Teleoperated robot control method based on event triggering and finite-time control

Abstract

This invention relates to a teleoperated robot control method based on event triggering and finite-time control, solving the problem of poor stability of teleoperated robots with asymmetric variable time delay, achieving finite-time position tracking convergence, and saving network communication resources. The method includes: sending the master-end position signal and the slave-end position signal via a zero-order hold to the master-end finite-time controller to calculate the control torque, which is then sent to the master-end robot to control its movement; sending the master-end position signal to the master-end event triggering decision-maker, which determines whether to use the communication network to transmit the current master-end position information based on the master-end event triggering strategy; sending the slave-end position signal and the master-end position signal via a zero-order hold to the slave-end finite-time controller to calculate the control torque, which is then sent to the slave-end robot to control its movement; and sending the slave-end robot position signal to the slave-end event triggering decision-maker, which determines whether to use the communication network to transmit the current slave-end position information based on the slave-end event triggering strategy.

Description

Technical Field

[0001] This invention relates to the field of control technology for teleoperated robots, and in particular to a control method for teleoperated robots based on event triggering and finite-time control. Background Technology

[0002] Teleoperated robots are a typical human-machine interface system, allowing operators to leverage the robot's high precision and consistency to complete numerous complex tasks. In many engineering applications, robots have already assisted or even replaced humans in performing tasks across various scenarios. However, in some constantly changing, complex, and hazardous work environments, remote operating systems remain necessary, enabling operators to perform tasks remotely. Currently, teleoperated robot systems are applied in many engineering fields, such as aerospace, deep-sea exploration, medical surgery, industrial operations, and inspection and operation in hazardous environments.

[0003] A teleoperated system can be considered as an interconnected system between two parts: a master robot and a master controller, and a slave robot and a slave controller. The operator applies forces to the master robot to move it, while the slave robot interacts with the environment and completes the task. Two-sided control is currently an important control architecture for teleoperated robots. For two-sided teleoperated robot systems, the master and slave controllers need to transmit status information to each other via a communication channel. Position tracking performance is an important evaluation criterion, which can be used to determine whether the slave robot can quickly and accurately reflect the operator's intentions, and whether the master robot can provide feedback on the slave robot's actions to the operator.

[0004] The transmission latency issue in the communication channels of teleoperated robots needs to be addressed to ensure the stability and position tracking performance of the closed-loop teleoperated robot system. On one hand, unavoidable communication latency reduces the stability of the closed-loop system; asymmetric latency can even lead to asynchrony in position tracking between the master and slave robots. On the other hand, data is packaged and exchanged within the communication network. Data transmission is discontinuous, involving packet loss and data congestion. Once a communication failure occurs, it will weaken the control performance of the teleoperated robot system and may even cause system instability. Summary of the Invention

[0005] The purpose of this invention is to provide a teleoperated robot control method based on event triggering and finite-time control to solve the problem of poor stability of asymmetric time-varying delay in bilateral teleoperated robot systems, and to achieve finite-time position tracking convergence, further simplifying the controller structure and saving data network communication resources.

[0006] To achieve the above objectives, this invention provides a teleoperated robot control method based on event triggering and finite-time control, the specific technical solution of which is as follows:

[0007] A teleoperated robot control method based on event-triggered communication and finite-time control includes the following steps:

[0008] On the main robot side:

[0009] The communication network acquires the slave position information uploaded to the communication network by the slave event trigger and sends the position information to the zero-order hold. If the slave event trigger does not upload the slave position information to the communication network, it means that no data transmission has occurred in the communication network at this time, and no new slave robot position signal has been sent to the zero-order hold.

[0010] The position information of the slave robot and the position information of the master robot output by the zero-order hold are sent to the master finite-time controller to calculate the control torque acting on the master robot;

[0011] The master robot moves under the control torque calculated by the master controller, thus obtaining the position information of the master robot.

[0012] The position information of the master robot is sent to the master event trigger. The master event trigger strategy in the master event trigger determines whether to send the current position information of the master robot into the communication network for transmission to the slave.

[0013] And / or, on the slave robot side:

[0014] The communication network acquires the master position information uploaded to the communication network by the master event trigger and sends the position information to the zero-order hold. If the master event trigger does not upload the master position information to the communication network, it means that no data transmission has occurred in the communication network at this time, and no new master robot position signal has been sent to the zero-order hold.

[0015] The position information of the master robot and the position information of the slave robot output by the zero-order hold are sent to the slave-end time controller to calculate the control torque applied to the slave robot.

[0016] The slave robot moves under the control torque calculated by the slave controller, and the position information of the slave robot is obtained.

