Intelligent connected vehicle platoon robust control method under communication failure condition

Abstract

The application aims at the networked automatic driving vehicle formation system under network attack, and proposes a robust control framework considering the change of communication topology structure. The random characteristics of deception attack in the formation system are modeled, and an adaptive event-triggered robust control model is adopted to ensure the reliability of the formation control. Finally, traffic flow simulation experiments are carried out to verify the effectiveness of the formation control model and the adaptive event-triggered strategy.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation systems technology, and specifically relates to a robust control method for intelligent connected vehicle platooning under communication failure conditions. Background Technology

[0002] Safety, energy conservation, and environmental protection have always been the core goals of road traffic system development. Over the past three decades, alongside the rapid growth of the automotive industry, motorized travel, while improving social efficiency, has also brought systemic problems such as road resource overload, frequent traffic accidents, and increased exhaust pollution. Compared to large-scale road network expansion and other engineering-based modifications, optimizing vehicle technical architecture and improving the operational efficiency of traffic facilities has become a more economical and sustainable solution. Against this backdrop, breakthroughs in intelligent connected and autonomous driving technologies have provided new directions for transportation system innovation. In an intelligent connected transportation environment, vehicles establish multi-layered data links with roadside facilities and cloud platforms through onboard communication units, constructing a collaborative "edge-cloud" three-dimensional perception architecture. This architecture not only supports real-time exchange of dynamic parameters such as speed and acceleration between vehicles and traffic elements, but also utilizes V2X (Vehicle to Everything) communication to acquire environmental information such as traffic signal timing and road anomalies. By integrating multi-dimensional perception data and collaborative decision-making algorithms, autonomous driving systems can dynamically optimize longitudinal following models and lateral obstacle avoidance logic, improving driving comfort while ensuring safety redundancy.

[0003] However, intelligent connected vehicle platooning still faces multiple real-world constraints in practical applications, with the core challenge lying in the deep coupling between the technological system and the traffic ecosystem. In mixed traffic scenarios, the behavioral differences between human-driven vehicles (HVs) and autonomous vehicle platoons can easily lead to dynamic game theory problems. The random lane-changing and conservative following behaviors of traditional vehicles can easily disrupt the spatiotemporal consistency of the platoon, and sudden interference events are more likely to affect the stability of the control strategy. These uncertain interactions not only enhance the nonlinear characteristics of the system but also force the cooperative control algorithm to dynamically balance safety and efficiency, placing higher demands on the robustness of the control architecture.

[0004] Meanwhile, the vulnerability of communication systems also poses a critical challenge. Complex environments such as urban canyons and tunnel obstructions can lead to V2X communication delays and packet loss, affecting the effective transmission of status information between vehicles. Once backbone node communication fails, the disturbance propagation mechanism in traditional control architectures may cause the accumulation of spacing errors between vehicles, or even lead to the disintegration of the formation structure. Therefore, the control strategy must have the ability to detect and handle communication anomalies, and be able to maintain the basic operational logic of the formation under communication constraints.

[0005] In conclusion, the platooning control of intelligent connected vehicles under unstable communication network conditions requires more attention. Summary of the Invention

[0006] In view of this, the present invention proposes a robust control method for intelligent connected vehicle platooning under communication failure conditions, which can maintain the stable operation of the platoon in communication unstable scenarios and avoid vehicle disintegration caused by brief communication failures, thereby avoiding frequent platoon reorganization. This provides certain theoretical guidance for improving vehicle traffic efficiency and reducing accident rate.

[0007] To solve the above-mentioned technical problems, the technical solution provided by the present invention is as follows:

[0008] Firstly, a robust control method for intelligent connected vehicle platooning under communication failure conditions includes the following steps:

[0009] Construct vehicle platooning models with different topologies under different communication conditions; wherein, the communication conditions include normal communication and communication failure.

[0010] A robust control model is established based on the vehicle platooning state equations of the aforementioned topology.

[0011] The stability of the robust control model is proven.

[0012] Define the algorithm flow; where LQR is used as the feedback controller when the vehicle formation is in a stable state; when the vehicle formation is in a robust state, a combination of robust control MPC and LQR feedback is adopted.

[0013] Specifically, the construction of vehicle platooning models with different topologies under different communication conditions includes:

[0014] Determine the vehicle platoon communication topology and car-following relationships, including:

[0015] Under normal communication conditions, the vehicles in the formation establish two-way communication with their adjacent vehicles to obtain information and send control commands; under communication failure conditions, the vehicle with communication failure executes the ACC control strategy; the vehicle behind the vehicle with communication failure passes over the vehicle with communication failure to establish communication with the vehicle in front, maintaining the function of information transmission and control command transmission.

