Robot adaptive leveling and paving control method and system based on complex slope surface

Abstract

The application discloses a robot self-adaptive leveling and paving control method and system based on a complex slope surface, relates to the technical field of robot control, and comprises the following steps: preprocessing multi-source sensing data of an acquired complex slope surface robot and a target paving structure, and constructing joint state variables and unified environment representation; based on the joint state variables and the unified environment representation, constructing a leveling and paving integrated collaborative optimization problem with the minimum comprehensive cost as a target, and solving to obtain collaborative control instructions; based on the collaborative control instructions, generating actions through an adaptive leveling control generation platform, generating mechanical arm execution actions through visual-mixed paving control, and acquiring execution state to update the joint state variables; and correcting mechanical arm control parameters according to a single paving completion quality index deviation. The application is used to solve the problems of coupling of leveling control and paving control and insufficiency of adaptability in a traditional robot control method.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and more specifically, to a robot adaptive leveling and paving control method and system based on complex slopes. Background Technology

[0002] As engineering construction expands into complex terrains and irregular working environments, the application of robots in construction scenarios such as road paving, slope protection, and retaining wall construction is gradually increasing. Compared to flat sites, complex slope environments typically feature large variations in slope gradient, uneven local undulations, and discontinuous working surfaces, placing higher demands on the stability and operational accuracy of robot platforms. Existing engineering robots mostly employ a control architecture where platform leveling control and working mechanism control are independent. While this control method can meet basic construction needs under simple or regular site conditions, in complex slope environments, there is a significant coupling between platform posture changes and the stress state of the working mechanism, making it difficult for independent control methods to coordinate the dynamic relationship between the two in a timely manner.

[0003] The existing technical solutions have at least the following technical problems: 1. The problem of insufficient coupling between leveling control and paving control; 2. The problem of difficulty in continuous adaptive leveling and paving control.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a robot adaptive leveling and paving control method and system based on complex slopes. The present invention constructs an integrated collaborative optimization problem of leveling and paving with the goal of minimizing the overall cost. It generates execution actions through adaptive leveling control and vision-based hybrid paving control, updates joint state variables, and corrects robotic arm control parameters, thus solving the problems of insufficient coupling and adaptation between leveling control and paving control in traditional robot control methods.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a robot adaptive leveling and paving control method based on complex slopes, comprising the following steps: preprocessing the multi-source perception data of the robot on the acquired complex slope and the target paving structure, and constructing joint state variables and a unified environmental representation; based on the joint state variables and the unified environmental representation, constructing an integrated collaborative optimization problem of leveling and paving with the goal of minimizing the overall cost, and solving it to obtain collaborative control commands; based on the collaborative control commands, generating platform actions through adaptive leveling control, generating robotic arm actions through vision-sensor hybrid paving control, and obtaining execution state updates to the joint state variables; and correcting the robotic arm control parameters according to the quality index deviation of a single paving operation.

[0007] In a preferred embodiment, the method for constructing the joint state variables includes: obtaining the platform pose of the robot body in the world coordinate system based on the body state data and global positioning data through a state estimation algorithm; calculating the manipulator pose based on the platform pose and according to the robot kinematic model; and constructing joint state variables based on the platform pose, the manipulator pose, and the body state data.

[0008] In a preferred embodiment, the method for constructing the unified environmental representation includes: uniformly converting local environmental data and contact force data to a preset robotic arm end-effector coordinate system; based on the robotic arm pose, initially converting the data in the tool coordinate system to the world coordinate system to obtain preliminary world data; performing point cloud registration between the preliminary world data and the target paving structure obtained through BIM modeling, and solving the fine registration transformation matrix of the preliminary world data relative to the world coordinate system; based on the fine registration transformation matrix, converting the local environmental data and contact force data in the tool coordinate system to the world coordinate system, and integrating the global environmental data and the target paving structure to construct a unified environmental representation.

[0009] In a preferred embodiment, the step of constructing a leveling and paving integrated collaborative optimization problem with the goal of minimizing the overall cost based on joint state variables and unified environmental representation, and solving it to obtain collaborative control commands, includes: constructing a coupled prediction model containing platform leveling dynamics and robotic arm paving dynamics as inputs using unified environmental representation and joint state variables, and predicting the robot's dynamic response within a preset time domain; based on the dynamic response, constructing a comprehensive cost function according to leveling energy consumption, paving trajectory error and task time cost, and solving it to obtain collaborative control commands with the goal of minimizing the overall cost; the collaborative control commands include a platform leveling posture sequence and a paving execution posture sequence; and calculating dynamic disturbance force based on a pre-constructed robotic arm dynamics model, combined with the paving execution posture sequence and robotic arm mass.

