Pilot-following AGV formation control method with composite controller structure

By introducing a sliding mode controller and feedback controller in the AGV formation, combined with the integral sliding mode surface, the impact of the acceleration changes of the leader's vehicle on the formation stability is solved, and the stable operation of the AGV formation under communication delay is achieved.

CN119065370APending Publication Date: 2024-12-03DALIAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411144516.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

In multi-AGV formations, the prior art fails to fully consider the impact of the acceleration changes in the leader's vehicle on formation stability, especially in the presence of communication delays and other interferences, resulting in formation stability being affected.

Method used

A pioneer with a composite controller structure-following AGV formation control method is designed, and by establishing a vehicle longitudinal kinematic model, introducing a sliding mode controller and a feedback controller based on tracking error, combined with the integral sliding mode surface, the leader AGV is eliminated. Acceleration disturbances and keeps formation stable under communication delays.

Benefits of technology

It effectively reduces the impact of the AGV acceleration disturbance on the formation stability of the leader's team, ensures the stable operation of the AGV formation during communication delay and acceleration changes, and verifies its effectiveness through simulation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065370A_ABST
    Figure CN119065370A_ABST
Patent Text Reader

Abstract

The invention provides a navigator-following AGV formation control method with a composite controller structure. The method comprises the following steps: describing an information flow between AGV formations through a communication topology matrix; taking the AGV as a mass point, and establishing a vehicle longitudinal kinematics model; setting a control target which can still maintain stable operation under the influence of communication delay and leader AGV acceleration disturbance; in order to eliminate the disturbance of the acceleration of the leading AGV, a sliding mode controller is introduced; introducing a feedback controller based on a tracking error in consideration of communication delay between vehicles; and a composite controller ui (t) is established, so that the whole formation is kept stable when the system is subjected to communication delay and the acceleration disturbance of the leading AGV. The invention provides a leader-following formation mode, and communication delay and acceleration disturbance are considered.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of formation control, and in particular to a leader-follower AGV formation control method with a composite controller structure. Background Art

[0002] Automated warehouses and factories are becoming more and more popular as a solution to meet the growing demand for industrial production, and automated guided vehicles (AGVs) play a key role in unmanned management. AGVs are widely used to autonomously transport goods, move large equipment, or perform dangerous tasks in harsh environments [1]. The working capacity of a single AGV is very limited, and multi-AGV collaboration has become a research hotspot due to its advantages such as high efficiency and flexibility.

[0003] Multi-AGV collaboration requires AGVs to travel in formation. Currently, the more commonly used formation control algorithms include: virtual structure method, behavior-based method, artificial potential field method, and leader-following method. Since the leader-following method can effectively solve the problems related to the stability, formation switching, system obstacle avoidance and adaptive control of multi-AGV formations, this formation method is widely used [2].

[0004] The problem of multi-AGV tracking is the focus of multi-AGV collaborative formation. However, in actual scenarios, there are often various interferences. Communication delay is also one of the main factors leading to system instability [3,4]. Reference [5] proposed a time window-based priority path planning algorithm to solve the delay problem of multi-AGVs, and designed a time window-based dynamic priority local path planning algorithm to solve the communication delay problem. Reference [6] studied the unloading scheduling problem of periodic tasks and proposed a multi-AGV periodic unloading optimization algorithm to reduce unloading conflicts and processing delays caused by resource competition. In addition, multi-AGV is also a type of multi-agent. From this perspective, reference [7] designed a predictive coordination controller for a multi-agent system with communication delay, which can actively compensate for communication delays and achieve stability and consensus at the same time. In reference [8], in order to make the agents form an ideal formation and run collaboratively at a constant speed, a coordinated formation control algorithm was designed, and the variable gain technology was used to eliminate the impact of communication delay on formation control.

[0005] All of the above studies only considered the impact of communication delays, without fully considering other interferences. However, when multiple AGVs are deployed in a leader-follower formation, the impact of the leader vehicle’s acceleration changes can be easily overlooked. The rate of change of acceleration is called “jitter”, sometimes also called “acceleration of acceleration”. For example, when driving at a constant speed on a planned path, when the path changes from a straight line to a curve, the acceleration will change instantly at the starting point of the curve[9]. This change will affect the stability of the formation.

[0006] Reference documents:

[0007] a) H. Andreasson et al., “Autonomous transport vehicle: Where we are and what is missing,” IEEE Robotics & Automation Magazine, vol. 22, pp. 64 - 75, Mar. 2015.