[0017] The position information of the slave robot is sent to the slave event trigger. The slave event trigger strategy determines whether to send the current position information of the slave robot to the communication network for transmission to the master.

[0018] The beneficial effects of this invention are as follows:

[0019] On the one hand, the designed controller combines proportional terms, damping terms, and adaptive methods. Compared with existing related control methods, the designed controller has a simple structure, low gain, and is easy to implement in practice. On the other hand, the designed controller includes a novel event-triggered communication strategy with non-integer power terms to ensure that synchronization errors converge within a finite time. Verification shows that the control method of this invention can achieve stable position synchronization control of master-slave robots and has fast position convergence performance. At the same time, the amount of data transmitted in the communication channel is also reduced, saving data transmission volume in the communication network. Attached Figure Description

[0020] The accompanying drawings, which constitute a part of this invention, are used to assist in the description of the invention. The content provided in the drawings and their relevant descriptions in this invention can be used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0021] Figure 1 This is a schematic diagram illustrating the structure of a teleoperated robot control method based on event-triggered communication and finite-time control according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings of the embodiments of the present invention. Those skilled in the art will be able to implement the present invention based on these descriptions.

[0023] Furthermore, the embodiments of the present invention described below are generally only some, not all, of the embodiments of the present invention. Therefore, all other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention.

[0024] like Figure 1 As shown, the bilateral teleoperated robot system with time-varying delay in this embodiment is a nonlinear time-varying system, described as follows:

[0025] The model expression for the master robot is:

[0026]

[0027] The model expression for the slave robot is:

[0028]

[0029] Where the subscript m represents the master robot, the subscript s represents the slave robot, and q m The joint positions of the main robot. The joint speed of the master robot, The joint acceleration of the master robot, q s For the joint positions of the slave robot, For the joint speed of the end-user robot, For the joint acceleration of the slave robot, M m (q m M s (q s Let R(n×n) be the symmetric positive definite inertia matrices of the master robot and the slave robot, respectively. The centrifugal force and Coriolis force of the master robot and slave robot, respectively, G m (q m ), G s (q s )∈R n The gravity terms, F, are for the master robot teleoperation system and the slave robot teleoperation system, respectively. m F s ∈R n τ represents the frictional and external disturbance torques, respectively. m , τ s ∈R n The control torques τ of the master robot controller and the slave robot controller are respectively. h , τ e ∈R n These are the master force and the slave force, respectively. The master force is the interaction force between the operator and the master robot, and the slave force is the interaction force between the slave robot and the environment. n represents the robot's degrees of freedom.

[0030] The above-mentioned teleoperated robot control method based on event-triggered communication and finite-time control includes the following steps:

[0031] (1) On the master robot side:

[0032] The communication network acquires the slave position information uploaded to the communication network by the slave event trigger and sends the position information to the zero-order hold. If the slave event trigger does not upload the slave position information to the communication network, it means that no data transmission has occurred in the communication network at this time, and no new slave robot position signal has been sent to the zero-order hold.

[0033] Due to the zero-order hold, when new slave robot position information arrives from the communication network, the zero-order hold outputs the new slave robot position information; when no new slave robot position signal is sent, the zero-order hold outputs the original slave robot position information.

[0034] The position information of the slave robot and the position information of the master robot output by the zero-order hold are sent to the master finite-time controller to calculate the control torque acting on the master robot;

[0035] The master robot moves under the control torque calculated by the master controller, thus obtaining the position information of the master robot.

[0036] The position information of the master robot is sent to the master event trigger. The master event trigger strategy in the master event trigger determines whether to send the current position information of the master robot into the communication network for transmission to the slave.

[0037] (2) On the slave robot side:

[0038] The communication network acquires the master position information uploaded to the communication network by the master event trigger and sends the position information to the zero-order hold. If the master event trigger does not upload the master position information to the communication network, it means that no data transmission has occurred in the communication network at this time, and no new master robot position signal has been sent to the zero-order hold.

[0039] Due to the zero-order hold, when new master robot position information arrives from the communication network, the zero-order hold outputs the new master robot position information; when no new master robot position signal is sent, the zero-order hold outputs the original master robot position information.

[0040] The position information of the master robot and the position information of the slave robot output by the zero-order hold are sent to the slave-end time controller to calculate the control torque applied to the slave robot.

[0041] The slave robot moves under the control torque calculated by the slave controller, and the position information of the slave robot is obtained.

[0042] The position information of the slave robot is sent to the slave event trigger. The slave event trigger strategy determines whether to send the current position information of the slave robot to the communication network for transmission to the master.

