Master-slave control method and system for a master-slave heterogeneous teleoperation system

Abstract

The present invention discloses a master-slave control method and system for a master-slave heterogeneous teleoperation system, including: an operator inputs the link parameters of the slave robot for forward kinematics and inverse kinematics analysis to obtain forward kinematics and inverse kinematics models; the operator modifies the master-slave mapping scale coefficient to achieve the movement of the slave robot with different efficiencies and accuracies; a variable-scale incremental master-slave mapping method is designed to realize the mapping from the position increment of the master robot to the position increment of the slave robot, and a BP neural network is designed to predict and compensate the master-slave following error, further reducing the master-slave following error. The present invention can not only realize the large-range movement of the slave robot in free space, but also ensure the movement positioning accuracy in the fine operation space, with strong versatility and convenient implementation.

Description

Technical Field

[0001] The present invention belongs to the field of teleoperation robot control, and particularly relates to a master-slave control method and system for a master-slave heterogeneous teleoperation system. Background Art

[0002] With the rapid development of robot technology, robots have been widely used in various fields of industrial production and daily life. However, in complex, changeable and dangerous application scenarios, such as nuclear power plants, chemical plants, underwater operations and aviation operation environments, due to the harsh working environment or endangerment to personnel safety, manual operation methods cannot be adopted, and due to the unstructured working site environment and limited robot intelligence level, full-autonomous operation methods of robots cannot be adopted. Teleoperation can precisely combine human intelligence with robots. The operator generates motion control signals by manipulating the master robot, transmits them to the slave robot through a communication network, and the slave robot moves and operates in a complex and dangerous environment according to the received instructions, and at the same time feeds back the working state, and the operator makes judgments and decisions based on the feedback information. Teleoperation not only protects the health and safety of personnel, but also expands the operation ability of people and improves the operation accuracy. At present, teleoperation technology has been widely applied in the nuclear industry, medical surgery, aviation field and marine field.

[0003] Teleoperation systems are divided into two configurations: master-slave homogeneous and master-slave heterogeneous according to whether the mechanical structures of the master robot and the slave robot are the same. Since the master robot and the slave robot of the master-slave heterogeneous teleoperation system are both independently designed, they have various forms and strong versatility, and have been widely applied. A key problem of the master-slave heterogeneous teleoperation system is that the working spaces of the master robot and the slave robot are inconsistent, so it is necessary to design a suitable master-slave mapping method to achieve master-slave space matching.

[0004] Common master-slave mapping methods for master-slave heterogeneous teleoperation systems include absolute mapping, incremental mapping, constant ratio mapping, variable ratio mapping, etc. Absolute mapping requires accurate modeling of the working spaces of the master robot and the slave robot to solve the mapping function. However, the shapes and sizes of the working spaces of the master and slave robots are quite different, and the modeling of the working space mapping function is complex, making it difficult to ensure mapping accuracy. The incremental mapping method does not need to establish the mapping relationship of the working spaces of the master robot and the slave robot, but establishes the mapping relationship between the master position increment and the slave position increment. Therefore, the mapping function can be paused at any time to adjust the position of the master robot, and the operation flexibility is strong. Constant ratio mapping maps the master increment to the slave increment through a fixed ratio coefficient, which is difficult to adapt to the motion efficiency and accuracy of different scenarios. Variable ratio mapping is that the operator sets the mapping ratio coefficient according to the feedback information from the slave end to achieve fast motion or fine motion, and the operation flexibility is strong.

[0005] When the slave robot avoids obstacles and performs operations, if the following error of the slave robot is large and the positioning accuracy is low, it is very easy to have a violent collision with obstacles or operation targets, resulting in safety accidents. Therefore, it is necessary to reduce the master-slave following error and ensure the high-precision end positioning of the slave robot. In addition, through a large number of experimental data, it is found that the following error of the slave robot at the next moment is closely related to the end motion speed, following error, and position increment at the current moment. Summary of the Invention

[0006] The purpose of the present invention is to provide a master-slave control method for a master-slave heterogeneous teleoperation system, which realizes the mapping from the small-range workspace of the master end to the large-range workspace of the slave end, can not only ensure the flexibility of the large-range movement of the slave robot in the free space, but also reduce the master-slave following error and ensure the accurate positioning during fine operations.