[0010] In a preferred embodiment, the step of generating platform execution actions based on cooperative control commands and adaptive leveling control includes: acquiring full-state measurement values ​​of the platform in each preset control cycle; constructing a platform state prediction equation based on the platform leveling attitude sequence and using dynamic disturbance force as feedforward compensation; and based on the platform state prediction equation, solving a preset time-domain platform leveling command sequence through model predictive control and executing the first command in the platform leveling command sequence.

[0011] In a preferred embodiment, the step of generating robotic arm actions through vision-force hybrid paving control includes: driving the robotic arm end effector toward the target contact point via visual servo control based on the paving execution posture sequence, and monitoring the end effector contact force in real time; when the end effector contact force exceeds a preset first threshold, solving the robotic arm pose compensation in each preset control cycle based on preset desired mechanical impedance parameters and desired contact force through impedance control equations; converting the robotic arm pose compensation into robotic arm joint control commands, and driving the robotic arm to execute them.

[0012] In a preferred embodiment, the step of correcting the robotic arm control parameters based on the quality index deviation of a single paving operation includes: obtaining the actual quality index and reference quality index of a single paving operation, and calculating the quality index deviation; when the quality index deviation exceeds a preset second threshold, correcting the robotic arm control parameters according to a preset mapping rule based on the type of the quality index deviation.

[0013] In a preferred embodiment, the method for obtaining the actual quality indicators includes: the actual quality indicators include surface flatness and the width of the joint between adjacent blocks; performing a three-dimensional scan of the current paved surface using sensors deployed at the end of a robotic arm to obtain point cloud data; fitting the point cloud data to obtain the plane normal vector and the point cloud center, calculating the normal distance from all points in the point cloud data to the fitted plane, and taking the maximum absolute value as the surface flatness; obtaining the edge feature point clouds at the joint between the current paved surface and the adjacent paved surface, and performing feature matching to obtain corresponding point pairs; calculating the Euclidean distance between all corresponding point pairs, and taking the average value as the width of the joint between adjacent blocks.

[0014] The robot adaptive leveling and paving control system based on complex slopes includes: a data acquisition module, used to preprocess the multi-source perception data of the robot on the complex slope and the target paving structure, and construct joint state variables and a unified environmental representation; a collaborative decision-making module, used to construct an integrated collaborative optimization problem of leveling and paving with the goal of minimizing the overall cost based on the joint state variables and the unified environmental representation, and solve it to obtain collaborative control commands; a motion control module, used to generate platform execution actions through adaptive leveling control based on collaborative control commands, generate robotic arm execution actions through vision-force hybrid paving control, and obtain execution status to update joint state variables; and a parameter learning module, used to correct the robotic arm control parameters based on the quality index deviation of a single paving operation.

[0015] The technical effects and advantages of this invention, which is a robot-based adaptive leveling and paving control method and system for complex slopes, are as follows: This invention constructs an integrated collaborative optimization problem for leveling and paving with the goal of minimizing overall cost, based on joint state variables and a unified environmental representation. This achieves unified decision-making for platform leveling control and paving operation control, avoiding control conflicts and operational instability caused by the two types of control being independent. Based on collaborative control commands, the platform executes actions through adaptive leveling control, and the robotic arm executes actions through vision-force hybrid paving control. The joint state variables are updated by acquiring the execution status, forming a control closed loop and achieving high-precision and robust execution of collaborative control commands. The robotic arm control parameters are corrected based on the deviation of the quality index of a single paving operation, enabling the robot to have adaptive learning capabilities for uncertainties such as changes in material properties and tool wear, achieving continuous self-improvement and long-term stable reliability of paving quality. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the robot adaptive leveling and paving control method based on complex slopes provided in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of a robot adaptive leveling and paving control system based on complex slopes, provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, Figure 1 The present invention provides a robot-based adaptive leveling and paving control method for complex slopes, comprising the following steps: S1 preprocesses the multi-source perception data of the robot on the complex slope with the target paving structure and constructs joint state variables and unified environmental representation.