[0008] b) Z. Wang, L. Wang, H. Zhang, L. Vlacic and Q. Chen, "Distributed Formation Control of Nonholonomic Wheeled Mobile Robots Subject to Longitudinal Slippage Constraints," in IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 51, pp. 2992 - 3003, May 2021.

[0009] c) M. Zheng, Z. Wang, L. Li, et al., “Finite - Time Generalized Projective Lag Synchronization Criteria for Neutral - Type Neural Networks with Delay,” Chaos Solitons & Fractals, vol. 107, pp. 195 - 203, Jan. 2018.

[0010] d) P. Ngoc, H. Trinh, “Novel Criteria for Exponential Stability of Linear Neutral Time - Varying Differential Systems,” IEEE Transactions on Automatic Control, vol. 61, pp. 1590 - 1594, Jun. 2016.

[0011] e) R. Tai, J. Wang, W. Chen. “A prioritized planning algorithm of trajectory coordination based on time windows for multiple AGV with delay disturbance,” Assembly Automation, vol. 5, pp. 753 - 768, Nov. 2019.

[0012] f) P. Liu, Z. Liu, J. Wang, ZF. Wu, P. Li, HJ. Lu, “Reinforcement learning empowered multi-AGV offloading scheduling in edge-cloud IIoT,” Journal of Cloud Computing - Advances Systems and Applications, vol. 1, Nov. 2022.

[0013] g) G.-P. Liu, “Coordination of Networked Nonlinear Multi-Agents Using a High-Order Fully Actuated Predictive Control Strategy,” in IEEE / CAA Journal of Automatica Sinica, vol. 9, pp. 615 - 623, April 2022.

[0015] h) L. Chen, C. LI, Y. Guo, G. Ma, Y. Li, B. Xiao, “Formation-containment control of multi-agent systems with communication delays,” ISA Transactions, vol. 128, pp. 32 - 43, 2022.

[0016] i) S. H. Schot, “Jerk: The time rate of change of acceleration,” American Journal of Physics, vol. 46, pp. 1090 - 1094, 1978. Summary of the Invention

[0017] According to the technical problems mentioned in the above background technology, a leader-follower AGV formation control method with a composite controller structure is provided.

[0018] The technical means adopted by the present invention are as follows:

[0019] A leader-follower AGV formation control method with a composite controller structure comprises the following steps:

[0020] Step 1: Describe the information flow between AGV formations through the communication topology matrix;

[0021] Step 2: Consider the AGV as a mass point and establish the longitudinal kinematic model of the vehicle;

[0022] Step 3: Set the control goal of maintaining stable operation under the influence of communication delay and leader AGV acceleration disturbance;

[0023] Step 4: In order to eliminate the disturbance of the leader AGV’s acceleration, a sliding mode controller is introduced;

[0024] Step 5: Considering the inter-vehicle communication delay, a feedback controller based on tracking error is introduced;

[0025] Step 6: Create the composite controller i (t), so that the entire formation remains stable when the system is disturbed by communication delays and the acceleration of the leader AGV.

[0026] Furthermore, in step 1, the information flow between the AGV formations is represented by the communication topology matrix H, and there exists: H = L + P;

[0027] Where L represents the Laplace matrix and P represents the adjacency matrix;

[0028] The Laplacian matrix L describes the directional connections between vehicles and is defined as:

[0029]

[0030] If the following vehicle i knows the information of the following vehicle j, then l ij = -1; otherwise l ij =0;

[0031] The adjacency matrix P represents the connection between the leader vehicle and the following vehicles and is defined as:

[0032]

[0033] If follower i is connected to the leader vehicle, then g i =1, otherwise, g i= 0; if vehicle j∈{0,…,n} is connected to vehicle i, it is called the neighbor of vehicle i; all neighbors form a set

[0034]

[0035] The communication topology matrix H is a non-singular matrix, and all its eigenvalues ​​are non-zero.

