectory: ； In the formula, For the first Each joint Angular acceleration at time t; These are inertial parameters; For stiffness gain; For the first The target position of each joint; For damping gain; For force feedback gain; For the first External forces acting on each joint; For the first Each joint angular velocity at time t; For the first Each joint The position at that moment; is the spatial vector of the joint.

7. The pose control method for an intelligent crushing robot according to claim 6, characterized in that: Position error ; for The actuator end-effector pose at a given moment. for The coordinates of the breakpoint at that moment; The error is converted into joint space adjustment amount using the Jacobian matrix. : ; In the formula is the generalized inverse of the Jacobian matrix; Optimized pose adjustment amount : .

8. The pose control method for an intelligent crushing robot according to claim 7, characterized in that: When comparing the virtual trajectory with the actual pose, Kalman filtering is used to fuse the input data and output the end effector pose signal. The pose error is formed by comparing it with the virtual trajectory. Trajectory optimization is performed; specifically, the following steps are included: S1. Define the state vector of the end-effector pose as follows: Then in State vector at time step It can be represented as: ； S2. Establish the basic mathematical equations for Kalman filtering: State transition equation: ; Observation equation: ; In the formula, Here is the state transition matrix. To control the input matrix, The joint angular velocity, For process noise, for The sensor measurement vector at any given time. For the observation matrix, For measuring noise; S3, Prediction Step: Make predictions based on the state of the previous time step; ； ； In the formula, Indicates based on Vector estimates at time t The state vector generated after predicting the value at time 1. Indicates the predicted state at the next moment. Represents the process noise covariance; S4. Update step: Update the status based on the observation data; ； ； ； In the formula: Indicates Kalman gain, This represents the updated state estimate; This represents the updated error covariance matrix. Represents the measurement noise covariance matrix. Represents the identity matrix.

9. The pose control method for an intelligent crushing robot according to claim 8, characterized in that: During the final attitude adjustment, based on the calculated pose error, different control commands are generated for the three different three-position four-way electro-hydraulic valves in the hydraulic module. This maps the end-position pose error to the control of the hydraulic cylinder and rotary motor, achieving virtual-real fusion adaptive trajectory optimization control. The specific details are as follows: The control command module is based on robust adaptive control principles and adjusts the position and orientation according to the optimized pose adjustment. The control voltage of the electro-hydraulic valve can be calculated using the following expression: proportional valve voltage : ; Robust item : ; Online update of unknown parameters : ; In the formula, For adaptive gain; , For compensation terms of unknown system parameters estimated online; Robust gain; It is a saturation function; Error limits; Adaptive learning rate; These are characteristic quantities related to system input, such as hydraulic cylinder displacement and speed.

Images