[0020] In this embodiment, the multi-source sensing data includes global positioning data, global environment data, local environment data, contact force data, and body state data; the joint state vector includes platform pose, robotic arm pose, platform rotational angular velocity, robotic arm joint position, and joint velocity.

[0021] In this embodiment, the method for constructing the joint state variables includes: S101, based on the body state data and global positioning data, obtains the platform pose of the robot body in the world coordinate system through a state estimation algorithm; S102, based on the platform pose, calculates the robot arm pose according to the robot kinematics model; S103 constructs joint state variables based on platform pose, robotic arm pose, and body state data.

[0022] In this embodiment S101, the body state data includes data that directly measures the robot's own state and the interaction between the body and the external environment.

[0023] It should be noted that the state estimation algorithm described in S101 and the robot kinematic model described in S102 are existing technologies, and will not be described again in this embodiment.

[0024] It should be noted that platform pose includes the robot's platform posture and platform positioning information.

[0025] In this embodiment, the method for constructing the unified environment representation includes: S104, converts local environmental data and contact force data to the preset coordinate system of the robotic arm end-effector; S105, based on the robot arm's pose, initially transforms the data in the tool coordinate system to the world coordinate system to obtain preliminary world data; S106, perform point cloud registration between the preliminary world data and the target paving structure obtained through BIM modeling, and solve the fine registration transformation matrix of the preliminary world data relative to the world coordinate system; S107, based on the fine registration transformation matrix, transforms the local environmental data and contact force data in the tool coordinate system to the world coordinate system, and integrates the global environmental data and the target paving structure to construct a unified environmental representation.

[0026] It should be noted that the global environmental data mentioned in S107 refers to the set of information used to perceive the macroscopic environment of the robot's working area; the local environmental data mentioned in S107 refers to the set of information used to perceive the microscopic geometric and physical interaction of the robotic arm's end-effector working area.

[0027] It should be noted that the target paving structure mentioned in S106 refers to the complete digital model obtained by modeling the entire target paving project through BIM. It contains all the geometric and semantic information of the project from the macro-topography to each paving block in the world coordinate system.

[0028] It should be noted that the point cloud registration described in S106 is an existing technical method, and will not be described again in this embodiment.

[0029] It should be noted that the construction of a unified environmental representation described in S107 solves the decision-making conflict problem caused by the dispersion of sensing sources and the lack of unified coordinate references in traditional systems, and provides a unique and reliable information source for integrated collaborative control.

[0030] S2, based on joint state variables and unified environment representation, constructs a leveling and paving integrated collaborative optimization problem with the goal of minimizing the comprehensive cost, and obtains the collaborative control command by solving it.

[0031] In this embodiment, the construction of a leveling and paving integrated collaborative optimization problem based on joint state variables and a unified environment representation, with the objective of minimizing the overall cost, and the solution to obtain the collaborative control command, includes: S201 uses a unified environmental representation and joint state variables as input to construct a coupled prediction model that includes platform leveling dynamics and robotic arm paving dynamics, and predicts the robot's dynamic response within a preset time domain. S202, based on dynamic response, constructs a comprehensive cost function according to leveling energy consumption, paving trajectory error and task time cost, and obtains the cooperative control command by minimizing the comprehensive cost; S203, the collaborative control instructions include a platform leveling attitude sequence and a paving execution attitude sequence; S204, based on a pre-built robotic arm dynamics model, combines the paving execution posture sequence with the robotic arm mass to calculate dynamic disturbance force.

[0032] In this embodiment S202, the formula for the comprehensive cost function is as follows:

[0033]

[0034]

[0035] In the formula, For the comprehensive cost function, For the overall cost of the process, The cost of terminal posture convergence, The start time of the task. For the end time of the operation, To balance the weighting of energy consumption, To level the control input vector, As the weight of the paving trajectory error, To account for paving trajectory errors. For the robotic arm pose, For reference, the paving path pose As a weight for the time cost of the task, For task time cost, The weight of the platform attitude error at the moment of operation terminal. The platform's posture at the moment of operation. The target platform attitude at the moment of operation.

