Vehicle cooperative queue control method based on dynamic event triggering

Abstract

The invention discloses a vehicle queue cooperative control method based on a dynamic event triggering strategy. The method comprises the following steps: step 1, establishing kinematic models of a leading vehicle and a following vehicle; 2, designing a following vehicle spacing strategy related to the speed of the leading vehicle; defining a tracking error and a state estimation error of the following vehicle to obtain a closed-loop error system; a distributed dynamic event triggering mechanism is designed according to the closed-loop error system; 3, when the triggering condition of the distributed dynamic event triggering mechanism is met, the following vehicles sample the states of the following vehicles and request neighbors to send the latest state to update the distributed event triggering controller, so that control input of all the following vehicles is dynamically updated, and it is guaranteed that tracking errors of the following vehicles are not amplified and propagated along the queue when the motorcade stably runs. Based on the dynamic event triggering strategy, the safety in the vehicle queue driving process is comprehensively improved, communication resources are saved, and the timeliness and accuracy of information interaction between vehicles are guaranteed.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of intelligent transportation, and particularly relates to a cooperative control technology of a vehicle platoon, and specifically relates to an application of a distributed control method based on a dynamic event triggering strategy in cooperative control of a vehicle platoon. BACKGROUND

[0002] With the vigorous development of intelligent transportation systems, cooperative control of vehicle platoons has become a research focus in the field and plays a key role in many practical application scenarios. Under the communication environment of a vehicle ad hoc network (VANET), data interaction and control between vehicles face a series of severe challenges. Under the VANET communication mode, the data measured by vehicles must be sampled and digitized before being shared, and almost all control protocols and algorithms are implemented on digital modules or computers that rely on discrete-time instantaneous calculation of control instructions. In order to ensure control effect, traditional control strategies often need to update control inputs frequently, which leads to a sharp increase in inter-vehicle data transmission, further exacerbating communication congestion. At the same time, the maneuverability of traditional inter-vehicle spacing strategies is poor. Common fixed spacing (CS) and fixed car head time (CTH) strategies are difficult to balance different traffic conditions and vehicle speeds for flexible adjustment. At low speed in congestion, a fixed large spacing will reduce the road capacity, making traffic congestion worse; at high speed, a fixed small spacing will increase the risk of vehicle collision. These traditional strategies cannot meet the needs of complex and variable traffic environments, greatly limiting the improvement of vehicle platoon driving performance.

[0003] Therefore, how to effectively reduce communication resource consumption and improve spacing strategy maneuverability while ensuring the accuracy and real-time performance of vehicle platoon control has become an important issue to be solved in current cooperative control of vehicle platoons. SUMMARY

[0004] The purpose of the present application is to overcome the deficiencies in the prior art and provide a vehicle platoon cooperative control method based on a dynamic event triggering strategy to solve the problems of traditional controllers not being suitable for discrete sampling environments, large communication resource consumption, and poor maneuverability of inter-vehicle spacing strategies.

[0005] Technical solution: A vehicle platoon cooperative control method based on a dynamic event triggering strategy, comprising the following steps:

[0006] Step 1: Abstract the leader vehicle and the follower vehicle in the vehicle platoon as nodes, describe the communication relationship between the vehicles according to algebraic graph theory to obtain the communication topology of the vehicle platoon, and collect the kinematic characteristics of the vehicles to establish kinematic models of the leader vehicle and the follower vehicle, respectively;

[0007] Step 2: design a following vehicle spacing strategy related to the speed of the leading vehicle based on the kinematic model; define the tracking error and state estimation error of the following vehicle to obtain a closed-loop error system; and design a distributed dynamic event triggering mechanism according to the closed-loop error system;

[0008] Step 3: design a distributed event-triggered controller; when the triggering condition of the distributed dynamic event triggering mechanism is met, the following vehicle samples its own state and requests the neighbors to send the latest state to update the distributed event-triggered controller, so that the control input of all following vehicles is dynamically updated, and the tracking error of the following vehicles is not amplified along the queue when the platoon is running stably.

