Human-machine operation collaboration method based on operation intention for precast component installation

Abstract

The present invention belongs to the technical field of heavy-duty assembly construction robots, and specifically discloses a human-machine operation collaboration method based on operation intention for precast component installation. The method disclosed by the present invention includes: First, an improved shear thickening fluid dynamics model and a robot kinematics and dynamics model are established, and a robot control model is built according to the two dynamics models; Next, data is collected to train the robot control model; Then, the operator continuously adjusts the positions of the sensor and the operation handle according to the actual on-site position at the end effector of the robot and then performs the installation operation; Finally, the operation data is continuously transmitted to the robot control model, the operation intention is analyzed, the parameters are optimized, and an action plan is obtained and executed until the end. The present invention can enable the robot to accurately capture the operator's operation intention, and improve the compliance and efficiency of the assembly operation. The present invention can be widely applied to human-machine collaboration in precast component installation.

Description

Technical Field

[0001] The present invention belongs to the technical field of heavy-duty assembly construction robots, and specifically relates to a human-machine operation collaboration method based on operation intention for precast component installation. Background Art

[0002] With the continuous advancement of urbanization and the rapid development of the economy, modern construction projects are increasingly tending towards large-scale, complex, and intelligent. In the field of building precast component installation, problems such as large size and heavy weight of building precast components are often encountered during the assembly of precast components.

[0003] At present, the assembly method of precast components mostly relies on heavy-duty assembly operation robots. Among them, the end effector of the operation robot is used to pick up and place precast components, and the operator controls the robot to install precast components through the operation handle on the end effector based on the force feedback method. In this operation method, sensors are usually installed at the end of the robotic arm. When dealing with medium and large-sized building precast components, due to the fixed position of the sensors, during the force interaction between the human and the robot, the operation position of the operator will be restricted accordingly, which is likely to lead to the situation of being blocked by the line of sight, bringing inconvenience and trouble to the operation process.

[0004] In the field of robot compliance control, impedance control highly depends on the accurate construction of the dynamic model. In the process of human-machine collaboration, it is often only possible to set fixed impedance parameters and it is difficult to flexibly adjust dynamically according to the real-time changes on site. For example, the Chinese invention patent with the application publication number CN115026820A discloses a human-machine collaborative assembly robot control system and control method. Although it has advantages such as low development difficulty and low cost, however, under complex working conditions, this method has obvious deficiencies. It fails to fully consider the impact of the operator's impedance model in the process of human-machine collaboration, which weakens the operator's on-site experience to a certain extent. Therefore, it is necessary to break through the inherent installation mode of traditional sensors, fully consider the impedance characteristics of people, and design a human-machine operation collaboration method based on operation intention. Summary of the Invention

[0005] The purpose of the present invention is to provide a human-machine operation collaboration method based on operation intention for precast component installation, establish and train a dynamic model for thickened fluid control, and then combine with sensors that can change positions along with the operation handle, so as to enable the robot to accurately capture the operation intention of the operator and improve the compliance and efficiency of the assembly operation.

[0006] The technical solutions adopted by the present invention to achieve the above purpose are as follows:

[0007] A human-machine operation collaboration method based on operation intention for precast component installation, including the following steps carried out in sequence:

[0008] S1. Establish an improved shear thickening fluid dynamics model and a robot kinematics and dynamics model, and build a robot control model based on the two dynamics models;

[0009] S2. Collect human impedance parameter data and robot dynamics parameter data of individuals in different height ranges and different weight levels in each link of the operation process;

[0010] S3. Input the human impedance parameter data and the robot dynamics parameter data into the robot control model, and use machine learning algorithms to train and improve the robot control model to generate a trained robot control model;

[0011] S4. The operator adjusts the positions of the sensor and the operation handle according to the actual on-site position at the end effector of the robot and then performs the installation operation;

[0012] S5. Receive the operation data transmitted on the operation handle, and transmit the operation data to the trained robot control model to generate operation intention data;

[0013] S6. Based on the operation intention data, adjust the human impedance parameters and the robot dynamics parameters in the trained robot control model, generate robot execution data and execute corresponding actions;

[0014] S7. Whether the operator has subsequent operations;

[0015] If so, return to step S4;

[0016] If not, proceed to step S8;

[0017] S8. End.

[0018] As a limitation, the control formula of the shear thickening fluid dynamics model includes:

[0019]

[0020]

[0021] In the formula, is the external force applied to the object, is the velocity of the controlled object, is the virtual inertia term, is the system output value, is the viscosity of the system, is the power term, is the acceleration of the controlled object.