[0043] The implementation process of the master robot's finite-time controller is described as follows:

[0044] The communication channel sends the signal to the master zero-order hold, and then the output signal is...

[0045] Combining the position and velocity signals of the master robot, the auxiliary variables of the master robot controller are constructed as follows:

[0046]

[0047] Where λ is the gain coefficient of the auxiliary variable of the main robot controller.

[0048] The finite-time controller for the master robot is:

[0049]

[0050] Where K represents the controller gain coefficient; α m The damping coefficient of the master controller is represented by σ, where 0 < σ < 1 is a non-integer power of the master controller. m The compensation term for the master robot controller is expressed as follows:

[0051]

[0052] Among them, G n (q m () is the gravity torque compensation term for the main robot; This represents the adaptive term in the RBF neural network; This represents the adaptive compensation term; The gain weight coefficient matrix of the RBF neural network for the master robot. The Gaussian function vector of the RBF neural network of the master robot. It can be obtained using the following expression:

[0053]

[0054]

[0055] Where l represents the number of intermediate nodes in the master-end RBF neural network, x m The input vector to the master-side RBF neural network, c mi The center vector of the i-th Gaussian function in the master-side RBF neural network, b m is the width of the Gaussian function.

[0056] They can be obtained respectively through the master-side controller learning law:

[0057]

[0058] Γ m1 ,Γ m2 The learning law coefficients of the master controller.

[0059] The controller of the master robot can be divided into three parts: the first part is -Ksig(s m ) σ The first part is mainly used for position error feedback control laws to achieve position tracking of the slave robot by the master robot; the second part is... Its main function is as a damping term, used to maintain the stability of the time-delay system; the third part is κ. m It is a system compensation item used to compensate for the robot's gravitational torque, frictional torque, and external torque.

[0060] The introduction of non-integer powers into the control law ensures the finite-time performance of the system's position control.

[0061] Whether the master position signal is sent to the communication network and transmitted to the slave robot is determined by the master event trigger.

[0062] The implementation process of the master robot event trigger is as follows:

[0063] The position signal output by the last master robot event trigger, after passing through a zero-order hold, is represented as: Combined with the current master robot position signal q sent to the master robot event trigger m The event triggering error e of the master robot can be calculated. mE (t) is

[0064]

[0065] When the main robot event triggers error e mE (t) satisfies the following event triggering strategy condition, expressed as:

[0066]

[0067] The master robot event trigger outputs signal q. m The location signal is then output to the communication network and transmitted to the slave side.

[0068] Regarding the event triggering strategy above, λ 12 , λ 22 >0 is the event trigger coefficient for the main robot. The event trigger coefficient for the main robot is expressed as:

[0069]

[0070] Where, k mE The event trigger weight of the main robot, d m This is the upper bound of the communication latency between the master robot and the slave robot.

[0071] The implementation process of the finite-time controller for the slave robot is described as follows:

[0072] The communication channel sends the signal to the slave-end zero-order hold, and then the output signal is...

[0073] Combining the position and velocity signals of the slave robot, construct the slave robot controller auxiliary variable as shown in the following expression:

[0074]

[0075] Where λ is the gain coefficient of the auxiliary variable of the slave robot controller.

[0076] The finite-time controller for the slave robot is:

[0077]

[0078] Where K represents the controller gain coefficient, which is the same as that in the master-side finite-time controller; α s The damping coefficient of the master controller is represented by σ, where 0 < σ < 1 is a non-integer power of the master controller. s The compensation term for the master robot controller is expressed as follows:

[0079]

[0080] Among them, G s (q s () is the gravity torque compensation term for the main robot; This represents the adaptive term of the RBF neural network in the slave robot. This represents the adaptive compensation term for the end-user robot. Here is the gain weight coefficient matrix of the RBF neural network for the slave robot. The Gaussian function vector is the RBF neural network vector of the slave robot. It can be obtained using the following formula:

[0081]

[0082]

[0083] Where l represents the number of intermediate nodes in the RBF neural network, x s Let c be the input vector of the slave-end RBF neural network. si Let b be the center vector of the i-th Gaussian function in the RBF neural network. s This is the width of the Gaussian function for the slave controller.

[0084] They can be obtained respectively through the following slave controller learning laws:

[0085]

[0086] Γ s1 ,Γ s2 The learning law coefficients are for the slave controller.

[0087] The controller of the slave robot can be divided into three parts: the first part is -Ksig(s s ) σThe first part is mainly used for position error feedback control laws to achieve position tracking of the slave robot to the master robot; the second part is... Its main function is as a damping term, used to maintain the stability of the time-delay system; the third part is κ. s It is a system compensation item used to compensate for the gravitational torque, frictional torque and external torque acting on the slave robot.