[0009] Further, the process of describing the communication relationship between vehicles according to algebraic graph theory to obtain the communication topology of the vehicle platoon comprises: considering that the length of the platoon is N+1, which includes 1 leading vehicle marked as i=0 and N following vehicles marked as i=1, 2,..., N; modeling the communication topology by using an undirected graph , wherein is a point set composed of nodes abstracted by the N following vehicles, v i represents the node abstracted by the i-th vehicle; is an edge set representing whether the vehicles i and j have communication capability; is an adjacency matrix used to represent the connection state between all following vehicles in the platoon; when there is a bidirectional communication connection between the vehicle i and the vehicle j, i.e. , then a ij =1; otherwise a i0 =0; meanwhile, a state connection diagonal matrix, i.e. a traction matrix , is defined between the leading vehicle and the following vehicles, wherein a i0 =1 if there is a communication connection between the following vehicle and the leading vehicle; otherwise a i0 =0; assuming that is an undirected graph, then the Laplacian matrix is defined as , wherein represents the degree of the node; the augmented undirected graph includes the original graph and an additional node 0, the information of the node 0 is only available for a subset of nodes in the original graph , and the node 0 does not receive information from any node in .

[0010] Further, the kinematic models of the leading vehicle and the following vehicle are respectively established as follows:

[0011] Kinematic model of the leading vehicle:

[0012]

[0013] p0(t) and v0(t) represent the displacement and velocity of the leading vehicle, respectively;

[0014] Kinematic model of the following vehicle:

[0015]

[0016] where p i (t), v i (t) and a i (t) represent the displacement, velocity and acceleration of the following vehicle i at time t, respectively, represents the control input of the i-th following vehicle, and τ is the inertia time constant in the vehicle powertrain system;

[0017] The state of the following vehicle is rewritten in vector form x i (t) = [p i (t) v i (t) a i (t)] T The kinematic equation of the following vehicle can be rewritten as:

[0018]

[0019] Further, the following vehicle spacing strategy related to the speed of the leading vehicle includes:

[0020] The expression of the dynamic spacing of the following vehicle is:

[0021]

[0022] where the expected spacing expression between the i-th vehicle and the j-th vehicle is d ij (t) = d i0 (t) - d j0 (t), where d i0 is the headway between the i-th following vehicle and the leading vehicle 0, d j0 (t) is the headway between the j-th following vehicle and the leading vehicle 0, d ij is the headway between the i-th following vehicle and the j-th following vehicle, b i0 and c i0 are spacing strategy parameters; s is a normal number representing the minimum expected distance between two adjacent vehicles at rest, and l is the vehicle body length; this following vehicle spacing strategy allows the following vehicle to maintain a smaller spacing at high speed to increase road capacity while maintaining a larger spacing at low speed to avoid collision risk.

[0023] Further, the definition follows the tracking error and state estimation error of the following vehicle, specifically:

[0024] The tracking error of the following vehicle i and the following vehicle j is defined as:

[0025] z i (t) = p i (t) - p j (t) + d ij (t) (13)

[0026] The tracking error variable of the displacement, speed and acceleration of the following vehicle i is defined as:

[0027]

[0028] The state estimation error of the displacement, speed and acceleration of the following vehicle i is defined as:

[0029]

[0030] Wherein, and is the displacement, speed and acceleration estimation of the i-th vehicle at time t;

[0031] The tracking error vector of the following vehicle i and the state estimation error vector are defined as:

[0032]

[0033] Further, the closed-loop error system is specifically:

[0034] Combined with the vehicle kinematics model, the closed-loop error system of the i-th following vehicle (using ε0(t) = 0 and e0(t) = 0) is as follows:

[0035]

[0036] In order to uniformly describe the state of the entire vehicle queue, the closed-loop error system of the entire vehicle team is constructed:

[0037]

[0038] Let ε(t) = [ε1(t) ε2(t) … ε N (t)] T And e(t) = [e1(t) e2(t) … e N (t)] T , then the compact form of the closed-loop error system can be obtained:

[0039]

[0040] The equation describes the dynamic evolution process of the whole vehicle platoon state error, wherein I N is an N-dimensional unit matrix, denotes the Kronecker product, and A and B are system matrices.