[0022] As a second limitation, the sensor and the operating handle described in step S4 are installed at one end of a telescopic rod; the other end of the telescopic rod is rotatably connected to the end effector of the robot.

[0023] As a third limitation, the human body impedance parameters include: human body motion damping coefficient, human body motion stiffness coefficient, and human body virtual mass.

[0024] As a fourth limitation, the robot dynamic parameters include: joint angle position parameters, joint angular acceleration, joint driving force, joint friction coefficient, joint stiffness coefficient, joint damping coefficient, and viscosity parameters of shear thickening fluid.

[0025] As a fifth limitation, the machine learning algorithm is the Transformer algorithm.

[0026] Due to the adoption of the above technical solutions, compared with the prior art, the technical progress achieved by the present invention lies in:

[0027] (1) The method of the present invention establishes and trains a robot control model, which can adaptively perform refined optimization processing on human body impedance parameters and robot dynamic parameters, enabling the robot to accurately capture the operation intention of the operator, and improving the compliance and efficiency of the assembly operation;

[0028] (2) In the method of the present invention, the sensor and the operating handle are installed at one end of a telescopic rod; the other end of the telescopic rod is rotatably connected to the end effector of the robot. The operator can adjust the positions of the sensor and the operating handle according to the actual position on site, solving the problem of line of sight occlusion caused by the fixed position of the sensor, and greatly improving the convenience and efficiency of the operation;

[0029] (3) In the method of the present invention, the mapping relationship between human body impedance parameters and robot dynamic parameters is mined, and a more accurate control strategy can be designed. When the robot executes tasks, it can interact with the external environment or the human body more smoothly and stably, reduce impact and jitter, and improve the fineness and safety of the operation;

[0030] (4) In the method of the present invention, the Transformer algorithm is used when selecting the machine learning algorithm, which can handle the complex non-linear relationship between human body impedance parameters and robot dynamic parameters, and improve the accuracy of the dynamic model.

[0031] The present invention belongs to the technical field of heavy-load assembly construction robots. By establishing and training a dynamic model for controlling thickening fluid, and combining with a sensor that can change its position along with the operating handle, the robot can accurately capture the operation intention of the operator, and improve the compliance and efficiency of the assembly operation. Brief Description of the Drawings

[0032] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention.

[0033] In the drawings:

[0034] Figure 1 is the process flow chart of the embodiment of the present invention;

[0035] Figure 2 is the framework schematic diagram of the robot control model in the embodiment of the present invention;

[0036] Figure 3 is the schematic diagram of human - machine cooperation in the embodiment of the present invention.

[0037] In the figure, 1 is the end - effector, 2 is the telescopic rod, 3 is the six - dimensional force sensor, and 4 is the operating handle. Detailed Embodiment

[0038] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only used to explain and illustrate the present invention and are not used to limit the present invention.

[0039] This embodiment is a human - machine operation cooperation method based on operation intention for precast component installation. As Figure 1 shown, this embodiment includes the following steps carried out in sequence:

[0040] S1. Establish an improved shear - thickening fluid dynamics model and a robot kinematics and dynamics model, and build a robot control model based on the two dynamics models.

[0041] Before establishing the shear - thickening fluid dynamics model and the robot kinematics and dynamics model, it is necessary to explore the mapping relationship between human impedance parameters and robot dynamics parameters. Among them, human impedance parameters include: human motion damping coefficient, motion stiffness coefficient, human virtual mass; robot dynamics parameters include: joint angle position parameter, joint angular acceleration, joint driving force, joint friction coefficient, joint stiffness coefficient, joint damping coefficient, and viscosity parameter of shear - thickening fluid.

[0042] The control formula of the shear - thickening fluid dynamics model includes:

[0043]

[0044]

[0045] In the formula, is the external force applied to the object, is the speed of the controlled object, is the virtual inertia term, is the system output value, is the viscosity of the system, is the power term, is the acceleration of the controlled object.

[0046] The obtained robot control model is as Figure 2 shown. The force feedback generated by the operator is combined with the force of the external environment to jointly form the total input force of the system. This input force is input into the robot control model, and the motion speed of the robot is calculated through calculation. Then, using the inverse kinematics and the mapping relationship of the velocity Jacobian, the motion speed is inversely solved into the velocity information of each joint of the robot. Finally, through the PID velocity controller, the driving force of each joint of the robot is calculated according to this velocity, and then the robot is driven to perform the corresponding motion.

[0047] S2. Collect the human impedance parameter data and robot dynamics parameter data of individuals in different height ranges and different weight levels in each link of the operation process.

[0048] S3. Input the human impedance parameter data and robot dynamics parameter data into the robot control model, and use machine learning algorithms to train and improve the robot control model to generate a trained robot control model.