[0016] Constructing the Laplace matrix of the vehicle platooning car-following topology includes:

[0017] make For vehicle platooning, a car-following adjacency matrix is ​​used. The following departure degree matrix for vehicle platooning represents the member vehicles that a vehicle is following when updating its state; The Laplace matrix is ​​used to calculate the state difference between car-following vehicle pairs in the vehicle platooning topology.

[0018] Reconstructing the vehicle formation dynamics model, including:

[0019] Constructing a vehicle platoon dynamics model:

[0020] ;

[0021] in, For intelligent connected vehicles The longitudinal position;

[0022] For intelligent connected vehicles speed;

[0023] For intelligent connected vehicles The acceleration;

[0024] This represents the time step of the discrete model.

[0025] For discrete-time indexing;

[0026] ;

[0027] when The time indicates the lead vehicle, whose status is not controlled by the formation model;

[0028] Construct the vehicle platoon state equations:

[0029] ;

[0030] Among them, vehicles The status of the vehicle and The position and velocity were calculated.

[0031] Combining the matrix form of the car-following topology and the Laplace matrix, and generalizing them to a general form, we construct the state expression for vehicle formation:

[0032] ;

[0033] in This refers to the longitudinal spacing between the vehicles in the formation;

[0034] The speed difference between the vehicles in the formation;

[0035] ;

[0036] Based on the expressions for the distance and speed difference between the vehicle and the vehicle in front, a state prediction matrix for vehicle formation is constructed:

[0037] ;

[0038] The results were:

[0039] ;

[0040] Constructing the error matrix Indexes used to represent vehicles with communication failures and lead vehicles:

[0041] ;

[0042] From the matrix, we can know , representing the vehicle index without state prediction error;

[0043] Rewrite the vehicle platoon state matrix:

[0044] ;

[0045] in, The predicted acceleration of an uncontrollable vehicle is expressed as:

[0046] ;

[0047] Errors between the spacing and speed difference of vehicles in the formation and the expected values:

[0048] ;

[0049] in, ;

[0050] Substitute the errors between the spacing and speed difference of vehicles in the formation and the expected values ​​into the state matrix of the vehicle formation to construct the state difference equation:

[0051] ;

[0052] After transposition, construct the state prediction equation:

[0053] ;

[0054] The expected speed difference between vehicles within the formation is zero. Substituting it into the above formula, we get:

[0055] ;

[0056] remember The above formula can be rearranged as follows:

[0057] .

[0058] Among them, matrix and The expression is: ;

[0059] The acceleration in the above formula Replace with acceleration increment ,get

[0060] ;

[0061] in, For state variables, each matrix is ​​defined as follows:

[0062] ;

[0063] Build H Predict the vehicle platooning state equations in the time domain:

[0064] ;

[0065] in, It combines the predicted state and error of uncontrollable vehicles;

[0066] The specific definitions of each state variable, control variable error term, and each matrix are as follows:

[0067] ;

[0068] ;

[0069] .

[0070] Specifically, the construction of a robust control model based on the vehicle platooning state equations of the aforementioned topology includes:

[0071] Set control model constraints, including:

[0072] Set upper and lower bound constraints:

[0073] ;

[0074] Set acceleration boundary constraints:

[0075] ;

[0076] Set vehicle platooning safety constraints:

[0077] ;

[0078] in, ;

[0079] ;

[0080] ;

[0081] ;

[0082] The optimization objectives for establishing a robust control model include:

[0083] State the optimization objective of the formation control model:

[0084] ;

[0085] in, ,

[0086] It is a symmetric penalty matrix;

[0087] For the state penalty submatrix;

[0088] The penalty submatrix for the control quantity;

[0089] The terminal penalty matrix is ​​used to ensure the stability of the control system;

[0090] based on Reconstruct the objective function from the state equations in the time domain:

[0091] ;

[0092] in, ;

[0093] ;

[0094] The objective of robust control optimization is:

[0095] ;

[0096] in, ;

[0097] This is the vector that perturbs the upper bound;

[0098] Rewriting security constraints includes:

[0099] Set the predicted location of uncontrollable vehicles:

[0100] ;

[0101] in, For the predicted acceleration of the uncontrollable vehicle in the prediction time domain, the safety constraints are rewritten as follows:

[0102] ;

[0103] The robust safety constraint expression is obtained as follows:

[0104] ;

[0105] in, ;

[0106] Relaxing security constraints includes:

[0107] ;

[0108] in, The relaxation term represents the error that may occur in controlling the model under strong disturbance conditions; at the same time, hard constraints are added to ensure a minimum safe distance.

[0109] ;

[0110] in, ;

[0111] Increase the objective function for relaxation terms To construct a complete control model, the penalty term is added to the objective function for the relaxation term. The penalty item is:

[0112] ;

[0113] in, Let be the penalty matrix for the relaxation term;

[0114] The complete control model is as follows:

[0115] ;

[0116] .