[0036] Furthermore, the vehicle longitudinal kinematic model is expressed as:

[0037]

[0038] Among them, z i (t), v i (t) and a i (t) represents the position, velocity and acceleration of the i-th AGV in the road coordinate system, i = 0, 1, ... n; i = 0 represents the leader; represents a constant greater than zero, representing the inertial delay of vehicle dynamics, u i (t) represents the control input;

[0039] make Represents the state vector of the vehicle; Derived, the vehicle longitudinal kinematic model is described as:

[0040]

[0041] Then, the tracking error of the i-th AGV is described as:

[0042]

[0043] Among them, δ i (t), and are the position error, velocity error and acceleration error of the i-th AGV, respectively; z0, v0 and a0 represent the position, velocity and acceleration of the leader vehicle in the road coordinate system, respectively; d i,0 Denotes the expected distance between the i-th AGV and the leader vehicle. Defines the distance d between the i-th and j-th AGVs i,j for;

[0044] d i,j =(ij)hv i (t)+(ij)d;

[0045] Where h is the headway distance, v i (t) represents the longitudinal speed of the i-th AGV in the road coordinate system, and d represents the minimum safety distance.

[0046] Furthermore, in the AGV formation, the state trajectory of the leader vehicle is set as the expected state trajectory of the followers; the state trajectory of the leader vehicle satisfies:

[0047]

[0048] Among them, v * (t) and a * (t) represent the desired velocity and acceleration respectively; τ(t) is the jitter, which represents the time rate of change of acceleration; τ(t) is bounded and satisfies |τ(t)|≤ι, where ι is a constant and ι>0.

[0049] Furthermore, the control target in step 3 is:

[0050] Each follower AGV maintains the same speed as the leader AGV and maintains the required spacing

[0051]

[0052] Among them, v i (t) represents the speed of the i-th AGV, v0(t) is the speed of the leader AGV, and z i (t) represents the longitudinal position of the i-th AGV, z i-1 (t) represents the longitudinal position of the i-1th AGV, d i,i-1 represents the expected distance between the i-th AGV and the i-1-th AGV.

[0053] Furthermore, the composite controller u i (t) Includes: Feedback controller based on following error and integral sliding mode controller

[0054]

[0055] in, represents tracking control, Its role is to alleviate the impact of jitter.

[0056] Furthermore, an integral sliding surface is introduced, wherein the integral sliding surface s i (t) is:

[0057]

[0058] in, is the acceleration error of the i-th AGV, Indicates the acceleration error The initial value of is the inertial delay of vehicle dynamics, a i(τ) represents the acceleration of the i-th AGV,

[0059] is the tracking controller;

[0060] In order to eliminate the influence of the leader AGV acceleration disturbance and ensure the AGV formation control performance, the sliding mode controller is:

[0061]

[0062] Among them, γ>ι and is a constant.

[0063] Furthermore, the feedback controller based on tracking error is:

[0064]

[0065] Where K = [k z k v k a ] represents the feedback gain vector; k z , k v and k a Both represent feedback gain; represents the tracking error vector; Δ i,j ∈[0,Δ) represents the inter-vehicle communication delay between the i-th AGV and the j-th AGV, and Δ>0 represents the upper bound of the time delay;

[0066] definition is the state vector of all following AGVs in the queue; the tracking error state equation of the AGV queue can be written as

[0067]

[0068] in represents the Kronecker product operation, and:

[0069]

[0070] The communication topology matrix is ​​decomposed into:

[0071] H=γΛγ -1 ;

[0072] in, is a non-singular matrix; the block diagonal matrix is ​​represented as:

[0073] Λ=diag(Λ1 … Λ m );

[0074] in,

[0075]

[0076] Λi The eigenvalue representing the topology is λ i The Jordan matrix of

[0077] The similarity transformation is defined as: but:

[0078]

[0079] The entire AGV formation can be decomposed into the following subsystems:

[0080]

[0081] Among them, Δ i represents the corresponding delay of the i-th subsystem; i=1,2,…,n,I3,I n and I Mi is a unit matrix of appropriate dimension. When the subsystem is stable, the entire formation system can be stable.

[0082] Compared with the prior art, the present invention has the following advantages:

[0083] The present invention provides a leader-follower formation mode, and takes into account communication delay and acceleration disturbance. Based on the third-order kinematic model, a distributed controller is designed. Combined with the feedback controller, an integral sliding surface is designed to reduce the influence of the leader vehicle's acceleration disturbance. Finally, its effectiveness is demonstrated through simulation experiments. Therefore, in future work, this method can be extended to more complex scenarios, such as curved driving of AGVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0085] Figure 1 It is a schematic diagram of the overall process of the present invention.

[0086] Figure 2 It is a freight AGV that adopts a leader-follower formation.

[0087] Figure 3 Schematic diagram of information transmission of LPF.