[0041] Further, the distributed dynamic event triggering mechanism comprises a triggering condition and an update law of a dynamic threshold of a following vehicle.

[0042] The triggering condition is:

[0043]

[0044] The update law of the dynamic threshold of the following vehicle is:

[0045]

[0046] The discrete triggering time is described as a monotonically increasing time sequence denotes the triggering time of the vehicle i for the k+1th time, and inf is the lower limit; the dynamic threshold of each following vehicle is η i (t), η i (0)>0, θ, β i and σ i are threshold parameters, and satisfy θ>0, β i >0, and 0<σ i <1; each following vehicle only communicates and updates its control input when the triggering condition is met; the Riccati equation PA+A T P-γPBB T P is the unique positive definite solution of the Riccati equation P = -I, the feedback gain matrix Γ = PBB T P is designed, and the matrix

[0047] μ is a design parameter.

[0048] Further, the designed distributed event-triggered controller is as follows:

[0049] denotes the control input of the following vehicle i, and the distributed event-triggered controller of the following vehicle i is constructed as follows:

[0050]

[0051] wherein K p , K v , and K a are feedback control gains; the matrix is defined, and (14) is rewritten as

[0052]

[0053] wherein, D ij = [d ij (t) 00] T , D i0 = [d i0 (t) 00] T .

[0054] Further, when the trigger condition of the distributed dynamic event trigger mechanism is met in step 3, the estimated value is replaced by the trigger value at the moment, and the control input of the i-th follower vehicle is:

[0055]

[0056] If there exists a control input u for any initial state x0(0) and x i (0) such that

[0057]

[0058] If the above is true for all follower vehicles, the entire vehicle platoon achieves the leader-follower consensus goal, which ensures that the tracking error of the follower vehicle does not amplify along the platoon when the platoon runs stably. The leader-follower consensus goal means that the follower vehicle needs to track the motion state (displacement, speed, etc.) of the leader vehicle over time, so that the difference between the motion state of all follower vehicles and the leader vehicle tends to zero. The leader-follower consensus goal enables all follower vehicles to follow the driving trajectory and speed change of the leader vehicle while maintaining the preset vehicle distance (or relative displacement) within the platoon, avoiding collisions or platoon dispersion, and ultimately achieving the "orderly and synchronized" motion of the entire platoon.

[0059] An electronic device comprising a processor and a memory; the memory is used to store a computer program, which is run by the processor to execute a vehicle platoon cooperative control method based on a dynamic event trigger strategy.

[0060] Beneficial effects: the speed-dependent follower vehicle spacing strategy adopted by the present application can dynamically adjust the vehicle spacing according to the real-time speed of the vehicle, reduce the possibility of rear-end accidents caused by improper spacing, and comprehensively improve the safety of the vehicle platoon during driving. The distributed dynamic event trigger mechanism is introduced, and only when the state error of each vehicle meets the trigger condition, communication is performed, saving communication resources. The distributed event trigger controller only needs to sample the neighbor state information when the trigger condition is met, and each follower vehicle updates its control input by receiving neighbor state information; effectively solving the problem of excessive resource consumption caused by high-frequency periodic broadcast sampling in traditional control methods, thereby ensuring the timeliness and accuracy of information interaction between vehicles. BRIEF DESCRIPTION OF DRAWINGS

[0061] ​​Figure 1 a vehicle communication topology of the present application;

[0062] Fig. 2(a) shows displacement tracking performance of a vehicle platoon, Fig. 2(b) shows speed tracking performance of a vehicle platoon, and Fig. 2(c) shows acceleration tracking performance of a vehicle platoon;

[0063] Fig. 3(a) shows dynamic variation curve of vehicle platoon spacing error, and Fig. 3(b) shows dynamic variation curve of vehicle platoon speed tracking error;

[0064] Figure 4 a control input diagram of a first following vehicle of the present application;

[0065] Figure 5 an event trigger time diagram of a first following vehicle of the present application. DETAILED DESCRIPTION

[0066] The technical solutions of the present application are described below clearly and completely according to the accompanying drawings and simulation examples.