[0007] The present invention adopts the following technical solutions: A master-slave control method for a master-slave heterogeneous teleoperation system, which is applied to a master-slave heterogeneous teleoperation system. The system includes a master robot and a slave robot with different mechanical structures, and a master-slave control system connecting the master robot and the slave robot. The master-slave control method includes the following steps:

[0008] (1) Kinematic modeling of the slave robot; According to the link parameters of the slave robot, forward kinematics and inverse kinematics analysis are carried out to obtain the forward kinematics and inverse kinematics models;

[0009] (2) Master-slave position control; Obtain the position increment of the master robot, calculate the expected position of the slave robot through the variable ratio incremental master-slave mapping method; and input the end movement speed, following error, and position increment of the slave robot into the BP neural network to output the predicted following error, and compensate the predicted error to the expected position; then input the compensated expected position into the inverse kinematics model of the slave robot to solve the expected joint angles, and finally realize joint angle control through the PD model;

[0010] (3) Master-slave adjustment; During the master-slave position control process, the opening and stopping of the master-slave position control function are detected in real time. If the function is enabled, return to step (2); When master-slave adjustment is required, it is necessary to first pause the master-slave position control function, and the slave robot stops moving. The master-slave adjustment includes modifying the master-slave mapping ratio coefficient and / or adjusting the position of the master robot; The specific modification of the master-slave mapping ratio coefficient is that the operator realizes the movement of the slave robot with different efficiencies and accuracies by modifying the master-slave mapping ratio coefficient.

[0011] Further, step (2) specifically includes:

[0012] Obtain the end position P m (t) of the master robot at the current moment t, and the end position Pm Obtain the end - position increment ΔP at the current moment t from (t - 1). m (t):

[0013] ΔP m (t)=P m (t)-P m (t - 1)

[0014] Map the ΔP of the master robot from time (t - 1) to t m (t) through variable - ratio incremental mapping to obtain the position increment ΔP of the slave robot from time t to (t + 1). s (t + 1)=KΔP m (t), where K is a 3×3 diagonal coefficient matrix for master - slave mapping; then add ΔP s (t + 1) to the desired position of the slave robot at the current moment to obtain the desired position at the next moment

[0015]

[0016] Input the joint angles of the slave robot at the current moment into the forward - kinematics model to solve for the end - position of the slave robot:

[0017] P s (t)=FK(θ s (t))

[0018] where P s (t) and θ s (t) represent the end - position and joint angles of the slave robot at the current moment respectively, and FK(·) represents the forward - kinematics solution function of the slave robot; from P s (t) and obtain the master - slave following error e(t) of the slave robot within the sampling time:

[0019]

[0020] Compensate the desired position of the slave robot with this error and the following error predicted by the BP neural network, then the calculation formula can be corrected to:

[0021]

[0022] ΔP s (t + 1)=ΔP m (t)-K -1 e(t)-K -1 e pred

[0023] where epred is the following error predicted by the BP neural network, and is input into the inverse kinematics model of the slave robot to solve for the desired joint angles:

[0024]

[0025] where represents the desired joint angles of the slave robot, and IK(·) represents the inverse kinematics solution function of the slave robot; finally, the control of the slave robot joint angles is achieved through the PD model.

[0026] Furthermore, the BP neural network prediction of the following error specifically includes:

[0027] A three-layer BP neural network model is established, including an input layer, a hidden layer, and an output layer; the inputs of the BP neural network are the movement speeds, following errors, and position increments of the end of the slave robot along the X, Y, and Z axes at the current moment. Therefore, the input layer has 9 neurons; the predicted output of the network is the master-slave following error at the next moment. Therefore, the output layer has 3 neurons; the number of neurons in the hidden layer is 20; the neuron activation function is the ReLu function, the optimization function uses Adam, and the hyperparameters are all the default parameters of the sklearn function library;

[0028] In the case of no error prediction and compensation by the BP neural network, the operator manipulates the master robot to move at different master-slave mapping ratio coefficients and speeds, controls the movement of the slave robot through the variable ratio incremental master-slave mapping method, obtains training data, and inputs the training data into the BP neural network for training to fit the regression function;

[0029] During the teleoperation process, the movement speeds, following errors, and position increments of the end of the slave robot along the X, Y, and Z axes at the current moment are input into the trained BP neural network to predict the following error e at the next moment pred and compensated into the desired position.