[0117] Specifically, the stability proof of the robust control model includes:

[0118] Transition matrix and control matrix Transform into:

[0119] ;

[0120] Prove the stability.

[0121] Specifically, the algorithm definition process includes:

[0122] Initialize all parameters;

[0123] When the vehicle platooning is in a stable state, LQR is used as the feedback controller, and the expression of the LQR controller is:

[0124] ;

[0125] in, For the feedback gain matrix, To expand control inputs;

[0126] When the vehicle formation is in a robust state, a scheme combining robust control MPC and LQR feedback is adopted: feedback control is used to track the trajectory, and the error between the current state and the planned state is calculated at each time step and corrected using the LQR controller. The state error at time step k is:

[0127] ;

[0128] in, for Tracking deviation at time steps; for The actual state of the time step; for The planning status of the time step;

[0129] In the expression of the LQR controller during the formation steady state Change to formation tracking deviation The tracking LQR controller expression is obtained as follows:

[0130] ;

[0131] in, To track and correct the control quantity;

[0132] Combine the planning input and feedback input into a control input:

[0133] .

[0134] In a second aspect, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the above-described method.

[0135] Thirdly, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the steps of the above-described method.

[0136] The beneficial effects of this invention are:

[0137] This invention addresses the risk of communication failures in real-world intelligent connected vehicle platooning systems by proposing a robust control framework that considers changes in communication topology. It models the stochastic characteristics of spoofing attacks in the platooning system and employs an adaptive event-triggered robust control model to ensure the reliability of platooning control. Attached Figure Description

[0138] The present invention includes the following figures:

[0139] Figure 1 This is a diagram of the vehicle platooning communication topology and following structure under normal communication conditions according to the present invention;

[0140] Figure 2 This is a diagram of the vehicle platooning communication topology and following structure under communication failure conditions according to the present invention.

[0141] Figure 3 This is a comparison chart of acceleration increment changes under continuous acceleration and deceleration conditions without communication failure, as presented in this invention.

[0142] Figure 4 This is a comparison diagram of acceleration changes under continuous acceleration and deceleration conditions without communication failure, as presented in this invention.

[0143] Figure 5 This is a comparison chart of acceleration increment changes under continuous acceleration and deceleration conditions in the context of communication failure according to the present invention;

[0144] Figure 6 This is a comparison chart of acceleration changes under continuous acceleration and deceleration conditions during communication failure, as presented in this invention.

[0145] Figure 7 This is a comparison chart of acceleration increment changes under rapid deceleration conditions without communication failure, as presented in this invention.

[0146] Figure 8 This is a comparison chart of acceleration changes under rapid deceleration conditions without communication failure, as presented in this invention.

[0147] Figure 9 This is a comparison chart of acceleration increment changes under rapid deceleration conditions due to communication failure, as presented in this invention.

[0148] Figure 10 This is a comparison chart of acceleration changes under rapid deceleration conditions due to communication failure, as presented in this invention.

[0149] Figure 11 This is a diagram showing the change in safe distance between vehicles in a platoon under continuous acceleration and deceleration conditions according to the present invention.

[0150] Figure 12 This is a diagram showing the change in safe distance between vehicles in a platoon under rapid deceleration conditions according to the present invention.

[0151] Figure 13This is a comparison diagram of the spatiotemporal diagrams of the normal scenario and the communication failure scenario of the present invention;

[0152] Figure 14 This is a comparison diagram of the model switching process under normal and communication failure scenarios according to the present invention;

[0153] Figure 15 The results are simulations of the nominal model under the communication failure scenario of this invention. Detailed Implementation

[0154] To make the objectives, advantages and features of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0155] In one specific embodiment, a robust control method for intelligent connected vehicle platooning under communication failure conditions includes the following steps:

[0156] Step 1: Determine the vehicle topology under different communication conditions, construct the Laplace matrix of the vehicle platooning and car-following topology, and reconstruct the vehicle platooning dynamics model.

[0157] Step 1.1: Determine the vehicle platoon communication topology and car-following relationships. Divide the vehicle platoon communication topology and car-following relationships into two categories: ① Normal communication ② Communication failure;

[0158] (1) Communication is normal

[0159] The vehicle platooning communication topology and car-following structure under normal communication conditions are as follows: Figure 1 As shown, vehicles in a platoon can establish two-way communication with their neighboring vehicles to obtain information and send control commands. The blue arrows in the diagram represent the following relationship of the vehicle platoon in the platoon control model. During the journey, each vehicle acquires the position, speed, and acceleration information of the vehicle in front, and calculates the distance and speed difference between itself and the vehicle in front based on this information.