[0088] Figure 4 Schematic diagram of AGV formation; (a) spacing; (b) speed; (c) acceleration; (d) speed error.

[0089] Figure 5 Schematic diagram of the results of AGV formation under pure feedback control; (a) spacing; (b) speed; (c) acceleration; (d) speed error. DETAILED DESCRIPTION

[0090] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0091] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0092] The formation control of freight AGVs is widely used in busy automated freight ports. Figure 1 As shown in the figure, n+1 AGVs on the transfer channel automatically form a queue by sharing their information (such as position, speed and acceleration) through the wireless communication network. Here, the index of the leader vehicle is i=0 and the index of the follower vehicle is i=1,2,…,n. Each AGV can only communicate with its adjacent AGVs to obtain information to update its control input. But in reality, the communication between vehicles is not perfect. Therefore, it is necessary to consider the impact of communication delay on the formation control performance.

[0093] The present invention provides a leader-follower AGV formation control method with a composite controller structure, comprising the following steps:

[0094] Step 1: Describe the information flow between AGV formations through the communication topology matrix. In general, the information flow between AGV formations is described by the inter-vehicle communication topology. Here, algebraic graphs are used to summarize the unified model of various communication topologies. The information flow between vehicles can be represented by the communication topology matrix H, and there exists H = L + P, where L is the Laplace matrix and P is the adjacency matrix. The Laplace matrix describes the directional connection between vehicles and is defined as:

[0095]

[0096] If the following vehicle i knows the information of the following vehicle j, then l ij = -1; otherwise l ij = 0. The adjacency matrix represents the connection between the leading vehicle and the following vehicles:

[0097]

[0098] If follower i is connected to the leader vehicle, then g i =1, otherwise, g i = 0. If vehicle j∈{0,…,n} is connected to vehicle i, it is called a neighbor of vehicle i. All its neighbors form a set

[0099]

[0100] To ensure platoon stability, all followers are either directly connected to the leader or indirectly connected to the leader through other followers. This means that H is a non-singular matrix with all eigenvalues ​​non-zero.

[0101] Step 2: Consider the AGV as a mass point and establish the longitudinal kinematic model of the vehicle; the port freight AGV can be regarded as an ordinary truck. Considering the AGV as a mass point, the longitudinal kinematic characteristics of its path coordinates can be described as:

[0102]

[0103] Among them, z i (t), v i (t) and a i (t) represent the position, velocity and acceleration of the i-th AGV (i=0, 1,…n, where i=0 represents the leader vehicle) in the road coordinate system. is a constant greater than zero, representing the inertial delay of vehicle dynamics, u i (t) is the control input.

[0104] let

[0105] can be described as:

[0106]

[0107] like Figure 1 As shown, the tracking error of the i-th AGV can be described as

[0108]

[0109] Among them, δ i (t), and are the position error, velocity error and acceleration error of the i-th AGV, respectively; z0, c0 and a0 represent the position, velocity and acceleration of the leader vehicle in the road coordinate system, respectively; d i,0 represents the expected distance between the i-th AGV and the leader vehicle. Then the distance d between the i-th and j-th AGVs is defined as i,j for,

[0110] d i,j =(ij)hv i (t)+(ij)d;

[0111] Where h is the headway, v is i (t) represents the longitudinal velocity of the i-th AGV in the road coordinate system, and d is the ideal static gap.

[0112] In a leader-follower formation, the state of the leader car is the desired state of the follower. Therefore, the state trajectory of the leader car satisfies the following conditions:

[0113]

[0114] where v * (t) and a * (t) denote the desired velocity and acceleration respectively. τ(t) is the jitter, which represents the time rate of change of acceleration. It should be noted that this jitter will inevitably degrade the control performance of the queue. Here τ(t) is bounded and satisfies |τ(t)|≤l, where ι is a constant and ι>0.

[0115] Step 3: Set the control goal of maintaining stable operation under the influence of communication delay and leader AGV acceleration disturbance; the goal is to achieve the following goals in the presence of communication delay and leader AGV acceleration disturbance:

[0116] Each follower AGV maintains the same speed as the leader AGV and maintains the required spacing:

[0117]

[0118] Among them, v i (t) represents the speed of the i-th AGV, v0(t) is the speed of the leader AGV, and z i (t) represents the longitudinal position of the i-th AGV, z i-1 (t) represents the longitudinal position of the i-1th AGV, d i,i-1 represents the expected distance between the i-th AGV and the i-1-th AGV.