[0030] On the other hand, the present invention also provides a master-slave control system for a master-slave heterogeneous teleoperation system, including a slave robot kinematic modeling module, a master-slave mapping ratio coefficient modification module, a master-slave adjustment module, and a master-slave position control module;

[0031] The slave robot kinematic modeling module performs forward kinematics and inverse kinematics analyses based on the link parameters of the slave robot to obtain forward kinematics and inverse kinematics models, and can calculate the end position of the slave robot according to the joint angles of the slave robot, or calculate the joint angles according to the end position;

[0032] The master-slave mapping ratio coefficient modification module allows an operator to modify the master-slave mapping coefficient based on the feedback of the slave robot's motion state. In free space, the ratio coefficient is increased to achieve faster movement over a larger range. When approaching an obstacle or performing an operation, the ratio coefficient is decreased to achieve finer movement over a smaller range and ensure the end-point positioning accuracy.

[0033] The master-slave adjustment module is used to adjust the position of the master robot and / or modify the master-slave motion mapping ratio coefficient through the master-slave mapping ratio coefficient modification module after stopping the master-slave position control function, and then restart the master-slave position control function to continue remotely operating the motion of the slave robot.

[0034] The master-slave position control module first uses the variable-ratio incremental master-slave mapping method to map the position increment of the master robot to the position increment of the slave robot, and then calculates the expected position of the slave robot. Then, it uses a BP neural network to predict the following error and compensate it to the expected position. Next, it obtains the expected joint angles through the slave robot kinematic modeling module. Finally, it uses a PD model to achieve joint angle control.

[0035] Compared with the prior art, the beneficial effects brought by the technical solution of the present invention are as follows:

[0036] 1. The present invention realizes the mapping from the small working space of the master robot to the large working space of the slave robot through incremental position control, and realizes the artificial adjustment of the motion efficiency and accuracy of the slave robot through variable-ratio control, improving the operation flexibility.

[0037] 2. The present invention feeds back the master-slave following error of the sampling time to reduce the following error, and further reduces the master-slave following error through the error prediction and compensation method based on the BP neural network, ensuring the positioning accuracy during fine operation.

[0038] 3. The present invention is easy to implement and has strong versatility, ensuring the operator's comfort, and can simultaneously meet the fast movement in free space and the precise positioning in the fine operation space. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic structural diagram of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention;

[0040] Figure 2 It is a schematic diagram of the master robot of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention;

[0041] Figure 3 It is a schematic diagram of the slave robot of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention;

[0042] Figure 4The flowchart of the master-slave control method for the master-slave heterogeneous teleoperation system according to the embodiment of the present invention.

[0043] Figure 5 The effect diagram of the master-slave control method for the master-slave heterogeneous teleoperation system according to the embodiment of the present invention. Specific embodiments

[0044] The feasibility of the implementation of the present invention and the specific details of the technical solution will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments described herein are only for explaining the present invention and do not limit the present invention.

[0045] Please refer to Figure 1 , Figure 1 which is the structural schematic diagram of the master-slave heterogeneous teleoperation system according to the embodiment of the present invention. A typical teleoperation system includes five parts: an operator, a master robot, a master-slave control system, a slave robot, and an operating environment. In the master-slave heterogeneous teleoperation system according to the embodiment of the present invention, the master robot adopts a Sigma.7 hand controller, the slave robot adopts a UR5 robotic arm, the master-slave control system uses a PC as a development platform, adopts a control method of upper and lower computers, and designs the master-slave control software based on the open-source robot operating system ROS framework. The upper computer of the master-slave control system communicates with the Sigma.7 force feedback hand controller through the USB interface and bus of the PC, receives the motion state of the Sigma.7 force feedback hand controller and inputs it to the master-slave control method for processing, outputs the motion control instruction of the UR5 robotic arm, and sends it through the Ethernet; after receiving the motion control instruction from the upper computer, the lower computer of the master-slave control system drives the UR5 robotic arm to complete the corresponding action; the lower computer sends the motion state of the UR5 robotic arm to the upper computer in real time.

[0046] Please refer to Figure 2 , Figure 2 which is the schematic diagram of the master robot of the master-slave heterogeneous teleoperation system according to the embodiment of the present invention. The Sigma.7 hand controller has a composite structure composed of a Delta parallel structure, an interactive series structure, and a clamping structure, and has seven degrees of freedom, and the parallel structure is used to achieve position control. The master device in the embodiment of the present invention can also be replaced by a Phantom Omni device of SensAble Technology Company or an Omega series product of ForceDimension Company in other embodiments.

[0047] Please refer to Figure 3 , Figure 3It is a schematic diagram of the slave robot of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention. UR5 is a six-degree-of-freedom robotic arm. A link coordinate system is modeled for it, and the link parameters shown in Table 1 are listed. The link parameters are input into the master-slave control method for forward kinematics and inverse kinematics analysis to obtain forward kinematics and inverse kinematics models.