[0160] (2) Communication failure

[0161] Vehicle platooning communication topology and car-following structure under communication failure conditions, as follows Figure 2 As shown, the red vehicles represent those with communication failures, unable to establish communication with adjacent vehicles and thus executing the ACC control strategy. In this situation, vehicles following the communication-deficient vehicle will pass it and establish communication with the vehicle in front, thereby maintaining information transmission and control command delivery. For following vehicles, information about the vehicle in front can still be obtained in real time through sensors, but its future movement cannot be detected.

[0162] Step 1.2: Construct the Laplace matrix of the vehicle platooning and car-following topology. Let... For vehicle platooning, a car-following adjacency matrix is ​​used. The following departure degree matrix for vehicle platooning represents the member vehicles that a vehicle is following when updating its state. The Laplace matrix for the car-following topology of vehicle platooning can be used to calculate the state difference between car-following vehicle pairs.

[0163] Step 1.3: Reconstruct the vehicle formation dynamics model.

[0164] (1) The vehicle formation dynamics model is as follows:

[0165] ;

[0166] in, For intelligent connected vehicles The longitudinal position; For intelligent connected vehicles speed; For intelligent connected vehicles The acceleration; This represents the time step of the discrete model. For discrete-time indexing; ;when The time indicates the lead vehicle, whose status is not controlled by the formation model;

[0167] (2) Construction of vehicle formation state equations

[0168] The distance and speed difference between the vehicle and the vehicle in front can be expressed as:

[0169] ;

[0170] In this equation, the vehicle The status of the vehicle and The position and velocity were calculated.

[0171] In a specific embodiment, taking a platoon of four intelligent connected vehicles as an example, the vehicle index is: i =0,1,2,3, Number of following vehicles N =3, i =0 indicates the lead vehicle.

[0172] The spacing between vehicles in a formation can be expressed as:

[0173] ;

[0174] Rearranging the matrix into matrix form, we get:

[0175] ;

[0176] This equation can be expressed using the Laplace matrix of a race-following topology:

[0177] ;

[0178] in, ;

[0179] Similarly, the speed difference of the formation can be obtained:

[0180] ;

[0181] Combining the matrix form of the carousel topology and the Laplace matrix, and generalizing it to a general form, we can obtain the state expression for the formation:

[0182] ;

[0183] in, This refers to the longitudinal spacing between the vehicles in the formation; The speed difference between the vehicles in the formation; ;

[0184] From the expressions for the distance and speed difference between the vehicle and the vehicle in front, the vehicle formation state expression can be derived into a vehicle formation state prediction matrix:

[0185] ;

[0186] The results were:

[0187] ;

[0188] Since the lead vehicle in a platoon is not controlled by the platoon model, its acceleration needs to be adjusted from the above formula. This is extracted from the equation. Furthermore, if a vehicle in the platoon experiences a communication failure, its following topology will change, and the communication-failed vehicle will be uncontrollable. Therefore, the states of both the communication-failed vehicle and the platoon leader cannot be accurately predicted; thus, their predicted states should be integrated and added to the above equation. For ease of description, an error matrix is ​​constructed. Indexes representing vehicles with communication failures and lead vehicles:

[0189] ;

[0190] Therefore, we know the matrix. , representing the vehicle index with no state prediction error.

[0191] The vehicle formation state matrix can be rewritten as:

[0192] ;

[0193] in, The predicted acceleration of an uncontrollable vehicle is expressed as:

[0194] ;

[0195] The rewritten vehicle formation state matrix uses the known matrix. reserve The controllable parts are transferred to the uncontrollable parts. This achieves decoupling between the control quantity and the prediction quantity.

[0196] Errors between the spacing and speed difference of vehicles in the formation and the expected values:

[0197] ;

[0198] in, ;

[0199] Substituting the above equation into the state matrix of the vehicle formation, we can derive the state difference equation:

[0200] ;

[0201] The state prediction equation is obtained by transposing the equations:

[0202] ;

[0203] Because the expected speed difference between vehicles within a platoon is zero, therefore Substituting it into the above formula, we get:

[0204] ;

[0205] remember The above formula can be rearranged as follows:

[0206] .

[0207] Among them, matrix and The expression is:

[0208] ;

[0209] The acceleration in the above formula Replace with acceleration increment ,

[0210] get:

[0211] ;

[0212] in, As state variables, each matrix is ​​defined as follows:

[0213] ;

[0214] Based on the above formula, we can obtain... H The state equation in the time domain of step prediction:

[0215] ;

[0216] in, This is a combination of the predicted state and error of an uncontrollable vehicle. The specific definitions of each state variable, control variable, error term, and matrix are as follows:

[0217] ;

[0218] ;

[0219] ;

[0220] Step 2: Based on the vehicle formation state equations considering the communication topology, design a robust control model and determine the model's constraints and optimization objectives.

