Lane change decision and time delay compensation control method and device based on prediction information

Abstract

This application relates to the field of vehicle technology, and particularly to a method and apparatus for lane-changing decision-making and delay compensation control based on predictive information. The method includes: acquiring traffic information collected by a roadside sensing unit; predicting traffic information of vehicles within the sensing range over a future period based on the traffic information, and constructing a multi-objective optimization problem of the target vehicle's future driving environment based on the predicted information; solving the multi-objective optimization problem to obtain the target vehicle's optimal expected speed and lane-changing decision time; and sending the expected speed and lane-changing decision time to the target vehicle. The target vehicle responds to the expected speed and lane-changing decision time using its own lane-changing safety judgment algorithm, lane-changing path planning algorithm, and delay compensation control algorithm to adjust its speed and change lanes in advance. This addresses the problems in related technologies, such as low matching degree between vehicle lane-changing planning and current traffic, and delays in the vehicle control system leading to poor control accuracy and stability, and low driving safety.

Description

Technical Field

[0001] This application relates to the fields of intelligent connected vehicle technology and autonomous lane changing technology of cloud control systems, and in particular to a lane changing decision and delay compensation control method and device based on predictive information. Background Technology

[0002] Lane changing is one of the basic driving behaviors of a vehicle. It requires consideration of the vehicle's lateral movement and the status of traffic in the target lane, which is quite complex. This makes it more likely that lane changing will conflict with other vehicles during the execution of the behavior, and has a significant impact on the traffic environment.

[0003] Among related technologies, an autonomous driving application based on road traffic prediction information can be used to focus on cruise technology, determining whether a lane change is necessary based on the current traffic environment faced by the vehicle.

[0004] However, when generating lane-changing decisions, if traffic vehicle prediction information is used, the methods in the relevant technologies need to expand the scope of vehicle information acquisition for prediction and decision calculation. Furthermore, deploying all algorithms on the onboard computer will increase vehicle costs and place a large computational load on the onboard computer. At the same time, the transmission of control commands and the execution of control commands by the actuators require time, resulting in time delays in the vehicle control system, affecting the vehicle control effect, causing vehicle vibration or even loss of control, and seriously affecting driving safety. Summary of the Invention

[0005] This application provides a lane-changing decision-making and delay compensation control method, device, cloud platform, and vehicle based on predictive information to solve problems in related technologies, such as low matching degree between vehicle lane-changing planning and current traffic, poor vehicle control accuracy and stability due to delay links in the vehicle control system, and high planning delay due to insufficient vehicle-side computing power, thereby affecting driving safety.

[0006] The first aspect of this application provides a lane-changing decision-making and delay compensation control method based on predictive information. The method is applied to a cloud platform and includes the following steps: acquiring traffic information collected by a roadside sensing unit; predicting traffic information of vehicles within the sensing range in the future based on the traffic information, and constructing a multi-objective optimization problem of the future driving environment of the target vehicle based on the predicted information; solving the multi-objective optimization problem to obtain the optimal expected speed and lane-changing decision time of the target vehicle in the future; and sending the expected speed and the lane-changing decision time to the target vehicle, wherein the target vehicle responds to the expected speed and the lane-changing decision time using a lane-changing safety judgment algorithm, a lane-changing path planning algorithm, and a delay compensation control algorithm deployed on the vehicle to adjust its speed and change lanes in advance.

[0007] Optionally, the step of predicting the predicted information of traffic vehicles within the perception range for a future period based on the traffic information includes: establishing a traffic vehicle state matrix based on the traffic information; inputting the information of the traffic vehicle state matrix into a micro-car-following model, and using the micro-car-following model to calculate the predicted state transition amount of the traffic vehicles; combining the state transition amount with the discrete state space equation, and iteratively deducing the predicted information of the traffic vehicles at each discrete moment in the future.

[0008] Optionally, the step of constructing a multi-objective optimization problem of the future driving environment of the target vehicle based on the predicted information includes: obtaining the vehicle's own state information and road traffic speed limit information stored on a cloud platform; constructing the multi-objective optimization problem based on the predicted information, the vehicle's own state information, and the road traffic speed limit information, wherein the multi-objective optimization problem includes a driving cost function and constraints, and the cost function reflects the optimization objectives of vehicle driving efficiency and comfort, as well as the requirements of following safety and lane-changing safety.

[0009] Optionally, solving the multi-objective optimization problem to obtain the target vehicle's optimal future expected speed and lane-changing decision time includes: obtaining the target vehicle's own acceleration sequence; decoupling the solution process of the vehicle's own acceleration sequence from the lane-changing decision sequence to discretize the lane subsequences, discretizing possible lane-changing decision sequences in continuous time and space into candidate subsequences; solving the optimal acceleration sequence corresponding to each candidate subsequence from the multi-objective optimization problem, filtering the optimal acceleration sequences, sorting and filtering the driving costs of the filtered candidate subsequences, and taking the subsequence corresponding to the lowest cost as the optimal lane-changing decision sequence, wherein the optimal lane-changing decision sequence includes the expected speed and lane-changing decision time.