[0048] Table 1. Link parameters of the slave robot of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention

[0049] Joint Connecting rod length / mm Connecting rod torsion angle / rad Connecting rod wheelbase / mm Joint angle / rad 1 0 π / 2 89.2 <![CDATA[θ1]]> 2 -425 0 0 <![CDATA[θ2]]> 3 -392 0 0 <![CDATA[θ3]]> 4 0 π / 2 109.3 <![CDATA[θ4]]> 5 0 π / 2 94.75 <![CDATA[θ5]]> 6 0 0 82.5 <![CDATA[θ6]]>

[0050] For this embodiment, the master-slave motion mapping essentially controls the end of the slave UR5 robotic arm by operating the master Sigma.7 hand controller, so that the position of the end of the UR5 robotic arm follows the movement of the Sigma.7 hand controller.

[0051] Please refer to Figure 4 , Figure 4 It is a flowchart of the master-slave control method of the master-slave heterogeneous teleoperation system according to an embodiment of the present invention. The specific steps of the master-slave control method implemented in the present invention are as follows:

[0052] Step 1: Input the link parameters of the slave robot shown in Table 1 for forward kinematics and inverse kinematics analysis to obtain the forward kinematics and inverse kinematics models of the slave robot.

[0053] Step 2: Turn on the master-slave position control function:

[0054] The operator manipulates the master robot to move, obtains the end position P m (t) of the master robot at the current moment t, and obtains the end position increment ΔP m (t) at the current moment t by comparing it with the end position P m (t - 1) at the previous moment (t - 1):

[0055] ΔP m (t) = P m (t) - P m (t - 1)

[0056] The ΔP m (t) of the master robot from time (t - 1) to t is mapped to the position increment ΔP s (t + 1) of the slave robot from time t to (t + 1) through variable ratio incremental mapping, where K is a 3×3 master-slave mapping diagonal coefficient matrix. Then, ΔP m (t + 1) is superimposed on the expected position of the slave robot at the current moment s to obtain the expected position at the next moment ​

[0057]

[0058] The joint angles of the slave robot at the current moment are input into the forward kinematics model to solve the end position of the slave robot:

[0059] P s (t) = FK(θ s (t))

[0060] where P s (t) and θ s (t) represent the end position and joint angles of the slave robot at the current moment respectively, and FK(·) represents the forward kinematics solution function of the slave robot.

[0061] From P s (t) and the master-slave following error e(t) of the slave robot within the sampling time can be obtained:

[0062]

[0063] Compensating this error for the desired position of the slave robot, then the calculation formula can be corrected to:

[0064]

[0065] ΔP s (t + 1) = ΔP m (t) - K -1 e(t)

[0066] A three-layer BP neural network model is established, including an input layer, a hidden layer and an output layer. The input of the BP neural network is the movement speeds along the X, Y, and Z axes, the following error, and the position increment of the end of the slave robot at the current moment. Therefore, the input layer has 9 neurons; the predicted output of the network is the master-slave following error at the next moment. Therefore, the output layer has 3 neurons; the number of neurons in the hidden layer is 20. The neuron activation function is the ReLu function, the optimization function uses Adam, and the hyperparameters are all the default parameters of the sklearn function library.

[0067] In the case of no error prediction and compensation by the BP neural network, the operator manipulates the master robot to move at different master-slave mapping ratio coefficients and speeds, controls the movement of the slave robot through the variable ratio incremental master-slave mapping method, obtains training data, and inputs the training data into the BP neural network for training to fit the regression function.

[0068] During the teleoperation process, the motion speeds, following errors, and position increments of the end-effector of the master robot along the X, Y, and Z axes at the current moment are input into the trained BP neural network to predict the following error e at the next moment. pred And it is compensated into the desired position:

[0069]

[0070] ΔP s (t + 1) = ΔP m (t) - K -1 e(t) - K -1 e pred

[0071] The is input into the inverse kinematic model of the slave robot to solve for the desired joint angles:

[0072]

[0073] Where represents the desired joint angles of the slave robot, and IK(·) represents the inverse kinematic solution function of the slave robot. Finally, the control of the slave robot joint angles is achieved through the PD model.

[0074] Step 3: When the master robot reaches the limit of the workspace, pause the master-slave position control function, the slave robot stops moving, and adjust the position of the master robot for convenient operation. The master-slave mapping ratio coefficient can also be modified to achieve different motion speeds. After completing the above operations, restart the master-slave position control function.