[0119] Step 4: In order to alleviate the disturbance of the leader AGV acceleration, a sliding mode controller is introduced; in order to alleviate the disturbance of the leader AGV acceleration, the following integral sliding mode surface is proposed here:

[0120]

[0121] in, is the acceleration error of the i-th AGV, Indicates the acceleration error The initial value of is the inertial delay of vehicle dynamics, a i (τ) represents the acceleration of the i-th AGV, It is a tracking controller.

[0122] For vehicle kinematics and integral sliding surface, the following controller can ensure that the AGV formation can achieve the ideal control state when starting, and is not affected by the acceleration disturbance of the leader AGV;

[0123]

[0124] where γ>ι is a constant.

[0125] To prove the sliding mode controller of step 4:

[0126] First, find the first-order derivative of the sliding surface function and substitute it into the equation:

[0127]

[0128] Then, choose the following Lyapunov function:

[0129]

[0130] Taking the first-order derivative of V2(t) we get:

[0131]

[0132] Then there is the following algebraic relationship:

[0133] -γ|s i (t)|-s i (t)τ(t)≤-γ|s i (t)|+l|s i (t)|;

[0134] The right side can be written as:

[0135]

[0136] Where γ-ι>0. Obviously, under the action of the controller, the integral sliding surface of each following vehicle is finite-time stable.

[0137] Step 5: Considering the communication delay between vehicles, a feedback controller based on tracking error is introduced; when the tracking error remains on the sliding surface, Substituting the controller (10) into the vehicle dynamics equation (5), we can obtain

[0138]

[0139] Taking into account the inter-vehicle communication delay, the present invention designs the following feedback controller based on tracking error for each following AGV:

[0140]

[0141] Where K = [k z k v k a ] is the feedback gain vector, k z , k v and k a is the feedback gain to be designed. is the tracking error vector. Δ i,j ∈[0,Δ) is the inter-vehicle communication delay between the i-th AGV and the j-th AGV, and Δ>0 is the upper bound of the time delay.

[0142] definition is the collective state of all the following AGVs in the queue. The tracking error state equation of the AGV queue can be written as

[0143]

[0144] in represents the Kronecker product operation, and

[0145]

[0146] In order to simplify the problem, the present invention further decomposes the tracking error state equation of the entire AGV queue into some independent simplified subsystems through similarity transformation. The transformation matrix involved is obtained by eigenvalue decomposition of the communication matrix. The communication matrix can be written as follows

[0147] H=γΛγ -1 ;

[0148] in is a non-singular matrix. The block diagonal matrix is ​​represented by Λ=diag(Λ1 … Λ m ).

[0149] in

[0150]

[0151] The eigenvalue of the topology is λ i The Jordan matrix.

[0152] Here, the present invention defines a similarity transformation The following form can be obtained:

[0153]

[0154] Since (22) is diagonally partitioned, the stability of the entire queue is equivalent to the following decomposed subsystems

[0155]

[0156] where Δ i , i=1,2,…,n represents the corresponding delay of the i-th subsystem.

[0157] For a system, the system is asymptotically stable if and only if the communication delay satisfies the following condition

[0158]

[0159] where Δ i The following relationship is satisfied:

[0160]

[0161] where P is a positive definite matrix, and

[0162] Q1=-[P(A i +λ i BK)+(A i +λ i BK) T P];

[0163]

[0164] Q3=P(λ i BK) 2 P -1 ((λ i BK) 2 ) T P;

[0165] Since the matrices A and All of them are block diagonal matrices. The present invention has:

[0166]

[0167] therefore:

[0168]

[0169] Without considering communication delay, if the controller gain satisfies the following formula, the system is Hurwitz stable.

[0170]

[0171] When Δ i =0, it can be written as:

[0172]

[0173] A-λ i The characteristic polynomial of BK is:

[0174]

[0175] According to the Ross-Herwitz stability criterion, when the controller gain satisfies, the matrix A-λ i BK is Hurwitz stable. When the inter-vehicle communication delay satisfies Δ i,j ∈[0,Δ), the dynamic stability of the tracking error can still be guaranteed.

[0176] For a multi-AGV formation that considers longitudinal dynamics, the controller is able to ensure that the entire formation operates stably when the system is disturbed by communication delays and acceleration of the leader AGV.