[0067] A vehicle platoon cooperative control method based on a dynamic event trigger strategy, comprising the following steps:

[0068] Step 1: abstract the leading vehicle and the following vehicles in the vehicle platoon as nodes, describe the communication relationship between the vehicles according to algebraic graph theory to obtain the communication topology of the vehicle platoon, and collect the kinematic characteristics of the vehicles to respectively establish the kinematic models of the leading vehicle and the following vehicles;

[0069] Step 2: combine the kinematic models to design a following vehicle spacing strategy related to the speed of the leading vehicle; define the tracking error and state estimation error of the following vehicle to obtain a closed-loop error system; and design a distributed dynamic event trigger mechanism according to the closed-loop error system;

[0070] Step 3: design a distributed event trigger controller; when the trigger condition of the distributed dynamic event trigger mechanism is met, the following vehicle samples its own state and requests the neighbors to send the latest state to update the distributed event trigger controller, so that the control input of all the following vehicles is dynamically updated, and the tracking error of the following vehicles is prevented from being amplified and propagated along the platoon when the platoon is stably running.

[0071] The kinematic model establishment process of Step 1 is as follows:

[0072] 1.1 Consider that the platoon length is N+1, which contains 1 leading vehicle (marked as i=0) and N following vehicles (marked as i=1, 2,..., N). A undirected graph is used to model the communication topology, where is the point set, is the edge set, is the adjacency matrix, which is used to represent the connection status between all following vehicles in the platoon. There is a bidirectional communication connection between vehicle i and vehicle j, i.e. a ij i,j = 1 otherwise a i0 i,j = 0. Meanwhile, define the status connection diagonal matrix between the leading vehicle and the following vehicles, i.e. the traction matrix a i0 i = 1; otherwise a i i = 0. Assume is an undirected graph, then the Laplacian matrix is defined as where denotes the degree of the node. The augmented undirected graph contains the original graph and an extra node (labeled 0), the information of node 0 is only available to a subset of nodes in the original graph , and node 0 does not receive information from any node in , define the matrix

[0073] 1.2 Consider a platoon consisting of N+1 vehicles driving on a straight road, where N following vehicles are equipped with VANET for wireless communication, transmitting their own state information, and updating the control input autonomously through a distributed event-triggered controller, so as to track the motion trajectory of the leading vehicle. Wherein, the kinematic model of the leading vehicle can be described as:

[0074]

[0075] p0(t) and v0(t) represent the displacement and speed of the leading vehicle, respectively;

[0076] Kinematic model of the following vehicle:

[0077]

[0078] where p i (t), v i (t) and a i (t) represent the displacement, speed and acceleration of the following vehicle i at time t, respectively, represent the control input of the i-th following vehicle. τ is the inertia time constant in the vehicle power transmission system.

[0079] 1.3 For convenience of representation, the state of the vehicle is rewritten as a vector form x i (t) = [p i (t) v i (t) a i (t)] TThen the vehicle kinematic equation can be rewritten as:

[0080]

[0081] Step 2 adopts a following vehicle spacing strategy related to the leader vehicle speed (VTH spacing strategy), which is very similar to the behavior of human drivers;

[0082] 2.1 The dynamic spacing expression of the following vehicle is:

[0083]

[0084] where the expected spacing expression between the ith vehicle and the jth vehicle is d ij (t) = d i0 (t) - d j0 (t), where d i0 is the headway between vehicle i and leader 0, d ij is the headway between vehicle i and vehicle j, b i0 and c i0 are the spacing strategy parameters to be designed; s is a positive constant representing the minimum expected distance between two adjacent vehicles at rest, and l is the vehicle body length. This following vehicle spacing strategy allows the following vehicle to maintain a smaller spacing at high speed and a larger spacing at low speed, thereby improving road capacity and avoiding collision risks.

[0085] 2.2 The leader-following consensus goal needs to be achieved for the vehicle platoon, i.e.