[0221] Step 2.1: Determine model constraints. First, the speed of controllable vehicles needs to meet the requirements of the upper and lower bounds. Then, considering the acceleration boundary constraints, due to the dynamic system, acceleration also needs to be constrained. Finally, the vehicle formation also needs to meet safety constraints.

[0222] (1) The speed of controllable vehicles must meet the upper and lower limits:

[0223] ;

[0224] in, This is the lower bound of the velocity; This is the upper bound of the velocity; It is a unit vector; For vehicle platooning velocity vector;

[0225] The upper and lower bound linear constraints are in the following forms:

[0226] ;

[0227] The specific form of the constraint matrix is ​​as follows:

[0228] ;

[0229] (2) Acceleration boundary constraints:

[0230] Due to limitations of the power system, the acceleration must satisfy the following formula:

[0231] ;

[0232] The constraint matrix is ​​as follows:

[0233] ;

[0234] (3) Vehicle platooning safety constraints:

[0235] Vehicle safety constraints k The expression for time is:

[0236] ;

[0237] in, for Predicted location of vehicles that are constantly out of control;

[0238] Safety constraint expressions for the entire prediction time domain:

[0239] ;

[0240] The results were:

[0241] ;

[0242] in,

[0243] ; ;

[0244] Step 2.2: Establish the optimization objective of the robust control model. The optimization objective of the formation control model can be expressed as:

[0245] ;

[0246] in, It is a symmetric penalty matrix; For the state penalty submatrix; The penalty submatrix for the control quantity; The terminal penalty matrix is ​​used to ensure the stability of the control system;

[0247] based on H The reconstructable objective function of the state equation in the prediction time domain is:

[0248] ;

[0249] in, The robust control optimization objective is obtained as follows: ;

[0250] in, This is the vector that perturbs the upper bound;

[0251] Step 2.3: Rewrite the safety constraints. The error terms in the original safety constraints need to be processed, so the safety constraints need to be rewritten as follows:

[0252] Predicted location of uncontrollable vehicles:

[0253] ;

[0254] in, For the predicted acceleration of the uncontrollable vehicle in the prediction time domain, the safety constraints are rewritten as follows:

[0255] ;

[0256] because Therefore, we can compress the above equation to obtain the robust safety constraint expression:

[0257] ;

[0258] in, ;

[0259] Step 2.4: Relax the constraints. To ensure that the control model always has a solution under all conditions, the safety constraints are relaxed as follows:

[0260] ;

[0261] in, This is the relaxation term, representing the possible error in controlling the model under strong disturbances. Hard constraints are also added to ensure a minimum safe distance.

[0262] ;

[0263] in, ;

[0264] Step 2.5: Increase the objective function for the relaxation term To construct a complete control model, a penalty term is added to the objective function for the relaxation term. The penalty item is:

[0265] ;

[0266] in, Here is the penalty matrix for the relaxation term; the complete control model is as follows:

[0267] ;

[0268] ;

[0269] Step 3: Perform stability verification on the multi-step predictive MPC control model to ensure that the model's control output can always restore the system from a disturbance state to a steady state. This includes:

[0270] The expressions for the state transition matrix and control matrix are:

[0271] ;

[0272] in, ,therefore The rank is .because Transition matrix and control matrix It can be transformed into:

[0273] ;

[0274] Prove the stability.

[0275] Step 4: Define the algorithm flow, which includes:

[0276] Before training the model, all parameters, including the vehicle platoon length, need to be initialized first. Prediction model prediction steps Vehicle blocking distance Vehicle safety distance wait.

[0277] When the vehicle platooning is in a stable state, LQR is used as the feedback controller, and the expression for the LQR controller is:

[0278] ;

[0279] in, This is the feedback gain matrix; To expand control inputs;

[0280] When the formation is in a robust state, a scheme combining robust control (MPC) and LQR feedback is adopted. The system uses feedback control to track the trajectory, calculates the error between the current state and the planned state at each time step, and corrects it using the LQR controller. The state error at time step k is:

[0281] ;

[0282] in, for Tracking deviation at time steps; for The actual state of the time step; for The planning status of the time step;

[0283] At this point, the expression for the LQR controller in the formation steady state will be... Change to formation tracking deviation The tracking LQR controller expression is obtained as follows:

[0284] ;

[0285] in, To track and correct the control quantity;

[0286] The final control input is a combination of the planning input and the feedback input:

[0287]

[0288] To verify the effectiveness of the robust control model that takes into account communication failures, a preferred embodiment of the present invention conducted a single-formation performance evaluation experiment and a traffic flow simulation experiment to analyze the model's performance in various scenarios.

[0289] Step 6: Analysis of model simulation results.