[0049] In this embodiment, the machine learning algorithm used is the Transformer algorithm. This reinforcement learning algorithm can handle the complex non-linear relationship between human impedance parameters and robot dynamics parameters, providing strong support for the accuracy and reliability of the dynamics model.

[0050] S4. The operator adjusts the positions of the sensor and the operating handle 4 according to the actual on-site position at the end effector 1 of the robot and then performs the installation operation.

[0051] As Figure 3 shown, when the operator installs precast components, he needs to stand at the end effector 1 of the robot to control the operating handle 4. In this embodiment, the sensor is a six-axis force sensor 3, and the six-axis force sensor 3 is installed on the operating handle 4. The operating handle 4 is rotationally connected to the end effector 1 of the robot through a telescopic rod 2. The telescopic rod 2 can rotate and extend on the end effector 1, thereby changing the positions of the operating handle 4 and the six-axis force sensor 3.

[0052] To ensure the installation effect, the operator needs to continuously adjust the steering and telescopic length of the telescopic rod 2 according to the actual on-site installation situation, so as to adjust the positions of the six-axis force sensor 3 and the operating handle 4.

[0053] S5. Receive the operation data transmitted on the operation handle 4, and transmit the operation data into the trained robot control model to generate operation intention data.

[0054] S6. Based on the operation intention data, adjust the human impedance parameters and robot dynamics parameters in the trained robot control model, generate robot execution data and execute corresponding actions.

[0055] The operator will generate operation data on the operation handle 4. Previously, the training of the robot control model has been completed based on the collected data. Based on the currently received operation data, the robot can accurately analyze the operation intention. By flexibly adjusting the human impedance parameters and robot dynamics parameters, the optimal action plan can be quickly obtained and the corresponding movement can be executed.

[0056] S7. Whether the operator has subsequent operations;

[0057] If so, return to step S4;

[0058] If not, proceed to step S8.

[0059] S8. End.

[0060] In summary, in this embodiment, by establishing and training a robot control model and combining with the six-axis force sensor 3 that can change positions along with the operation handle 4, the robot can accurately capture the operation intention of the operator, improving the compliance and efficiency of the assembly operation.

Claims

1. A human-machine collaborative method for prefabricated component installation based on operation intention, characterized in that: The process includes the following steps: S1. Establish an improved shear thickening fluid dynamics model and a robot kinematics dynamics model, and build a robot control model based on the two dynamics models; S2, collect human body impedance parameter data and robot dynamic parameter data of individuals of different height ranges and weight levels in each link of the operation process; S3, inputting the human body impedance parameter data and the robot dynamic parameter data into the robot control model, and using the machine learning algorithm to train and improve the robot control model to generate a trained robot control model; S4. The operator performs the installation operation after adjusting the position of the sensor and the operating handle of the robot end effector according to the actual position on site; S5, receiving the operation data transmitted from the operating handle, transmitting the operation data to the trained robot control model, and generating operation intention data; S6. Based on the operation intention data, adjust the human body impedance parameters and robot dynamic parameters in the trained robot control model, generate robot execution data and perform corresponding actions; S7. Whether the operator has any follow-up operations; If yes, return to step S4; If not, proceed to step S8; S8. End.

2. According to claim 1, a human-machine collaborative operation method based on operation intention for prefabricated component installation is characterized in that: The control formula of the shear thickening fluid dynamics model includes: ； ； In the formula, is the external force applied to the object, is the speed of the controlled object, is the fictitious inertia term, is the system output value, is the viscosity of the system, is a power term, is the acceleration of the controlled object.

3. The human-machine collaborative operation method based on operation intention for prefabricated component installation according to claim 1 is characterized in that: In step S4, the sensor and the operating handle are installed at one end of a telescopic rod; the other end of the telescopic rod is rotatably connected to the end effector of the robot.

4. A human-machine collaborative operation method based on operation intention for prefabricated component installation according to any one of claims 1 to 3, characterized in that: The human body impedance parameters include: human body motion damping coefficient, human body motion stiffness coefficient and human body virtual mass.

5. A human-machine collaborative operation method based on operation intention for prefabricated component installation according to any one of claims 1 to 3, characterized in that: The robot dynamics parameters include: joint angle position parameters, joint angle acceleration, joint driving force, joint friction coefficient, joint stiffness coefficient, joint damping coefficient and viscosity parameter of shear thickening fluid.

6. A human-machine collaborative operation method based on operation intention for prefabricated component installation according to any one of claims 1 to 3, characterized in that: The machine learning algorithm is the Transformer algorithm.

Images