[0086]

[0087] 2.3 To ensure that the vehicle platoon maintains chordal stability during driving, i.e., to avoid the gradual amplification of tracking errors as the queue is transmitted backward;

[0088] Define the tracking error between the following vehicle i and the following vehicle j as:

[0089] z i (t) = p i (t) - p j (t) + d ij (t)

[0090] Define the tracking error variables of the displacement, speed, and acceleration of the following vehicle i as:

[0091]

[0092] Define the state estimation error of the displacement, speed, and acceleration of the following vehicle i as:

[0093]

[0094] where, and are the displacement, velocity and acceleration estimates of the ith vehicle at time t within the interval between the two consecutive triggering instants;

[0095] Combining the kinematic model of the vehicle, the closed-loop error system of the ith follower vehicle is given as follows:

[0096]

[0097] Continuing the definition of the tracking and state estimation error vectors of the ith follower vehicle:

[0098]

[0099] The closed-loop error system of the ith follower vehicle is given as follows:

[0100]

[0101] 2.4 To uniformly describe the state of the entire vehicle platoon, we extend the perspective from a single vehicle i to the entire platoon. We can construct the closed-loop error system of the entire platoon. Further, the closed-loop error system of the entire platoon is given as follows (using ε0(t) = 0 and e0(t) = 0):

[0102]

[0103] Let ε(t) = [ε1(t) ε2(t) … ε N (t)] T and e(t) = [e1(t) e2(t) … e N (t)] T then the compact form of the closed-loop error system is obtained as follows:

[0104]

[0105] This equation describes the dynamic evolution of the state error of the entire platoon, where I N is the N-dimensional identity matrix, denotes the Kronecker product, and A, B are system matrices.

[0106] 2.5 The discrete triggering instants can be described as a monotonically increasing time sequence denotes the time instant of the k+1th triggering of vehicle i; to ensure that the tracking error does not gradually amplify as it is passed back through the platoon during vehicle travel, a distributed dynamic event-triggering mechanism is designed for each follower vehicle as follows:

[0107]

[0108] where η i (t) is a dynamic threshold variable and η i (0) > 0, other threshold parameters satisfy θ > 0, β i > 0, 0 < σ i < 1.

[0109] Step 3 designs a distributed event-triggered controller, the process is as follows:

[0110] 3.1 Construct a distributed event-triggered controller, denotes the control input of vehicle i, and the designed control protocol is as follows:

[0111]

[0112] where, and are the displacement, velocity and acceleration estimates of the ith vehicle at the time interval between the adjacent two triggering instants; K p , K v , K a are feedback control gains to be designed. Define the matrix D ij = [d ij (t) 0 0] T , D i0 = [d i0 (t) 0 0] T , then the above formula can be rewritten as:

[0113]

[0114] 3.2 The dynamic event-triggering mechanism ensures that for t ∈ [0, ∞), there is:

[0115]

[0116] Moving the formula, we get:

[0117]

[0118] When θ = 0, we get η i (t) > 0; when θ ≠ 0, we get

[0119] Combined with the distributed dynamic event-triggering mechanism, we get:

[0120]

[0121] And then for all t ∈ [0, ∞), there is η i (t) ≥ η i(0) x exp(-β-1 / θ), and η i (0)>0, then η i (t)>0.Thus, the distributed dynamic event-triggered mechanism requires that the vehicle only calculates the dynamic threshold η i (t), communicates and updates its control input when the triggering condition is satisfied, which solves the problem that the traditional controller can only update the occupation by periodic sampling and occupies the communication resources.

[0122] 3.3 At the moment when the triggering value replaces the estimated value, the control input of the i-th vehicle is:

[0123]

[0124] If for any initial state x0(0) and x i (0), there exists a control input that can make:

[0125]

[0126] For all vehicles, the entire queue reaches the leader-following consensus, wherein i=1, 2,..., N.

[0127] The assumptions and lemmas used in the application are as follows:

[0128] Assumption 1: The system is controllable.

[0129] Assumption 2: is a simple graph, that is, a ii =0.