[0075] The present invention proposes a master-slave control method applicable to a master-slave heterogeneous teleoperation system. Through incremental position control, the mapping from the small workspace of the master robot to the large workspace of the slave robot is realized. Through variable ratio control, the motion efficiency and accuracy of the slave robot can be adjusted manually, improving the operation flexibility. Through the error prediction and compensation method based on the BP neural network, the master-slave following error is reduced, ensuring the positioning accuracy during fine operations.

[0076] Please refer to Figure 5 , Figure 5 which is the effect diagram of the master-slave control method of the master-slave heterogeneous teleoperation system in the embodiment of the present invention. Among them, Figure 5 (a) in it is the diagram without BP neural network prediction error compensation, Figure 5Among them, (b) is the BP neural network prediction error compensation diagram. The master robot is manipulated to remotely operate the slave robot to move twice. The two manipulation speeds are nearly equal, and the data sampling time is 0.01 s. BP neural network prediction error compensation is not used once, and BP neural network prediction error compensation is used the other time. By comparing the two diagrams, it can be seen that the BP neural network prediction compensation reduces the master-slave following error and improves the motion and positioning accuracy of the end of the slave manipulator.

[0077] On the other hand, the present invention also provides a master-slave control system for a master-slave heterogeneous teleoperation system, including a slave robot kinematic modeling module, a master-slave mapping proportionality coefficient modification module, a master-slave adjustment module, and a master-slave position control module;

[0078] The slave robot kinematic modeling module performs forward kinematics and inverse kinematics analysis based on the link parameters of the slave robot to obtain forward kinematics and inverse kinematics models, and can calculate the end position of the slave robot according to the joint angles of the slave robot, or calculate the joint angles according to the end position; the specific implementation process of this module refers to the implementation steps of a master-slave control method for a master-slave heterogeneous teleoperation system.

[0079] The master-slave mapping proportionality coefficient modification module is for the operator to modify the master-slave mapping coefficient according to the feedback motion state of the slave robot, increase the proportionality coefficient in the free space to achieve faster movement in a larger range; when approaching an obstacle or performing an operation, reduce the proportionality coefficient to achieve finer movement in a smaller range and ensure the end positioning accuracy; the specific implementation process of this module refers to the implementation steps of a master-slave control method for a master-slave heterogeneous teleoperation system.

[0080] The master-slave adjustment module is used to adjust the position of the master robot and modify the master-slave motion mapping proportionality coefficient through the master-slave mapping proportionality coefficient modification module after stopping the master-slave position control function, and then restart the master-slave position control function to continue remotely operating the slave robot to move; the specific implementation process of this module refers to the implementation steps of a master-slave control method for a master-slave heterogeneous teleoperation system.

[0081] The master-slave position control module first uses the variable ratio incremental master-slave mapping method to map the position increment of the master robot to the position increment of the slave robot, and then calculates the expected position of the slave robot; then uses the BP neural network to predict the following error and compensate it to the expected position, and then obtains the expected joint angles through the slave robot kinematic modeling module, and finally uses the PD model to achieve joint angle control. The specific implementation process of this module refers to the implementation steps of a master-slave control method for a master-slave heterogeneous teleoperation system.

[0082] The above embodiments are used to explain the present invention rather than limit the present invention. Any modifications and changes made to the present invention within the spirit and scope of the claims of the present invention fall within the protection scope of the present invention.