[0088] The introduction of non-integer powers into the control law ensures the finite-time performance of the system's position control.

[0089] Whether the master position signal is sent to the communication network and transmitted to the master robot is determined by the slave event trigger.

[0090] The process of implementing event triggers for the slave robot is as follows:

[0091] The position signal output from the last trigger of the slave robot event is represented by a zero-order hold as follows: Based on the current master robot position signal qs sent to the master robot event trigger, the master robot event triggering error e is calculated. sE (t) is

[0092]

[0093] When the main robot event triggers error e sE (t) satisfies the following event triggering strategy condition, expressed as:

[0094]

[0095] The slave robot's event trigger outputs signal qs, and this position signal is then transmitted to the master via the communication network. For the event triggering strategy above, λ2, λ 22 >0 represents the event trigger coefficient of the slave robot. The event triggering coefficient for the slave robot is expressed as:

[0096]

[0097] Where, k sE d is the trigger weight for the slave robot event. s This is the upper bound of the communication delay between the slave robot and the master robot.

[0098] The stability of the teleoperated robot system using the control method of this invention will be demonstrated and analyzed using the Lyapunov function.

[0099] The Lyapunov-Krasovskii functional is selected as:

[0100] V = V1 + V2 + V3 + V4 + V5

[0101]

[0102]

[0103]

[0104]

[0105]

[0106] Taking its derivative and substituting it into the control and adaptive laws of the master-slave robot, we can obtain...

[0107] Considering the need to satisfy the time-triggered strategy in event triggers, we can further obtain...

[0108]

[0109] in

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] Integrating the above inequality, we can obtain

[0116]

[0117] Furthermore, we can obtain q m -q s e mE e sE , and Both are bounded.

[0118] Based on the error definition triggered by the event, we can obtain

[0119]

[0120]

[0121] Therefore, it is possible to obtain further and Both are bounded.

[0122] Based on the dynamic model of the teleoperated robot system, we can obtain

[0123]

[0124] therefore, It is bounded; using Barbalat's corollary, we have...

[0125] Considering the event triggering error, its derivative is:

[0126]

[0127] Using the definition of the derivative, we have the following formula.

[0128]

[0129] Further, we can obtain

[0130]

[0131] Therefore, there is no condition for an event to be triggered an infinite number of times.

[0132] Finally, based on the above analysis, the finite-time tracking performance of the system can be demonstrated.

[0133] Construct the Lyapunov function as follows:

[0134]

[0135] in,

[0136] Substituting the control law and the adaptive learning law, and combining them with the scaling relations of the inequalities, we can obtain...

[0137]

[0138] in,

[0139]

[0140] Ξ=ξ m1 +ξ m2 +ξ s1 +ξ s2

[0141] Where, ξ m1 ξ m2 ξ s1 and ξ s2 All are positive numbers and satisfy the following conditions:

[0142]

[0143]

[0144]

[0145]

[0146] Thus, the position tracking of the closed-loop system exhibits the characteristic of finite-time convergence.

[0147] As can be seen, the control method of this embodiment can effectively realize the position tracking control of an asymmetric time-varying time-delay teleoperated robot system.

[0148] The foregoing has described the relevant content and embodiments of the present invention. Without departing from the design concept of the present invention, those skilled in the art can implement the present invention based on the above description. All other embodiments obtained by those skilled in the art based on the above content of the present invention without creative effort should fall within the scope of protection of the present invention.

Claims

1. A teleoperated robot control method based on event triggering and finite-time control, characterized in that, include: On the main robot side: The communication network acquires the slave position information uploaded to the communication network by the slave event trigger and sends the position information into the zero-order hold; The slave robot position information and master robot position information output from the zero-order hold are fed into the master finite-time controller to calculate the control torque acting on the master robot; the master finite-time controller is: in, This represents the controller gain coefficient. The joint speed of the master robot, The auxiliary variables of the constructed master robot controller are represented as follows: in, The output signal of the master zero-order hold The gain coefficient of the auxiliary variable of the main robot controller; This represents the damping coefficient of the master controller. The main controller is a non-integer power. The compensation term for the master robot controller is expressed as follows: in, The gravitational torque compensation term for the main robot; This represents the adaptive term in the RBF neural network; This represents the adaptive compensation term; The gain weight coefficient matrix of the RBF neural network for the master robot. The Gaussian function vector of the RBF neural network of the master robot; It can be obtained through the following formula: Where l represents the number of intermediate nodes in the master-end RBF neural network, x m The input vector to the master-side RBF neural network, c mi The center vector of the i-th Gaussian function in the master-side RBF neural network, b m The width of the Gaussian function; , The following are obtained using the master-side controller learning laws: , The learning law coefficients of the master controller; The master robot moves under the control torque calculated by the master finite-time controller, and the position information of the master robot is obtained. The position information of the master robot is sent to the master event trigger. The master event trigger strategy in the master event trigger determines whether to send the current position information of the master robot into the communication network for transmission to the slave. And / or, on the slave robot side: The communication network acquires the master location information uploaded to the communication network by the master event trigger and sends the location information into the zero-order hold; The position information of the master robot and the position information of the slave robot output by the zero-order hold are sent to the slave finite-time controller to calculate the control torque acting on the slave robot; The slave robot moves under the control torque calculated by the slave finite-time controller, and the position information of the slave robot is obtained. The position information of the slave robot is sent to the slave event trigger. The slave event trigger strategy determines whether to send the current position information of the slave robot to the communication network for transmission to the master.