[0130] Lemma 1: For the undirected graph Laplacian matrix has the following properties: is a semi-positive definite matrix;

[0131] Lemma 2: If is connected, and at least one node in can access node 0, then

[0132] Lemma 3: If assumption 1 is true, the following algebraic Riccati equation (ARE) PA+A T P-γPBB T P=-I has a unique positive definite solution P, wherein A and B are system matrices, I n is an n-order unit matrix, design the feedback gain matrix Γ=PBB T P, and μ is a parameter to be designed.

[0133] Lemma 4 (Barbalat’s Lemma): If φ(t) is uniformly continuous and monotonically increasing for t≥0, then

[0134]

[0135]

[0136] The stability of the system is verified, including the stability of the closed-loop error system, the avoidance of Zeno behavior in the distributed dynamic event-triggered mechanism, and the chordal stability. The process is as follows:

[0137] A Lyapunov function V(t) is selected for the tracking error

[0138]

[0139] According to Lemma 3, V(t) is a positive definite matrix. Taking the derivative of V(t) with respect to time t and substituting and Γ=PBB T P, we get:

[0140]

[0141] Since and Γ≥0, we can use Young’s inequality to get:

[0142]

[0143] According to Lemma 3, it is not difficult to calculate:

[0144]

[0145] Note that:

[0146]

[0147] We get:

[0148]

[0149] where α=min i＝1,2,…,N {(1-σ i ) / λ max (P),β i}. Since 0<σ i <1, we have α>0. Using the above formula, we further get V(t)≤e -αt V(0).

[0150] For any vehicle i∈N, assume its current triggering time is​ We need to prove that the length of its next event interval is strictly positive. According to e i The definition of (t) is in the trigger interval. In the middle, for ||e i Taking the derivative of (t)||, we obtain its right-hand derivative as follows:

[0151]

[0152] In the formula Since it has been proven that the error converges, there exists a positive constant ψ. i Meet the conditions

[0153] When t∈[0,∞), D + ||e i (t)||≤||Ae i (t)||+ψ i

[0154] It is not difficult to deduce e i (t) in Upper bound of the interval:

[0155]

[0156] At the (k+1)th trigger, we have:

[0157]

[0158] again but exist:

[0159]

[0160] Summarized as follows:

[0161]

[0162] but:

[0163]

[0164] Substitution have:

[0165]

[0166] Summarized as follows:

[0167]

[0168] but Therefore, Zeno behavior does not exist for any vehicle.

[0169] Since The vehicle acceleration error is bounded in the controller design, (Uniformly bounded), it can be concluded that the first derivative is uniformly continuous. i (0) is zero in the ideal state,

[0170]

[0171] That is Square integrable. Because Uniformly continuous and integral limited, according to Barbalat's lemma, it can be obtained Similarly, it can be obtained

[0172] If the initial error satisfies ( is an arbitrarily small amount). Since z i (t) is continuous and converges to zero, there must be a constant υ>0 such that sup|z i (t)|<υ for t∈[0,∞).

[0173] In order to verify the effectiveness of the proposed distributed dynamic event-triggered control algorithm, the present application carries out numerical simulation experiment based on MATLAB platform. The simulation setting includes 1 leader vehicle and 5 follower vehicles (i.e. N=5), the leader vehicle travels at a constant speed v0=20 m / s, the vehicle body length l=3 m, and the minimum expected distance s=40 m when the adjacent two vehicles are stationary. The initial displacement of the follower vehicles is randomly distributed within the interval [-200, 0] m, and the initial speed is randomly set within the range [0, 20] m / s. The time constant τ=0.75 s. The controller parameter settings are: σ=0.9, θ=0.1, and β=0.01. The simulation duration T=60 s, the time step Δt=0.05 s, and the communication topology adopts LBD structure.

[0174] The simulation results are shown in the figures. Figure 2(a) shows the displacement tracking performance of the vehicle queue. The simulation results show that there is a displacement deviation (the maximum deviation is about 15 m) in the initial stage of the follower vehicles, but all follower vehicles quickly converge to the expected displacement within 20 s.

