Multi-wheeled mobile robot formation control method

Through the dynamic event triggering mechanism and extended state observer, combined with the guidance law of the predetermined time tunnel performance function, the problem of insufficient transient overshoot and tracking error convergence performance in the multi-wheeled mobile robot formation is solved, and the efficient anti-interference performance of the system is achieved and error convergence effect.

CN118897546APending Publication Date: 2024-11-05HAINAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410935125.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-05

AI Technical Summary

Technical Problem

Traditional preset performance control methods have transient overshoot problems and insufficient tracking error convergence performance in multi-wheeled mobile robot formations, and anti-interference technology results in waste of resources based on time triggering.

Method used

The dynamic event trigger mechanism is used to build an extended state observer, and a preset performance function and guidance law of predetermined time tunnel are designed to obtain control law through error conversion technology to improve transient and steady-state performance.

Benefits of technology

The tracking error converges within the predefined time and improves the transient performance of the multi-wheeled mobile robot system. Through dynamic events, the coordinated controller of the extended state observer and the preset performance function of the predetermined time tunnel is ensured that the system is on the surface. Stable performance for internal uncertainty and external interference.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118897546A_ABST
    Figure CN118897546A_ABST
Patent Text Reader

Abstract

The invention discloses a multi-wheeled mobile robot formation control method, which comprises the following steps of: setting a spanning tree with a virtual leader robot as a root node and a plurality of follower robots as nodes, and establishing a model of wheeled mobile robots; based on a dynamic event triggering mechanism, a dynamic event triggering expansion state observer is constructed, and the dynamic event triggering expansion state observer is used for estimating internal parameter uncertainty and external interference of each model; in order to ensure that the plurality of follower robots are cooperatively controlled under the guidance of the virtual leader robot, a preset time tunnel preset performance function is designed; designing a guidance law based on a preset performance function of a preset time tunnel; and obtaining control laws of the plurality of wheeled mobile robots based on a guidance law of a preset performance function of a preset time tunnel and a dynamic event triggering expansion state observer. According to the method, it can be guaranteed that the tracking error converges within the predefined time, and the performance of the transient stage can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of robot formation control, and in particular to a multi-wheeled mobile robot formation control method. Background Art

[0002] In the past few decades, the cooperative control of wheeled mobile robots (WMRs) has attracted increasing attention from the scientific community because of its important application prospects in search, firefighting, agriculture and other fields. According to the task type and the number of leaders, the coordinated control can be further divided into consensus control, inclusion control, formation control and cluster control. WMR is a nonlinear multi-coupled system, so it is crucial to enhance the transient and steady-state performance of multi-WMR systems.

[0003] Preset performance control (PPC) is a common method to improve the robustness of system performance. So far, many research results based on PPC have emerged. For mobile robot systems, a global performance constraint control method is designed. For four-wheel independent drive skid-steer mobile robots, a sequentially prescribed time tracking controller is designed. For autonomous ground vehicles, a prescribed performance fault-tolerant controller is proposed. However, the above research results only consider the PPC of single-wheel mobile robots. In contrast, multi-WMR formations can complete some more complex tasks. At the same time, this will further test the robustness of system performance. Many research results on cooperative control based on PPC have been reported. For non-holonomic multi-mobile robot systems, a performance-guaranteed centralized leader-follower control strategy is proposed. In order to achieve rigid formation maneuvers, a fixed-time prescribed performance method is proposed. Although the above results can make the tracking error converge within the prescribed performance constraint range, there are still two points that need to be improved: 1) In the transient stage, the traditional PPC still causes the overshoot problem of the multi-WMR formation; 2) The traditional PPC does not consider the convergence performance of the tracking error.

[0004] When performing tasks, each WMR is affected by internal uncertainties and external disturbances. So far, many fruitful studies have been proposed in the control field to estimate these internal uncertainties and external disturbances. For example: nonlinear extended state observer (ESO), radial basis function neural network, echo state network predictor. However, these anti-interference techniques are based on time triggering, which will inevitably lead to the waste of transmission resources.

[0005] To solve the above problems, a multi-wheeled mobile robot formation control method based on dynamic event triggering mechanism and tunnel performance guarantee is proposed. Summary of the invention

[0006] In order to solve the above technical problems, the present invention proposes a multi-wheeled mobile robot formation control method, which can not only ensure that the tracking error converges within a predefined time, but also improve the performance in the transient stage.