[0036] In this embodiment S204, the specific formula for calculating the dynamic interference force is as follows:

[0037] In the formula, The dynamic disturbance force generated by the movement of the robotic arm on the platform. The total mass of the robotic arm, This indicates that the force is a reaction force. Let be the instantaneous acceleration of the center of mass of the entire robotic arm system in the world coordinate system.

[0038] In this embodiment, the method for obtaining the instantaneous acceleration described in the dynamic disturbance force calculation formula includes: By using the inverse dynamics model of the robotic arm, the paving execution trajectory is converted into a joint angle trajectory; Based on the joint angle trajectory and its derivative, the trajectory of the center of mass is calculated by combining the mass and center of mass distribution information of each link with the dynamic model of the robotic arm. The instantaneous acceleration is obtained by taking the second derivative of the trajectory of the center of mass with respect to time.

[0039] It should be noted that by controlling the predictive model, the leveling energy consumption, paving trajectory error, and task time cost are incorporated into an optimization problem for solution, outputting cooperative control commands. This design enables the system to proactively predict the interference of the robotic arm's movement on the platform and plan compensation in advance, overcoming the limitations of independent planning and ensuring optimal global performance from the decision-making level.

[0040] S3, based on collaborative control commands, generates platform execution actions through adaptive leveling control, generates robotic arm execution actions through vision-force hybrid paving control, and obtains execution status to update joint state variables.

[0041] In this embodiment, the action executed by the adaptive leveling control generation platform based on cooperative control commands includes: S301: Acquire the full-state measurement values ​​of the platform within each preset control cycle, and construct the platform state prediction equation based on the platform leveling attitude sequence and with dynamic disturbance force as feedforward compensation. S302, based on the platform state prediction equation, uses model predictive control to solve the preset time domain platform leveling instruction sequence and executes the first instruction in the platform leveling instruction sequence.

[0042] In this embodiment S301, the platform state prediction equation formula is as follows:

[0043] In the formula, The time derivative of the platform state vector. For the platform state vector, The system state matrix, To control the input variable vector, To control the input matrix, For the interference input matrix, This refers to the dynamic disturbance force generated by the movement of the robotic arm on the platform.

[0044] It should be noted that the method for obtaining the platform state vector in the platform state prediction equation includes: fusing and estimating the obtained full-state measurement values ​​of the platform through a state estimation algorithm, and then solving the equation.

[0045] It should be noted that through feedforward compensation, the platform can quickly and smoothly track the optimal posture sequence while the robotic arm is operating dynamically, effectively suppressing internal coupling interference and ensuring the stability of the operation process.

[0046] In this embodiment, the step of generating robotic arm actions through vision-based hybrid paving control includes: S303, based on the paving execution posture sequence, drives the end effector of the robotic arm to move toward the target contact point through visual servo control, and monitors the end effector contact force in real time; S304, When the end contact force exceeds the preset first threshold, the robot arm pose compensation is solved in each preset control cycle based on the preset desired mechanical impedance parameters and desired contact force through the impedance control equation. S305 converts the robot arm pose compensation into robot arm joint control commands and drives the robot arm to execute them.

[0047] It should be noted that the preset desired mechanical impedance parameters mentioned in S304 include the desired inertia matrix, the desired damping matrix, and the desired stiffness matrix.

[0048] It should be noted that visual servo control is an existing technology and will not be described in detail in this embodiment.

[0049] In this embodiment S304, the impedance control equation is specifically as follows:

[0050] In the formula, Let be the desired inertia matrix. Let be the desired damping matrix. Let be the desired stiffness matrix. Let be the second derivative of the pose error. The first derivative of the pose error. Let be the pose error vector. For end contact force, The desired contact force.

[0051] It should be noted that through visual servo control and impedance control, the robot can ensure high-precision approach in the non-contact stage, and in the contact stage, it can perform stable physical operations according to the expected force and impedance characteristics, thus ensuring the physical interaction quality of the paving process.

[0052] S4, adjusts the robotic arm control parameters based on the quality index deviation of a single paving operation.

[0053] In this embodiment, the step of correcting the robotic arm control parameters based on the quality index deviation of a single paving operation includes: S401, obtain the actual quality indicators and reference quality indicators of a single paving operation, and calculate the deviation of the quality indicators; S402, when the deviation of the quality index exceeds the preset second threshold, the control parameters of the robotic arm are corrected according to the type of the quality index deviation and the preset mapping rule.