Claims

1. A master-slave control method for a master-slave heterogeneous teleoperation system, characterized in that, This method is applied to a master-slave heterogeneous teleoperation system, which includes a master robot and a slave robot with different mechanical structures, and a master-slave control system connecting the master robot and the slave robot. The master-slave control method includes the following steps: (1) Kinematic modeling of the slave robot; according to the link parameters of the slave robot, forward kinematic and inverse kinematic analyses are carried out to obtain forward kinematic and inverse kinematic models; (2) Master-slave position control; obtain the position increment of the master robot, calculate the desired position of the slave robot through the variable ratio incremental master-slave mapping method; and input the end movement speed, following error and position increment of the slave robot into the BP neural network to output the predicted following error, and compensate the predicted error to the desired position; then input the compensated desired position into the inverse kinematic model of the slave robot to solve the desired joint angles, and finally realize joint angle control through the PD model. The specific process is as follows: Obtain the end position \(P\) of the master robot at the current moment \(t\) m (t), and the end position \(P\) at the previous moment \((t - 1)\) m (t - 1) to obtain the end position increment \(\Delta P\) at the current moment \(t\) m (t): ΔP m (t) = P m (t) - P m (t - 1) The ΔP of the master robot from time (t - 1) to t m (t) is mapped by variable - ratio incremental mapping to the position increment ΔP s (t + 1) of the slave robot from time t to (t + 1) = KΔP m (t), where K is a 3×3 diagonal coefficient matrix of master - slave mapping; then ΔP s (t + 1) is superimposed on the expected position of the slave robot at the current moment to obtain the expected position at the next moment Input the joint angles of the slave robot at the current moment into the forward kinematic model to solve the end position of the slave robot: P s F(θ s (t)) Where P s (t) and θ s (t) represent the end - effector position and joint angles of the slave robot at the current moment respectively, and FK(·) represents the forward kinematics solution function of the slave robot; from P s (t) and the master - slave following error e(t) of the slave robot within the sampling time is obtained: Compensate this error and the following error predicted by the BP neural network from the desired position of the end robot, then The calculation formula can be corrected to: ΔP s (t + 1)=ΔP m (t)-K -1 e(t)-K -1 e pred where e pred is the following error predicted by the BP neural network, and is input into the inverse kinematics model of the slave robot to solve for the desired joint angles: Among them represents the desired joint angle of the slave robot, and IK(·) represents the inverse kinematics solution function of the slave robot; finally, the control of the joint angle of the slave robot is achieved through the PD model; The BP neural network prediction of the following error specifically includes: Establish a three-layer BP neural network model, including an input layer, a hidden layer and an output layer; the input of the BP neural network is the movement speed, following error and position increment of the end of the slave robot along the X, Y and Z axes at the current moment, so the input layer has 9 neurons; the predicted output of the network is the master-slave following error at the next moment, so the output layer has 3 neurons; the number of neurons in the hidden layer is 20; the neuron activation function is the ReLu function, the optimization function adopts Adam, and the hyperparameters are all the default parameters of the sklearn function library; In the case of no error prediction and compensation by the BP neural network, the operator manipulates the master robot to move at different master-slave mapping ratio coefficients and speeds, controls the movement of the slave robot through the variable ratio incremental master-slave mapping method, obtains training data, and inputs the training data into the BP neural network for training to fit the regression function; During the teleoperation process, the motion speeds, following errors, and position increments of the end of the slave robot along the X, Y, and Z axes at the current moment are input into the trained BP neural network to predict the following error e at the next moment pred and compensated into the desired position; (3) Master-slave adjustment; during the master-slave position control process, the start and stop of the master-slave position control function are detected in real time. If the function is started, return to step (2); when master-slave adjustment is required, the master-slave position control function needs to be paused first, and the slave robot stops moving. The master-slave adjustment includes modifying the master-slave mapping ratio coefficient and / or adjusting the position of the master robot; the specific modification of the master-slave mapping ratio coefficient is that the operator realizes the movement of the slave robot with different efficiencies and precisions by modifying the master-slave mapping ratio coefficient.

2. A master-slave control system of a master-slave heterogeneous teleoperation system for implementing the master-slave control method described in claim 1, characterized in that, It includes a kinematic modeling module of the slave robot, a master-slave mapping ratio coefficient modification module, a master-slave adjustment module and a master-slave position control module; The kinematic modeling module of the slave robot is based on the link parameters of the slave robot, conducts forward kinematic and inverse kinematic analyses to obtain forward kinematic and inverse kinematic models, and can calculate its end position according to the joint angles of the slave robot, or calculate the joint angles according to the end position; The master-slave mapping ratio coefficient modification module allows the operator to modify the master-slave mapping coefficient based on the feedback of the slave robot's motion state. The ratio coefficient is increased in free space to achieve faster movement over a larger range; when approaching an obstacle or performing an operation, the ratio coefficient is decreased to achieve finer movement over a smaller range and ensure the end-effector positioning accuracy. The master-slave adjustment module is used to adjust the position of the master robot and / or modify the master-slave motion mapping ratio coefficient through the master-slave mapping ratio coefficient modification module after stopping the master-slave position control function, and then restart the master-slave position control function to continue the teleoperation of the slave robot's movement. The master-slave position control module first uses the variable ratio incremental master-slave mapping method to map the position increment of the master robot to the position increment of the slave robot, and then calculates the desired position of the slave robot; then uses a BP neural network to predict the following error and compensate it to the desired position, and then obtains the desired joint angles through the kinematic modeling module of the slave robot. Finally, the PD model is used to achieve joint angle control.

Images