[0290] (1) Performance evaluation of a single formation

[0291] The single-formation performance evaluation was used to verify the anti-disturbance capability of the single-vehicle formation when subjected to external disturbances. The disturbance scenarios used in the simulation experiments were continuous acceleration and deceleration of the preceding vehicle and rapid acceleration and deceleration. The former was to test the formation control model's ability to track disturbances, and the latter was to test the safety of the control model. The simulation experimental parameters are shown in Table 1.

[0292] Figure 3 and Figure 4 The acceleration variation curves of the robust model and the nominal model are shown respectively under the condition of continuous acceleration and deceleration of the preceding vehicle, without communication failure, demonstrating the ability to resist disturbances in this scenario.

[0293] Figure 5 and Figure 6 The control performance of the robust model and the nominal model is shown respectively when a vehicle in the platoon experiences communication failure. Vehicles 2 and 4 experience communication failure at the 5th second and are no longer controlled by the platooning model; instead, they update their states using the IDM model. Because the steady-state distance of the IDM model after the communication failure is greater than the expected distance of the vehicles in the platooning control model, a deceleration operation is performed. As can be seen from the figure, after the communication failure event, the dynamics of the vehicle model differ significantly from the original cooperative platooning control model, resulting in a sharp change in acceleration and causing significant disturbance to subsequent vehicles. It can be observed that both the robust model and the nominal model promptly adjust the states of subsequent vehicles after the communication failure event, restoring the vehicle platoon to a stable operating state. Compared to continuous acceleration and deceleration scenarios, rapid acceleration and deceleration scenarios better demonstrate the control model's ability to ensure safety.

[0294] Figure 7 and Figure 8 The acceleration variation curves of the robust model and the nominal model are shown separately under conditions of rapid acceleration and deceleration, without communication failure. The figure shows the acceleration variation curves of the lead vehicle moving at -4... m / s 2 The system decelerates at its maximum deceleration, allowing subsequent vehicles in the platoon to track smoothly and effectively suppress disturbances. All vehicles use the same initial state; therefore, compared to the nominal control model, the robust model performs a slight deceleration at the beginning of the experiment to adaptively adjust the robust spacing.

[0295] Figure 9 and Figure 10 This illustrates the control performance under communication failure conditions. Vehicle 2 and Vehicle 4 experience a communication failure at the 5th second, resulting in a sudden change in their states and causing disturbances to following vehicles. As shown in the figure, both the nominal control model and the robust control model respond promptly to the state change of the communication-failed vehicles. However, the robust control model, due to the need to reserve a larger safety margin, exhibits a larger response amplitude.

[0296] Figure 11 Figures -a and 11-b illustrate the changes in safe spacing between vehicles under continuous acceleration / deceleration and rapid acceleration / deceleration conditions, respectively. In the disturbance experiment, communication between vehicles 2 and 4 failed at the 5th second, altering their dynamic characteristics. Therefore, the vehicle spacing in the figure also changed significantly at the 5th second. Figure 11 As shown in sub-figure (a), when the vehicle in front suddenly decelerates, the spacing of the nominal model decreases rapidly, but the spacing of the robust control model decreases less and then quickly returns to its original state. Furthermore, due to the consideration of the uncertainty of the vehicle in front's state during communication failure, the reserved spacing of the robust control model is significantly larger than that of the nominal control model. Compared to continuous acceleration and deceleration scenarios, the change in safe spacing is more pronounced in rapid acceleration and deceleration scenarios.

[0297] like Figure 12 As shown, when the lead vehicle decelerates rapidly, the spacing between vehicles within the platoon is relatively small. Therefore, when a communication failure occurs, the smaller spacing causes the vehicle experiencing the communication failure to decelerate more, resulting in a stronger disturbance to following vehicles. Figure 12 When vehicles decelerate significantly, both control models follow the deceleration of the vehicle in front, but the robust control model has a significantly larger safety margin in terms of vehicle spacing. When the spacing of the nominal model approaches the safe spacing, the robust model still has a large margin of safety to prepare for possible uncertainties in the future.

[0298] Table 1 Simulation Parameter Settings for Single Formation Performance Evaluation

[0299]

[0300] (2) Traffic flow simulation experiment

[0301] To verify the performance of the robust control model in traffic flow simulation experiments, traffic flow simulation experiments were conducted, and the experimental parameters are shown in Table 2. In the simulation experiments, intelligent connected vehicles forming a platoon were based on preset probabilities. p f ∈ [0 , 1] Communication failure occurred. This section uses... p f =0 . 05. Considering that if the lead or tail vehicle in the formation experiences a communication failure, it can leave the formation directly without affecting the status of other vehicles in the formation, the location of the communication failure vehicle is completely randomized in this experiment to simplify the experiment, but it will not appear at the positions of the lead or tail vehicle. When a vehicle experiences a communication failure, the failure time will last [1s, 10s].