[0177] Since the sliding surface starts running on it at the initial time, this means that the system state of each following vehicle will not leave the sliding surface in continuous time, that is, when also lead to

[0178]

[0179] When t>0, the present invention has

[0180] τ(t)=-γsgn(s i (t));

[0181] That is, by the equivalent value [-γsgn(0)] eq The jitter is compensated. The stability of the feedback controller has been discussed. Therefore, the controller ensures that the entire system remains stable when subjected to communication delays and acceleration disturbances.

[0182] It is worth noting that the proposed integral sliding mode controller consists of a feedback controller and adaptive controller This means that the introduced adaptive controller can effectively expand the pure feedback controller to adaptively alleviate the impact of acceleration jitter.

[0183] Step 6: Create the composite controller i (t), ensuring that the entire formation can maintain stable operation when affected by communication delays and acceleration disturbances of the leader AGV. i (t) Includes: Feedback controller based on following error and integral sliding mode controller

[0184]

[0185] in, represents tracking control, Its role is to alleviate the impact of jitter.

[0186] Embodiment 1:

[0187] The effectiveness of the proposed control method will be verified by numerical experiments. In this paper, a simulation model of an AGV formation consisting of 5 AGVs (1 leader AGV and 4 follower AGVs) is established in Simulink. All AGVs travel in a straight line. In order to be more realistic, a nonlinear dynamic model is used in the simulation. According to the literature

[15] , the longitudinal nonlinear dynamic model of a four-wheel fuel vehicle can be expressed as

[0188]

[0189] Where m i Indicates the quality of AGV, C A,i represents the comprehensive air resistance coefficient, g represents the acceleration of gravity, f represents the comprehensive rolling resistance, T i Indicates the actual driving / braking torque, T i,des represents the desired driving / braking torque, R i Indicates the tire radius, represents the mechanical efficiency of the transmission system. These parameter values ​​are listed in Table I.

[0190] Table I. Vehicle parameters used in the simulation

[0191]

[0192]

[0193] The vehicle's driving force / braking force can be obtained from the control input by the following conversion

[0194]

[0195] Incorporating the required driving / braking torque into the nonlinear vehicle dynamics results in a linear vehicle kinematic model.

[0196] The controller gain is K = [0.17 1.4 1.7]. The initial positions of the AGVs in the formation are as follows: i (0) = [45, 30, 15, 0], which means that the initial spacing error is not zero, and the initial position of the leader AGV is selected as z0(0) = 60. The initial speed of all AGVs is set to 2m / s, and the initial acceleration is set to 0m / s 2 In the proposed integral sliding mode controller, the present invention sets the parameter γ = 0.01. The present invention uses the lead-follower (LPF) topology to verify the effectiveness of the scheme. Each follower receives information from the leader and the front car. The information transmission of this topology is as follows Figure 3 As shown. For LPF, its communication topology matrix is:

[0197]

[0198] Assuming that the communication delay of all AGVs is the same, according to Lemma 1, the communication delay is set to Δ i,j =0.05. Assume that the AGVs are homogeneous, that is, the parameters of each vehicle are consistent.

[0199] The experiment was divided into two groups. The first group was the method proposed by the present invention. The experimental results are shown in Figure 4 The second group is pure feedback control, and the experimental results are shown in Figure 5 shown.

[0200] Figure 4 The experimental results of the method proposed in the present invention show that at the moment when the acceleration of the leader AGV changes frequently, the method proposed in the present invention can keep the distance in a stable state and the speed error can also be kept within an acceptable range. Figure 5 The experimental results under pure feedback control are shown. Under the same conditions, although the spacing can eventually be maintained within the stable range, due to the lack of compensation for acceleration disturbance, the acceleration of the following vehicle does not present a smooth curve, and there is some jitter, so the speed error is also significantly higher than the method proposed in the present invention. Moreover, after being disturbed, the convergence speed of all curves is slower than that of the method proposed in the present invention.

[0201] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A leader-follower AGV formation control method with a composite controller structure, characterized in that: The following steps are involved: Step 1: Describe the information flow between AGV formations through the communication topology matrix; Step 2: Consider the AGV as a mass point and establish the longitudinal kinematic model of the vehicle; Step 3: Set the control goal of maintaining stable operation under the influence of communication delay and leader AGV acceleration disturbance; Step 4: In order to eliminate the disturbance of the leader AGV’s acceleration, a sliding mode controller is introduced; Step 5: Considering the inter-vehicle communication delay, a feedback controller based on tracking error is introduced; Step 6: Create the composite controller i (t), so that the entire formation remains stable when the system is disturbed by communication delays and the acceleration of the leader AGV.