[0007] In order to achieve the above object, the technical solution of the present invention is as follows:

[0008] A multi-wheeled mobile robot formation control method comprises the following steps:

[0009] Assume a spanning tree having a virtual leader robot as a root node and including a plurality of follower robots as nodes, wherein the virtual leader robot and the plurality of follower robots are all wheeled mobile robots, and establish a model of the wheeled mobile robot;

[0010] Based on the dynamic event triggering mechanism, a dynamic event triggered extended state observer is constructed, wherein the dynamic event triggered extended state observer is used to estimate the internal parameter uncertainty and external interference of each model;

[0011] In order to ensure the coordinated control of several follower robots under the guidance of the virtual leader robot, a preset performance function of the predetermined time tunnel is designed; through the error conversion technology, a guidance law based on the preset performance function of the predetermined time tunnel is designed;

[0012] Through anti-interference technology, the control law of the wheeled mobile robot is obtained based on the guidance law of the preset performance function of the predetermined time tunnel and the dynamic event-triggered extended state observer.

[0013] Preferably, the step of establishing a model of the wheeled mobile robot comprises the following:

[0014] Mathematical modeling is performed on several wheeled mobile robots, and the kinematic model of the i-th wheeled mobile robot is expressed as follows:

[0015]

[0016] According to the Newton-Euler formula, the dynamic model of the i-th wheeled mobile robot can be derived as follows:

[0017]

[0018] Among them, R i is a rotation matrix, η i =[x i ,y i ,ψ i ] T represents the position and yaw angle of the wheeled mobile robot, v i =[u i ,v i ,w i] T represents the surge velocity, sway velocity and yaw rate of the wheeled mobile robot, Indicates ν i The first derivative of τ i Represents the control input, M i represents the control input matrix, f i Represents the total disturbance consisting of internal parameter uncertainty and external disturbance.

[0019] Preferably, the dynamic event triggering extended state observer is constructed based on the dynamic event triggering mechanism, including the following:

[0020] Dynamic event triggering mechanism, the formula is as follows:

[0021]

[0022] in, Indicates v i Non-periodic sampling of Represents the velocity vector updated at the current time. and represents the current update time and the next update time. R represents a real number set, is the non-periodic sampling error, considering the following dynamic threshold Λ based on the distributed tracking error i , the formula is as follows:

[0023]

[0024] in and is a positive constant, E i is the distributed tracking error designed later, ||E i || represents the norm of the distributed tracking error. Then, the dynamic event-triggered extended state observer is designed as follows:

[0025]

[0026] Preferably, in order to ensure that a plurality of follower robots under the guidance of the virtual leader robot are cooperatively controlled, a preset performance function of a predetermined time tunnel is designed; and a guidance law based on the preset performance function of a predetermined time tunnel is designed through error conversion technology, including the following:

[0027] Define a distributed kinematic tracking error E i for:

[0028]

[0029] Among them, p ij =p ir -p jr, η r is a desired trajectory. If the i-th wheeled mobile robot can receive the position and heading information of the j-th wheeled mobile robot, then a ij =1; otherwise, a ij = 0; if the i-th wheeled mobile robot can receive the position and heading information of the virtual leader, a i0 =1, otherwise, a i0 =0;

[0030] By taking the derivative of (7), we can get

[0031]

[0032] in represents the communication between the i-th wheeled mobile robot and its neighbors and the virtual leader, N represents the number of followers, and Respectively represent η r and p ij In order to improve the transient and steady-state performance of the multi-wheeled mobile robot system, the following constraints are imposed on the tracking error:

[0033]

[0034] in, E ik and The preset performance function representing the predetermined time tunnel is defined as follows:

[0035]

[0036] in

[0037]

[0038] in, E ik,0 =E ik 0 is the initial error value, ik,0 =Ξ ik 0 and Ξ ik,∞ =lim t→∞ Ξ ik (t are initial value and steady-state value respectively, represents the convergence speed, T ik represents the expected convergence time of the tracking error; sgn(E ik,0 represents a symbolic function, when E ik,0 = 0, sgn(E ik,0 =0; when E ik,0 >0, sgn(E ik,0 =1; when E ik,0 <0, sgn(Eik,0 = -1;

[0039] Through error conversion, the converted error is represented as follows:

[0040]

[0041] Among them, Θ ik is an error transfer function, defined as

[0042] Taking the derivative of (12) we can get:

[0043]

[0044] The main variables in formula (12) are summarized as follows:

[0045]

[0046] For clarity, define and E i =[ E ix , E iy , E iψ ] T , then rewrite (13) into vector form:

[0047]

[0048] in and and and Substituting (8) into (14) yields:

[0049]

[0050] In order to stabilize the error dynamics of (15) and allow the distributed tracking error to converge to the predetermined limit at the predetermined time, a guidance law based on a preset performance function of the predetermined time tunnel is designed as follows:

[0051]

[0052] where k i1 represents the control gain of the guidance layer, Represents the transposed matrix of a rotation matrix.