[0175] In the steady stage (t>30s), the displacement tracking error is less than 0.5m, and the vehicle platoon keeps a stable inter-vehicle distance. Fig. 2(b) and Fig. 2(c) show the tracking performance of velocity and acceleration, respectively. The velocity tracking error converges to zero within 5s, and the steady-state velocity error is less than 0.1m / s. The acceleration response is smooth and without overshoot, and the control input amplitude is reasonable, which indicates that the designed distributed event-triggered controller has good dynamic response characteristics and ensures the stability of vehicle driving. Fig. 3(a) gives the dynamic curve of inter-vehicle distance error. The maximum inter-vehicle distance error occurs at t≈3 (about 1.2m), and all following vehicles achieve inter-vehicle distance error convergence within 15s. It is worth noting that the tracking error presents a decay trend when it propagates backward along the vehicle platoon, which meets the chordal stability condition. In the steady stage, the inter-vehicle distance error tends to zero, which verifies the effectiveness of the present application. Figure 4 The control input curve of the first following vehicle is plotted. In the initial stage (t<10s), the maximum absolute value of control input reaches 5m / s 2 to quickly eliminate the initial deviation; in the transition stage (10s<t<30s), the control amplitude decays exponentially; after entering the steady stage (t>30s), the control input tends to zero. Figure 5 The event-triggered time diagram of the first following vehicle is counted. The results show that the total trigger number is 56 (not including the initial trigger), the average trigger interval is 1.1s, the minimum trigger interval is 0.4s, and the maximum trigger interval is 2.8s. In the convergence stage (t<30s), the trigger frequency is relatively high (about 1.2 times / s) to ensure rapid convergence; while in the steady stage (t>30s), the trigger frequency is significantly reduced to 0.3 times / s, which effectively reduces the communication load.

Claims

1. A vehicle queue cooperative control method based on a dynamic event triggering strategy, characterized in that, Includes the following steps: Step 1: Abstract the leader vehicle and following vehicles in the convoy as nodes, describe the communication relationship between vehicles according to algebraic graph theory, obtain the communication topology of the vehicle convoy, and collect the kinematic characteristics of the vehicles to establish kinematic models of the leader vehicle and following vehicles respectively. Step 2: Using a kinematic model, design a following vehicle spacing strategy that is related to the speed of the leading vehicle; Define the tracking error and state estimation error of the following vehicle to obtain the closed-loop error system; and design a distributed dynamic event triggering mechanism based on the closed-loop error system. Step 3: Design a distributed event-triggered controller; when the triggering conditions of the distributed dynamic event triggering mechanism are met, the following vehicle samples its own state and requests its neighbors to send the latest state to update the distributed event triggering controller, so that the control inputs of all following vehicles are dynamically updated, ensuring that the tracking error of the following vehicle does not amplify and propagate along the queue when the convoy is running stably.

2. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The process of describing the communication relationships between vehicles based on algebraic graph theory to obtain the communication topology of the vehicle queue includes: considering a queue length of N+1, containing 1 leader vehicle, labeled i=0; and N following vehicles, labeled i=1,2,...,N; using an undirected graph... To model the communication topology, where The set of points consists of nodes abstracted from N following vehicles, v i This represents the node abstracted from the i-th vehicle; Let be the edge set, representing whether vehicles i and j have communication capabilities; It is an adjacency matrix used to represent the connection state between all following vehicles in the queue; when there is a bidirectional communication connection between vehicle i and vehicle j, that is... a ij =1; otherwise a ij =0; Simultaneously, define the diagonal matrix of the state connection between the leading vehicle and the following vehicles, i.e., the traction matrix. a i0 This indicates whether there is a communication connection between the following vehicle and the lead vehicle; if so, then a i0 =1; otherwise a i0 =0; Assume For an undirected graph, the Laplacian matrix is ​​defined as follows: in Determine the degree of a node; augmenting undirected graph Includes original image And an additional node 0, whose information is only relevant to the original graph. A subset of nodes in the dataset is available, and node 0 does not receive data from... Information about any node in the process.

3. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The kinematic models for the leading vehicle and the following vehicle are established as follows: Kinematic model of the leading vehicle: p0(t) and v0(t) represent the displacement and velocity of the lead vehicle, respectively; Kinematic model of following vehicle: Where, p i (t), v i (t) and a i (t) represent the displacement, velocity, and acceleration of the following vehicle i at time t, respectively. τ represents the control input of the i-th following vehicle, and τ is the inertial time constant of the vehicle's powertrain system; Rewrite the state of the following vehicle in vector form x. i (t)=[p i (t)v i (t)a i (t)] T Then the kinematic equations of the following vehicle can be rewritten as: in 4. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The following vehicle spacing strategy related to the speed of the leading vehicle includes: The expression for the dynamic distance between following vehicles is: The expression for the expected distance between the i-th car and the j-th car is d. ij (t)=d i0 (t)-d j0 (t), where d i0 To track the headway between vehicle i and the leading vehicle 0, d j0 (t) represents the headway between the tracking vehicle j and the leading vehicle 0, and d ij To track the headway between vehicle i and vehicle j, b i0 and c i0 is the spacing strategy parameter; s is a positive constant representing the minimum expected distance between two adjacent vehicles when they are stationary, and l is the vehicle length.

5. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The definition of the tracking error and state estimation error of the following vehicle is as follows: The tracking error between following vehicle i and following vehicle j is defined as: z i (t)=p i (t)-p j (t)+d ij (t) (5) Define the tracking error variables for the displacement, velocity, and acceleration of the following vehicle i as follows: Define the state estimation errors of displacement, velocity, and acceleration of following vehicle i as: in, and To estimate the displacement, velocity, and acceleration of the i-th vehicle within the interval between two adjacent triggering times; The tracking error vector and state estimation error vector of the following vehicle i are defined as follows:

6. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The closed-loop error system includes the closed-loop error system of the i-th following vehicle: To provide a unified description of the state of the entire vehicle platoon, a closed-loop error system for the entire platoon is constructed: Let ε(t) = [ε1(t)ε2(t)…ε N (t)] T and e(t)=[e1(t)e2(t)…e N (t)] T Then we can obtain a compact form of the closed-loop error system: This equation describes the dynamic evolution of the state error of the entire vehicle platoon, where I N It is an N-dimensional identity matrix. Let A and B represent the Kronecker product, where A and B are system matrices.

7. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The distributed dynamic event triggering mechanism includes the update law of triggering conditions and the dynamic threshold of following vehicles; The triggering condition is: The update law for the following vehicle dynamic threshold is: Discrete trigger times are described as a monotonically increasing time series. Let represent the (k+1)th trigger time of vehicle i, where inf is the infimum; the dynamic threshold for each following vehicle is η. i (t), η i (0)>0, θ, β i and σ i Let θ be a threshold parameter, satisfying θ > 0, β i >0, 0<σ i <1; Each following vehicle communicates and updates its own control input only when the triggering condition is met; Ricardi equation PA+A T P-γPBB T The unique positive definite solution P = -I is given by designing the feedback gain matrix Γ = PBB. T P, defining the matrix μ is the design parameter.

8. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, The designed distributed event-triggered controller is as follows: The control input representing the vehicle i is used to construct a distributed event-triggered controller for following vehicle i as follows: Where K p K v K a To control the gain through feedback; define a matrix. Then equation (14) can be rewritten as: where, D ij = [d ij (t) 0 0] T , D i0 = [d i0 (t) 0 0] T .

9. The vehicle queue cooperative control method based on a dynamic event triggering strategy according to claim 1, characterized in that, In step 3, when the triggering conditions of the distributed dynamic event triggering mechanism are met, If the time-triggered value replaces the estimated value, then the control input for the i-th following vehicle is: If for any initial state x0(0) and x i (0), control input exists. Make If all following vehicles meet the requirement, the entire vehicle convoy achieves the leader-follower consistency goal, ensuring that the tracking error of the following vehicles does not amplify and propagate along the convoy when the convoy is running stably. The leader-follower consistency goal is that, as time goes by, the following vehicles need to track the movement state of the leader vehicle so that the difference between the movement states of all following vehicles and the leader vehicle approaches zero.

10. An electronic device, characterized in that, It includes a processor and a memory; the memory is used to store a computer program that is executed by the processor to perform the method described in any one of claims 1-9.

Images