2. A leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: In step 1, the information flow between AGV formations is represented by the communication topology matrix H, and there exists: H = L + P; Where L represents the Laplace matrix and P represents the adjacency matrix; The Laplacian matrix L describes the directional connections between vehicles and is defined as: If the following vehicle i knows the information of the following vehicle j, then l ij = -1; otherwise l ij =0; The adjacency matrix P represents the connection between the leader vehicle and the following vehicles and is defined as: If follower i is connected to the leader vehicle, then g i =1, otherwise, g i = 0; if vehicle j∈{0,…,n} is connected to vehicle i, it is called the neighbor of vehicle i; all neighbors form a set The communication topology matrix H is a non-singular matrix, and all its eigenvalues ​​are non-zero.

3. A leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: The vehicle longitudinal kinematic model is expressed as: Among them, z i (t), v i (t) and a i (t) represents the position, velocity and acceleration of the i-th AGV in the road coordinate system, i = 0, 1, ... n; i = 0 represents the leader; represents a constant greater than zero, representing the inertial delay of vehicle dynamics, u i (t) represents the control input; make Represents the state vector of the vehicle; Derived, the vehicle longitudinal kinematic model is described as: Then, the tracking error of the i-th AGV is described as: Among them, δ i (t), and are the position error, velocity error and acceleration error of the i-th AGV, respectively; z0, v0 and a0 represent the position, velocity and acceleration of the leader vehicle in the road coordinate system, respectively; d i,0 Denotes the expected distance between the i-th AGV and the leader vehicle. Defines the distance d between the i-th and j-th AGVs i,j for: d i,j =(i-j)hv i (t)+(i-j)d; Where h is the headway, v is i (t) represents the longitudinal speed of the i-th AGV in the road coordinate system, and d represents the minimum safety distance.

4. A leader-follower AGV formation control method with a composite controller structure according to claim 3, characterized in that: In the AGV formation, the state trajectory of the leader vehicle is set as the expected state trajectory of the followers; the state trajectory of the leader vehicle satisfies: Among them, v * (t) and a * (t) represent the desired velocity and acceleration respectively; τ(t) is the jitter, which represents the time rate of change of acceleration; τ(t) is bounded and satisfies |τ(t)|≤ι, where ι is a constant and l>0.

5. The leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: The control objectives in step 3 are: Each follower AGV maintains the same speed as the leader AGV and maintains the required spacing Among them, v i (t) represents the speed of the i-th AGV, v0(t) is the speed of the leader AGV, and z i (t) represents the longitudinal position of the i-th AGV, z i-1 (t) represents the longitudinal position of the i-1th AGV, d i,i-1 represents the expected distance between the i-th AGV and the i-1-th AGV.

6. A leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: The composite controller u i (t) Includes: Feedback controller based on following error and integral sliding mode controller in, represents tracking control, Its function is to compensate for the impact of acceleration disturbance.

7. The leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: The integral sliding surface s i (t) is: in, is the acceleration error of the i-th AGV, Indicates the acceleration error The initial value of is the inertial delay of vehicle dynamics, a i (τ) represents the acceleration of the i-th AGV, is the tracking controller; In order to eliminate the influence of the leader AGV acceleration disturbance and ensure the AGV formation control performance, the sliding mode controller is: Among them, γ>ι and is a constant.

8. The leader-follower AGV formation control method with a composite controller structure according to claim 1, characterized in that: The feedback controller based on tracking error is: Where K = [k z k v k a ] represents the feedback gain vector; k z , k v and k a Both represent feedback gain; represents the tracking error vector; Δ i,j ∈[0,Δ) represents the inter-vehicle communication delay between the i-th AGV and the j-th AGV, and Δ>0 represents the upper bound of the time delay; definition is the state vector of all following AGVs in the queue; the tracking error state equation of the AGV queue can be written as: in represents the Kronecker product operation, and: The communication matrix is ​​decomposed into: H=γΛγ -1 ; in, is a non-singular matrix; the block diagonal matrix is ​​represented as: Λ=diag(Λ1 … Λ m ); in, Λ i The eigenvalue representing the topology is λ i The Jordan matrix of The similarity transformation is defined as: but: The entire AGV formation can be decomposed into the following subsystems: Among them, Δ i Represents the corresponding delay of the i-th subsystem; i=1,2,…,n.