[0053] Preferably, the guidance law based on the preset performance function of the predetermined time tunnel and the dynamic event-triggered extended state observer to obtain the control law of the wheeled mobile robot includes the following:

[0054] In order to obtain the dynamic control law, the velocity layer tracking error is defined as follows: ie =v i -α i , by ie The derivative is:

[0055]

[0056] in Represents α i The first derivative of

[0057] In order to stabilize the error dynamics of (17), a control law based on dynamic event-triggered extended state observer is proposed as follows:

[0058]

[0059] where k i2 Represents the control gain of the control layer.

[0060] Based on the above technical scheme, the beneficial effects of the present invention are as follows: the present invention establishes a preset time tunnel preset performance collaborative controller to achieve the desired formation effect. At the guidance layer, in order to enhance the transient performance and steady-state performance of the system, a preset time tunnel preset performance function (PTTPP) is designed. Through the error conversion technology, a guidance law based on PTTPP is designed to stabilize the distributed tracking error. At the control layer, in order to simultaneously deal with internal uncertainty and external interference, a dynamic event triggered extended state observer (DETESO) is designed. Through anti-interference technology, a control law based on DETESO is constructed, which can not only ensure that the tracking error converges within a predefined time, but also improve the performance of the transient stage. The effectiveness and reliability of the proposed method are verified through theoretical analysis and a multi-vehicle experimental platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a flow chart of a multi-wheeled mobile robot formation control method in an embodiment;

[0062] Figure 2 is a vehicle body-earth reference coordinate system in a multi-wheeled mobile robot formation control method in an embodiment;

[0063] Figure 3 is a three-vehicle trajectory diagram of a multi-wheeled mobile robot formation control method in an embodiment;

[0064] Figure 4is a three-vehicle control input diagram of a multi-wheeled mobile robot formation control method in an embodiment;

[0065] Figure 5 is interference information estimated based on DETESO in one embodiment;

[0066] Figure 6 is interference information estimated based on DETESO in one embodiment;

[0067] Figure 7 This is a comparison chart between DETESO and ETESO. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0069] like Figure 1 As shown, this embodiment provides a multi-wheeled mobile robot formation control method, the simulation software used is MATLAB2022, the operating environment is Windows, the experimental platform consists of a motion capture system, three Mecanum wheel cars and a zigbee network, and the specific processing process is as follows:

[0070] Assumption 1: The graph theory contains a spanning tree with a virtual leader as the root node.

[0071] Step 1: Mathematically model the wheeled mobile robot (WMR). The kinematics of the i-th WMR is expressed as follows:

[0072]

[0073] According to the Newton-Euler formula, the dynamics of the ith WMR can be derived as follows:

[0074]

[0075] Among them, R i is a rotation matrix. η i =[x i ,y i ,ψ i ] T Represents the position and yaw angle of WMR. i =[u i ,v i ,w i ] T Represents the surge speed, sway speed and yaw rate of the WMR. Indicates v i The first derivative of τ i Indicates control input. irepresents the control input matrix. i Represents the total disturbance consisting of internal parameter uncertainty and external disturbance.

[0076] Step 2: In order to estimate the overall disturbance, and considering that the nonlinear extended state observer (ESO) can estimate the external disturbance using only the speed information without the need for additional sensors, the traditional ESO is constructed as follows:

[0077]

[0078] Among them, ε i is a positive gain matrix. The variable and Respectively represent ν i and f i The estimated value of . and express and Next, in order to reduce the resource consumption from the sensor to the controller, the following dynamic event triggering mechanism is adopted:

[0079]

[0080] in, Indicates ν i Non-periodic sampling. Represents the velocity vector updated at the current time. and represents the current update time and the next update time. R represents a real number set. is the non-periodic sampling error. Consider the following dynamic threshold Λ based on the distributed tracking error i :

[0081]

[0082] in and is a positive constant. i is the distributed tracking error designed later. ||E i || represents the norm of the distributed tracking error. Then, the dynamic event-triggered extended state observer (DETESO) is designed as follows:

[0083]

[0084] Step 3: In order to realize the cooperative control under the guidance of the virtual leader, a guidance law based on the scheduled time tunnel preset performance function (PTTPP) and a control law based on the dynamic event triggered extended state observer (DETESO) are designed. The details are as follows:

[0085] Define a distributed tracking error E i for:

[0086]

[0087] Among them, p ij =p ir -p jr represents the relative position distance and heading distance between the i-th WMR and the j-th WMR. ir and p jr Respectively represent the position and heading distance between the i-th and j-th WMR and the virtual leader. η j represents the position and heading of the jth WMR. ij Indicates the communication between the i-th WMR and the j-th WMR. i0 Represents the communication between the i-th WMR and the virtual leader. r is a desired trajectory. If the i-th WMR can receive the position and heading information of the j-th WMR, then a ij =1; otherwise, a ij = 0. If the i-th WMR can receive the position and heading information of the virtual leader, a i0 =1, otherwise, a i0 =0.