[0302] Table 2 Traffic flow simulation experiment parameter settings

[0303] parameter Value Maximum formation length 10 Number of iterations 4000 Simulation time step 0.1 s Main lane vehicle arrival frequency 1800 vehicles / hour Main lane length 3000 m ramp location 1500 m Speed ​​limit on main lane -4 km / h Vehicle Expected Speed 80 km / h

[0304] Figure 13 The simulation spatiotemporal diagrams for normal scenarios (no communication failure) and communication failure scenarios are shown for various CAV penetration rates. As can be seen from the diagrams, the spatiotemporal diagrams for different CAV penetration rates are basically consistent in both normal and communication failure scenarios. This indicates that, in the communication failure scenario, the control model proposed in this section effectively suppresses traffic flow disturbances caused by the abrupt state changes after communication failure by considering the uncertainties of the failed vehicles.

[0305] Figure 14 This figure illustrates the model switching process between normal (no communication failure) and communication failure scenarios at various CAV penetration rates. As shown in the figure, in the normal scenario, vehicle platooning employs feedback control in most stable operating states, switching to a robust control model and tracking mode under abrupt state changes such as near ramps. However, in the communication failure scenario, the randomness of vehicle communication failures leads to continuous changes in the vehicle platoon state, significantly disrupting stable operation. Figure 14 Figures (b), (d), and (f) show that when a vehicle communication failure occurs, the vehicle formation can promptly switch to a robust control model for replanning, and then adopt a tracking mode to reduce errors and suppress disturbances. Therefore, the vehicle formation can still operate smoothly and safely even when communication fails, proving the effectiveness of the robust control model and control framework proposed in this chapter.

[0306] In addition, the results of the nominal model (the model without robust handling) were compared in the communication failure scenario. Figure 15The figures show the simulation results of the nominal model under communication failure scenarios at 80% and 100% penetration rates. As can be seen from the figures, there is a significant performance gap between the nominal model and the robust control model under communication failure scenarios. Figure 15 Figure (c) shows that in a fully CAV scenario, communication failure causes some vehicles to revert to their original functions, resulting in inconsistent vehicle dynamics within the platoon and thus causing some degree of traffic congestion. Figure 14 In both normal scenarios and communication failure scenarios using robust control, there is almost no traffic congestion at 100% penetration.

[0307] Table 3 lists traffic flow indicators under different scenarios and models. The table shows that robust control in normal scenarios and robust control models considering changes in communication topology in communication failure scenarios achieve similar results across all indicators. However, in communication failure scenarios using the nominal control model, because robust redundancy is not provided for communication failure risks, all indicators significantly lag behind the results of robust control at different CAV penetration rates.

[0308] Table 3 Comparison of traffic flow indicators under different scenarios and control models

[0309]

[0310] It should be noted that any process or method description in the embodiments can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order according to the functions involved, as should be understood by those skilled in the art to which the embodiments of the invention pertain.