2. The teleoperated robot control method based on event triggering and finite-time control according to claim 1, characterized in that: The model expression for the master robot is: The model expression for the slave robot is: Where the subscript m represents the master robot, the subscript s represents the slave robot, and q m The joint positions of the main robot. The joint acceleration of the master robot, For the joint positions of the slave robot, For the joint speed of the end-effector robot, For the joint acceleration of the slave robot, Let be the symmetric positive definite inertia matrices of the master robot and the slave robot, respectively. These are the centrifugal force and Coriolis force of the master robot and the slave robot, respectively. These are the gravity terms for the master robot teleoperation system and the slave robot teleoperation system, respectively. These represent frictional and external disturbance torques, respectively. These are the control torques of the master robot controller and the slave robot controller, respectively. These are the master force and the slave force, respectively. The master force is the interaction force between the operator and the master robot, and the slave force is the interaction force between the slave robot and the environment. n represents the robot's degrees of freedom.

3. The teleoperated robot control method based on event triggering and finite-time control according to claim 1, characterized in that: The master-side event trigger is: The position signal output from the last master robot event trigger, after passing through a zero-order hold, is represented as follows: Combined with the current master robot position signal q sent to the master robot event trigger m The event triggering error of the master robot was calculated. for: When the main robot event triggers an error The event triggering strategy conditions are expressed as follows: The master robot event trigger outputs signal q. m The location signal is then output to the communication network and transmitted to the slave end; for the above event triggering strategy, The event trigger coefficient for the main robot. The event trigger coefficient for the main robot is expressed as follows: in, The trigger weight of the main robot event. This is the upper bound of the communication latency between the master robot and the slave robot.

4. The teleoperated robot control method based on event triggering and finite-time control according to claim 2, characterized in that: The slave-end finite-time controller is: in, This represents the controller gain coefficient, which is the same as that in the master-side finite-time controller. The auxiliary variables of the constructed slave robot controller are represented as follows: in, The output signal of the zero-order hold is from the end. This is the gain coefficient of the auxiliary variable of the slave controller, which is the same as that of the master controller; This represents the damping coefficient of the slave controller. For slave controllers, the power of a non-integer power. The compensation term for the slave robot controller is expressed as follows: in, For the gravitational torque compensation term of the slave robot; This represents the adaptive term of the RBF neural network in the slave robot. This represents the adaptive compensation term for the end-user robot. Here is the gain weight coefficient matrix of the RBF neural network for the slave robot. The Gaussian function vector of the RBF neural network of the slave robot; It can be obtained through the following formula: Where l represents the number of intermediate nodes in the RBF neural network, x s Let c be the input vector of the slave-end RBF neural network. si Let b be the center vector of the i-th Gaussian function in the RBF neural network. s The width of the Gaussian function for the slave controller; , The following are obtained using the slave controller learning laws: , The learning law coefficients are for the slave controller.

5. The teleoperated robot control method based on event triggering and finite-time control according to claim 1, characterized in that: The slave event trigger is: The position signal output from the last trigger of the slave robot event is represented by a zero-order hold as follows: Combined with the current master robot position signal q sent to the master robot event trigger s The event triggering error of the master robot was calculated. for: When the main event triggers an error The event triggering strategy conditions are expressed as follows: Then the slave robot event trigger outputs signal q s The location signal is then output to the communication network and transmitted to the master terminal; for the above event triggering strategy, For the event trigger coefficient of the slave robot, The event triggering coefficient for the slave robot is expressed as follows: in, As the trigger weight for the slave robot event, This is the upper bound of the communication delay between the slave robot and the master robot.

Images