[0088] By taking the derivative of (7), we can obtain:

[0089]

[0090] in represents the communication of the i-th WMR with its neighbors and the virtual leader, and N represents the number of followers. and Respectively represent η r and p ij In order to improve the transient and steady-state performance of the multi-WMR system, the following constraints are imposed on the tracking error:

[0091]

[0092] in, E ik and represents a preset performance function based on a predetermined time tunnel, which is defined as follows:

[0093]

[0094] in

[0095]

[0096] in, E ik,0 =E ik 0 is the initial error value. ik,0 =Ξ ik 0 and Ξ ik,∞ =lim t→∞ Ξ ik (t are the initial value and steady-state value, respectively. Indicates the convergence speed. T ik Indicates the expected convergence time of the tracking error. ik,0 represents a symbolic function, when E ik,0 = 0, sgn(E ik,0 =0; when E ik,0 >0, sgn(E ik,0 =1; when E ik,0 <0, sgn(E ik,0 =-1.

[0097] Through error conversion, the converted error is represented as follows:

[0098]

[0099] Among them, Θ ik is an error transfer function, defined as

[0100] Next, we can derive (12):

[0101]

[0102] in Indicates E ik , E ik The main variables in formula (13) are summarized as follows:

[0103]

[0104] For clarity, define and E i =[ E ix , E iy , E iψ ] T Then rewrite equation (13) into vector form:

[0105]

[0106] in and and and Substituting (8) into (14) yields:

[0107]

[0108] In order to stabilize the error dynamics of (15) and allow the distributed tracking error to converge to the predetermined limit within the predetermined time, a guidance law based on PTTPP is designed:

[0109]

[0110] where k i1 represents the control gain of the guidance layer. Represents the transposed matrix of a rotation matrix.

[0111] Step 4: In the following study, in order to obtain the dynamic control law, the velocity layer tracking error is defined as follows: ie =v i -α i . By v ie The derivative is:

[0112]

[0113] in Represents α i The first derivative of .

[0114] In order to stabilize the error dynamics of (17), a control law based on DETESO is proposed:

[0115]

[0116] where k i2 Represents the control gain of the control layer.

[0117] The experimental verification analysis proves that the entire closed-loop system including DETESO, the guidance law based on PTTPP and the control law based on DETESO is stable in input state and Zeno behavior can be avoided through cascade stability analysis and input state stability. The effectiveness and reliability of the proposed multi-vehicle platoon control method based on the tunnel performance guarantee of the dynamic event trigger mechanism are verified by experiments. The vehicle body-earth reference coordinate system (representing the distributed cooperative mode of N WMRs guided by a reference trajectory, such as Figure 2 As shown in the figure), the trajectory diagram (shows that under the proposed controller, the three vehicles can smoothly form the expected formation. And the performance in the transient stage is much better than that of the preset performance control method, such as Figure 3 ), control torque diagram (shows the actual control signals of the three WMRs, such as Figure 4 ), interference estimation diagram (describes the estimation performance of DETESO, such as Figure 5 As shown in the figure), the trigger frequency diagram (shows the trigger interval of the speed under the dynamic event trigger mechanism, such as Figure 6 As shown in the figure), and the comparison diagram (DETESO is compared with the event-triggered extended state observer (ETESO). It can be seen from the figure that the approximation effect of DETESO is better than that of ETESO, which reflects the superiority of DETESO. Figure 7 shown).

[0118] The above are only preferred embodiments of the embodiments of the present application and are not intended to limit the embodiments of the present application. For those skilled in the art, the embodiments of the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the embodiments of the present application shall be included in the protection scope of the embodiments of the present application.

Claims

1. A multi-wheeled mobile robot formation control method, characterized in that: The steps include: Assume a spanning tree having a virtual leader robot as a root node and including a plurality of follower robots as nodes, wherein the virtual leader robot and the plurality of follower robots are all wheeled mobile robots, and establish a model of the wheeled mobile robot; Based on the dynamic event triggering mechanism, a dynamic event triggered extended state observer is constructed, wherein the dynamic event triggered extended state observer is used to estimate the internal parameter uncertainty and external interference of each model; In order to ensure the coordinated control of several follower robots under the guidance of the virtual leader robot, a preset performance function of the predetermined time tunnel is designed; Through the error conversion technique, a guidance law based on a preset performance function of a predetermined time tunnel is designed; Through anti-interference technology, the control law of the wheeled mobile robot is obtained based on the guidance law of the preset performance function of the predetermined time tunnel and the dynamic event-triggered extended state observer.