[0054] In this embodiment, the method for obtaining the actual quality indicators includes: S403, the actual quality indicators include surface flatness and the width of the gap between adjacent blocks; S404 uses sensors deployed at the end of the robotic arm to perform a three-dimensional scan of the current paved surface and acquire point cloud data; S405, Fit the point cloud data to obtain the plane normal vector and the point cloud center, calculate the normal distance from all points in the point cloud data to the fitted plane, and take the maximum absolute value as the surface flatness. S406, Obtain the edge feature point cloud of the joint between the current paved body and the adjacent paved body, and perform feature matching to obtain the corresponding point pairs; S407, calculate the Euclidean distance between all corresponding point pairs and take the average value as the gap width between adjacent blocks.

[0055] In this embodiment S403, the surface flatness calculation formula is as follows:

[0056] In the formula, For surface flatness, This is a collection of point clouds on the surface of a single paved block. To fit using the least squares method, The obtained plane normal vector, for The first in A three-dimensional point, for The coordinates of the center point.

[0057] In this embodiment S403, the formula for calculating the gap width between adjacent blocks is as follows:

[0058] In the formula, The width of the gap between adjacent blocks. The number of corresponding point pairs involved in the calculation. The function for calculating Euclidean distance is... For the current paving block One edge feature point, For the adjacent pavement Each corresponding edge feature point.

[0059] In this embodiment S402, the specific formula for calculating the corrected robotic arm control parameters is as follows:

[0060]

[0061] In the formula, For the slope of the complex broken surface, Code the type of basic materials. The updated value of the parameter. The value before the parameter was updated. This is the amount of process parameter correction. The preset learning rate parameter, For the preset mapping matrix, For actual quality indicators, For reference quality indicators.

[0062] It should be noted that by correcting the control parameters of the robotic arm based on the deviation of the quality indicators of a single paving operation, optimal operating experience under different working conditions can be accumulated, continuously improving the quality consistency and robustness of subsequent operations, and enabling the robot to have long-term adaptability to cope with uncertainties such as material variation and tool wear.

[0063] Example 2, Figure 2 The present invention provides a robot adaptive leveling and paving control system for complex slopes, comprising: The data acquisition module is used to preprocess the multi-source perception data of the complex slope robot and the target paving structure, and to construct joint state variables and unified environmental representation. The collaborative decision-making module is used to construct an integrated collaborative optimization problem of leveling and paving with the goal of minimizing the overall cost based on joint state variables and unified environmental representation, and to obtain collaborative control commands by solving the problem. The motion control module is used to generate platform actions based on cooperative control commands through adaptive leveling control, generate robotic arm actions through vision-force hybrid paving control, and obtain execution status updates to joint state variables. The parameter learning module is used to correct the control parameters of the robotic arm based on the deviation of the quality indicators after a single paving operation.

[0064] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0065] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0066] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0067] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0068] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0069] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A robot-based adaptive leveling and paving control method for complex slopes, characterized in that, Includes the following steps: The multi-source perception data of the robot on the complex slope were preprocessed with the target paving structure, and joint state variables and unified environmental representation were constructed. Based on joint state variables and unified environment representation, a coordinated optimization problem of leveling and paving integration with the goal of minimizing comprehensive cost is constructed, and the coordinated control command is obtained by solving the problem. Based on collaborative control commands, the platform executes actions through adaptive leveling control, and the robotic arm executes actions through vision-force hybrid paving control, and the execution status is obtained to update the joint state variables; The control parameters of the robotic arm are adjusted based on the deviation of the quality indicators after a single paving operation.

2. The robot adaptive leveling and paving control method based on complex slopes according to claim 1, characterized in that, The method for constructing the joint state variables includes: Based on the body state data and global localization data, the platform pose of the robot body in the world coordinate system is obtained through a state estimation algorithm. Based on the platform pose, the robot arm pose is calculated according to the robot kinematics model; Based on the platform pose, robotic arm pose, and body state data, a joint state variable is constructed.

3. The robot adaptive leveling and paving control method based on complex slopes according to claim 2, characterized in that, The method for constructing the unified environment representation includes: The local environmental data and contact force data are uniformly converted to the preset coordinate system of the robotic arm end effector. Based on the robot arm's pose, the data in the tool coordinate system is initially transformed to the world coordinate system to obtain preliminary world data; The preliminary world data is registered with the target paved structure obtained through BIM modeling, and the fine registration transformation matrix of the preliminary world data relative to the world coordinate system is solved. Based on the precise registration transformation matrix, the local environmental data and contact force data in the tool coordinate system are transformed to the world coordinate system, and the global environmental data and target paving structure are integrated to construct a unified environmental representation.