[0311] It should be noted that the logic and / or steps in the embodiments, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0312] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0313] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0314] Furthermore, the functional modules described in this invention can be integrated into a single processing module, or each module can exist as a separate physical entity, or two or more modules can be integrated into a single module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0315] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0316] The above embodiments have provided a detailed description of the technical solutions of the present invention. Obviously, the present invention is not limited to the described embodiments. Based on the embodiments of the present invention, those skilled in the art can make various changes, but any changes that are equivalent or similar to the present invention fall within the scope of protection of the present invention. Contents not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A robust control method for intelligent connected vehicle platooning under communication failure conditions, characterized in that, Includes the following steps: Construct vehicle platooning models with different topologies under different communication conditions; wherein, the communication conditions include normal communication and communication failure. A robust control model is established based on the vehicle platooning state equations of the aforementioned topology. The stability of the robust control model is proven. Define the algorithm flow; where, when the vehicle formation is in a stable state, LQR is used as the feedback controller; when the vehicle formation is in a robust state, a scheme combining robust control MPC and LQR feedback is adopted. The robust control model constructed based on the vehicle platooning state equations of the aforementioned topology includes: Set control model constraints, including: Given the prediction time domain is Set upper and lower bound constraints for velocity: ; L = IF , representing the vehicle index without state prediction error. As a lower bound for velocity, This is the upper bound of the velocity; Set acceleration boundary constraints: ; For maximum deceleration, This is the maximum acceleration; Set vehicle platooning safety constraints: ； in, ; ; and , Vehicle congestion spacing; For safe headway; The optimization objectives for establishing a robust control model include: State the optimization objective of the formation control model: ； ； It is a symmetric penalty matrix; Let be the penalty submatrix of the state; The penalty submatrix for the control quantity; The terminal penalty matrix is ​​used to ensure the stability of the control system; based on H Reconstruct the objective function from the state equations in the time domain: ； in, ; ； The objective of robust control optimization is: ； in, ; Specifically, it is expressed as follows: ； ； ； in ,matrix A,B,C Derived from the vehicle dynamics formula, it is specifically expressed as: ； This is the vector that perturbs the upper bound; Rewriting security constraints includes: Set the predicted location of uncontrollable vehicles: ； in, For the predicted acceleration of uncontrollable vehicles in the prediction time domain, It is a unit vector; Rewrite the safety constraints as follows: ； To approximate the upper bound of the perturbation, we compress the above equation to obtain the robust safety constraint expression: ； in, ; Relaxing security constraints includes: ； in, The relaxation term represents the error that may occur in controlling the model under strong disturbance conditions; at the same time, hard constraints are added to ensure a minimum safe distance. ； in, ; Increase the objective function for relaxation terms To construct a complete control model, the penalty term is added to the objective function for the relaxation term. The penalty item is: ； in, Let be the penalty matrix for the relaxation term; The complete control model is as follows: ； ； ， 。 2. The method as described in claim 1, characterized in that, The construction of vehicle platooning models with different topologies under different communication conditions includes: Determine the vehicle platoon communication topology and car-following relationships, including: Under normal communication conditions, the vehicles in the formation establish two-way communication with their adjacent vehicles to obtain information and send control commands; under communication failure conditions, the vehicle with communication failure executes the ACC control strategy; the vehicle behind the vehicle with communication failure passes over the vehicle with communication failure to establish communication with the vehicle in front, maintaining the function of information transmission and control command transmission. Constructing the Laplace matrix of the vehicle platooning car-following topology includes: make For vehicle platooning, a car-following adjacency matrix is ​​used. The following departure degree matrix for vehicle platooning represents the member vehicles that a vehicle is paying attention to when updating its state; The Laplace matrix is ​​used to calculate the state difference between car-following vehicle pairs in the vehicle platooning topology. Reconstructing the vehicle formation dynamics model, including: Constructing a vehicle platoon dynamics model: ； ； in, For intelligent connected vehicles The longitudinal position; For intelligent connected vehicles speed; For intelligent connected vehicles The acceleration; ； when The time indicates the lead vehicle, whose status is not controlled by the formation model; Construct the vehicle platoon state equations: ； Among them, vehicles i The status of the vehicle i and i- The position and velocity of 1 were calculated; Combining the matrix form and Laplace matrix of the car-following topology, and generalizing them to a general form, we construct the state expression for vehicle formation: ； in, This refers to the longitudinal spacing between the vehicles in the formation; The speed difference between the vehicles in the formation; ； Based on the expressions for the distance and speed difference between the vehicle and the vehicle in front, a state prediction matrix for vehicle formation is constructed: ； The results were: ； Constructing the error matrix Indexes used to represent vehicles with communication failures and lead vehicles: ； From the matrix, we can know L = IF , representing the vehicle index without state prediction error; Rewrite the vehicle platoon state matrix: ； in, The predicted acceleration of an uncontrollable vehicle is expressed as: ； Errors between the spacing and speed difference of vehicles in the formation and the expected values: ； Substitute the errors between the spacing and speed difference of vehicles in the formation and the expected values ​​into the state matrix of the vehicle formation to construct the state difference equation: ； After transposition, construct the state prediction equation: ； The expected speed difference between vehicles within the formation is zero. Substituting it into the above formula, we get: ； remember The above formula can be rearranged as follows: ； Among them, matrix and The expression is: ； The acceleration in the above formula Replace with acceleration increment ,get ； in, For state variables, each matrix is ​​defined as follows: ； Build H Predict the vehicle platooning state equations in the time domain: ； in, It combines the predicted state and error of uncontrollable vehicles; The specific definitions of each state variable, control variable error term, and each matrix are as follows: ； ； 。 3. The method as described in claim 1, characterized in that, The stability proof of the robust control model includes: Transition matrix and control matrix Transform into: ； Prove the stability.

4. The method as described in claim 1, characterized in that, The defined algorithm flow includes: Initialize all parameters; When the vehicle platooning is in a stable state, LQR is used as the feedback controller, and the expression of the LQR controller is: ； in, This is the feedback gain matrix; To expand control inputs; When the vehicle formation is in a robust state, a scheme combining robust control MPC and LQR feedback is adopted: feedback control is used to track the trajectory, and the error between the current state and the planned state is calculated at each time step and corrected using the LQR controller. The state error at time step k is: ； in, for k Tracking deviation at time steps; for k The actual state of the time step; for k The planning status of the time step; In the expression of the LQR controller during the formation steady state Change to formation tracking deviation The tracking LQR controller expression is obtained as follows: ； in, To track and correct the control quantity; Combine the planning input and feedback input into a control input: 。 5. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-4.

Images