2. A multi-wheeled mobile robot formation control method according to claim 1, characterized in that: The model of the wheeled mobile robot is established, including the following: Mathematical modeling is performed on several wheeled mobile robots, and the kinematic model of the i-th wheeled mobile robot is expressed as follows: According to the Newton-Euler formula, the dynamic model of the i-th wheeled mobile robot can be derived as follows: Among them, R i is a rotation matrix, η i =[x i ,y i ,ψ i ] T represents the position and yaw angle of the wheeled mobile robot, ν i =[u i ,v i ,w i ] T represents the surge velocity, sway velocity and yaw rate of the wheeled mobile robot, Indicates v i The first derivative of τ i Represents the control input, M i represents the control input matrix, f i Represents the total disturbance consisting of internal parameter uncertainty and external disturbance.

3. A multi-wheeled mobile robot formation control method according to claim 2, characterized in that: The dynamic event triggering extended state observer is constructed based on the dynamic event triggering mechanism, including the following: Dynamic event triggering mechanism, the formula is as follows: in, Indicates v i Non-periodic sampling of Represents the velocity vector updated at the current time. and represents the current update time and the next update time. R represents a real number set, is the non-periodic sampling error, considering the following dynamic threshold Λ based on the distributed tracking error i , the formula is as follows: in and is a positive constant, E i is the distributed tracking error designed later, ||E i || represents the norm of the distributed tracking error. Then, the dynamic event-triggered extended state observer is designed as follows:

4. A multi-wheeled mobile robot formation control method according to claim 3, characterized in that: In order to ensure the coordinated control of several follower robots under the guidance of the virtual leader robot, a preset performance function of a predetermined time tunnel is designed; through the error conversion technology, a guidance law based on the preset performance function of the predetermined time tunnel is designed, including the following: Define a distributed kinematic tracking error E i for: Among them, p ij =p ir -p jr , η r is a desired trajectory. If the i-th wheeled mobile robot can receive the position and heading information of the j-th wheeled mobile robot, then a ij =1; otherwise, a ij = 0; if the i-th wheeled mobile robot can receive the position and heading information of the virtual leader, a i0 =1, otherwise, a i0 =0; By taking the derivative of (7), we can get in represents the communication between the ith wheeled mobile robot and its neighbors and the virtual leader, N represents the number of followers, and Represent η and p respectively ij In order to improve the transient and steady-state performance of the multi-wheeled mobile robot system, the following constraints are imposed on the tracking error: in, E ik and The preset performance function representing the predetermined time tunnel is defined as follows: in in, E ik,0 =E ik (0) is the initial error value, ik,0 =Ξ ik (0) and Ξ ik,∞ =lim t→∞ Ξ ik (t) are initial value and steady-state value, respectively. represents the convergence speed, T ik represents the expected convergence time of the tracking error; sgn(E ik,0 ) represents the symbolic function, when E ik,0 = 0, sgn(E ik,0 )=0; when E ik,0 >0, sgn(E ik,0 )=1; when E ik,0 <0, sgn(E ik,0 )=-1; Through error conversion, the converted error is represented as follows: Among them, Θ ik is an error transfer function, defined as Taking the derivative of (12) we can get: The main variables in formula (12) are summarized as follows: For clarity, define and Then rewrite (13) into vector form: in and and and Substituting (8) into (14) yields: In order to stabilize the error dynamics of (15) and allow the distributed tracking error to converge to the predetermined limit at the predetermined time, a guidance law based on a preset performance function of the predetermined time tunnel is designed as follows: where k i1 represents the control gain of the guidance layer, Represents the transposed matrix of a rotation matrix.

5. A multi-wheeled mobile robot formation control method according to claim 4, characterized in that: The guidance law based on the preset performance function of the predetermined time tunnel and the dynamic event triggering extended state observer to obtain the control law of the wheeled mobile robot include the following: In order to obtain the dynamic control law, the velocity layer tracking error is defined as follows: ie =v i -α i , by ie The derivative is: in Represents α i The first derivative of In order to stabilize the error dynamics of (17), a control law based on dynamic event-triggered extended state observer is proposed as follows: where k i2 Represents the control gain of the control layer.