4. The robot adaptive leveling and paving control method based on complex slopes according to claim 3, characterized in that, Based on joint state variables and a unified environment representation, a coordinated optimization problem integrating leveling and paving is constructed with the objective of minimizing the overall cost, and the coordinated control commands are obtained by solving the problem, including: Using a unified environmental representation and joint state variables as input, a coupled prediction model is constructed that includes platform leveling dynamics and robotic arm paving dynamics, and the dynamic response of the robot is predicted within a preset time domain. Based on dynamic response, a comprehensive cost function is constructed according to the leveling energy consumption, paving trajectory error and task time cost, and the cooperative control command is obtained by minimizing the comprehensive cost. The collaborative control commands include a platform leveling attitude sequence and a paving execution attitude sequence; Based on a pre-built robotic arm dynamics model, dynamic disturbance force is calculated by combining the paving execution posture sequence with the robotic arm mass.

5. The robot adaptive leveling and paving control method based on complex slopes according to claim 4, characterized in that, The action executed by the adaptive leveling control generation platform based on the cooperative control command includes: The platform's full-state measurement values ​​are acquired within each preset control cycle. Based on the platform's leveling attitude sequence and with dynamic disturbance force as feedforward compensation, a platform state prediction equation is constructed. Based on the platform state prediction equation, the platform leveling instruction sequence in the preset time domain is solved by model predictive control, and the first instruction in the platform leveling instruction sequence is executed.

6. The robot adaptive leveling and paving control method based on complex slopes according to claim 5, characterized in that, The action of the robotic arm generated by the vision-based hybrid paving control includes: Based on the paving execution posture sequence, the robotic arm end effector is driven to move toward the target contact point through visual servo control, and the end effector contact force is monitored in real time. When the end contact force exceeds the preset first threshold, the robot arm pose compensation is solved in each preset control cycle based on the preset desired mechanical impedance parameters and desired contact force through the impedance control equation. The robot arm pose compensation is converted into robot arm joint control commands, which then drive the robot arm to execute them.

7. The robot adaptive leveling and paving control method based on complex slopes according to claim 6, characterized in that, The process of correcting the robotic arm control parameters based on the quality index deviation of a single paving operation includes: Obtain the actual and reference quality indicators for a single paving operation, and calculate the deviation of the quality indicators. When the deviation of the quality index exceeds the preset second threshold, the control parameters of the robotic arm are corrected according to the type of quality index deviation and the preset mapping rules.

8. The robot adaptive leveling and paving control method based on complex slopes according to claim 7, characterized in that, The method for obtaining the actual quality indicators includes: The actual quality indicators include surface flatness and the width of the gap between adjacent blocks; The surface of the current paved structure is scanned in three dimensions by sensors deployed at the end of the robotic arm to obtain point cloud data; The plane normal vector and the point cloud center are obtained by fitting the point cloud data. The normal distance from all points in the point cloud data to the fitted plane is calculated, and the maximum absolute value is taken as the surface flatness. The edge feature point clouds at the joints between the current paved section and the adjacent paved section are obtained respectively, and feature matching is performed to obtain the corresponding point pairs; Calculate the Euclidean distance between all corresponding point pairs and take the average value as the gap width between adjacent blocks.

9. A system using the robot adaptive leveling and paving control method for complex slopes as described in any one of claims 1-8, comprising: The data acquisition module is used to preprocess the multi-source perception data of the complex slope robot and the target paving structure, and to construct joint state variables and unified environmental representation. The collaborative decision-making module is used to construct an integrated collaborative optimization problem of leveling and paving with the goal of minimizing the overall cost based on joint state variables and unified environmental representation, and to obtain collaborative control commands by solving the problem. The motion control module is used to generate platform actions based on cooperative control commands through adaptive leveling control, generate robotic arm actions through vision-force hybrid paving control, and obtain execution status updates to joint state variables. The parameter learning module is used to correct the control parameters of the robotic arm based on the deviation of the quality indicators after a single paving operation.

Images