[0010] Optionally, the lane-changing safety judgment algorithm includes: after receiving a lane-changing instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, a lane-changing operation is performed; otherwise, a downgraded safety following operation is performed, slowing down and following the vehicle in front; the lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes based on the change in driving state; the straight-line selection uses the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, using the vehicle position receiving the lane-changing decision as the starting point coordinates of the lane-changing, setting a path optimization function based on vehicle comfort and lane-changing efficiency to calculate the optimal lane-changing time, and setting the lane-changing process as a uniform speed lane-changing, deriving... The coordinates of the lane change endpoint are obtained; the lane change path is obtained by fitting the coordinates of the lane change start point and the lane change endpoint using a polynomial curve; the time delay compensation control algorithm includes: building a longitudinal dynamic model of the vehicle, calculating the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and performing longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; building a lateral dynamic model of the vehicle, analyzing the lateral dynamic model using two degrees of freedom, calculating the lateral control law through feedforward control and a linear quadratic regulator, establishing a time delay lateral control model based on the time delay elements in the control system, and using the state augmentation method to transform the time delay system into a time delay-free system to achieve lateral time delay compensation.

[0011] A second aspect of this application provides a lane-changing decision-making and delay compensation control method based on predictive information. The method is applied to a vehicle and includes the following steps: obtaining the vehicle's desired speed and lane-changing decision time from a cloud platform; wherein the cloud platform obtains traffic information collected by a roadside sensing unit, predicts traffic information within the sensing range for a future period based on the traffic information, constructs a multi-objective optimization problem for the vehicle's future driving environment based on the predicted information, solves the multi-objective optimization problem to obtain the vehicle's optimal desired speed and lane-changing decision time; and utilizes a lane-changing safety judgment algorithm, a lane-changing path planning algorithm, and a delay compensation control algorithm deployed in the vehicle to respond to the desired speed and lane-changing decision time, thereby adjusting the speed for lane-changing in advance.

[0012] A third aspect of this application provides a lane-changing decision-making and delay compensation control device based on predictive information. The device is applied to a cloud platform and includes: a first acquisition module for acquiring traffic information collected by a roadside sensing unit; a prediction module for predicting traffic information within the sensing range over a future period based on the traffic information, and constructing a multi-objective optimization problem of the target vehicle's future driving environment based on the predicted information, solving the multi-objective optimization problem to obtain the target vehicle's optimal expected speed and lane-changing decision time; and a distribution module for distributing the expected speed and lane-changing decision time to the target vehicle, wherein the target vehicle responds to the expected speed and lane-changing decision time using its own lane-changing safety judgment algorithm, lane-changing path planning algorithm, and delay compensation control algorithm to adjust its speed and change lanes in advance.

[0013] Optionally, the prediction module is further configured to: establish a traffic vehicle state matrix based on the traffic information; input the information of the traffic vehicle state matrix into a micro-car-following model, and use the micro-car-following model to calculate the predicted state transition amount of the traffic vehicle; combine the state transition amount with the discrete state space equation to iteratively deduce the predicted information of the traffic vehicle at each discrete moment in the future.

[0014] Optionally, the prediction module is further configured to: obtain the vehicle status information of the target vehicle and the road traffic speed limit information stored on the cloud platform; construct the multi-objective optimization problem based on the prediction information, the vehicle status information and the road traffic speed limit information, wherein the multi-objective optimization problem includes a driving cost function and constraints, and the cost function reflects the optimization objectives of vehicle driving efficiency and comfort, and the requirements of car-following safety and lane-changing safety.

[0015] Optionally, the prediction module is further configured to: obtain the acceleration sequence of the target vehicle; decouple the acceleration sequence of the target vehicle from the lane-changing decision sequence solution process to discretize the lane subsequence, and discretize the possible lane-changing decision sequences in continuous time and space into candidate subsequences; solve the optimal acceleration sequence corresponding to each candidate subsequence from the multi-objective optimization problem, and filter the optimal acceleration sequence; sort and filter the driving costs of the filtered candidate subsequences, and take the subsequence corresponding to the lowest cost as the optimal lane-changing decision sequence, wherein the optimal lane-changing decision sequence includes the expected speed and the lane-changing decision time.

[0016] Optionally, the sending module is further configured to: after receiving the lane change instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, then the lane change operation is performed; otherwise, a downgraded safety following operation is performed, and the vehicle slows down to follow the vehicle in front.

[0017] Optionally, the sending module is further configured to: switch between lane-changing and straight-line planning modes based on the change in driving state; select the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, use the vehicle position receiving the lane-changing decision as the starting point coordinates of the lane-changing, set a path optimization function based on vehicle comfort and lane-changing efficiency to calculate the optimal lane-changing time, set the lane-changing process as a uniform speed lane-changing, and derive the lane-changing endpoint coordinates; obtain the lane-changing path by fitting the starting point coordinates and the endpoint coordinates of the lane-changing using a polynomial curve.

[0018] Optionally, the sending module is further configured to: build a longitudinal dynamics model of the vehicle, calculate the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and perform longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; build a lateral dynamics model of the vehicle, analyze the lateral dynamics model using two degrees of freedom, calculate the lateral control law through feedforward control and a linear quadratic regulator, establish a time-delayed lateral control model based on the time delay elements in the control system, and use the state augmentation method to transform the time-delayed system into a time-delay-free system to achieve lateral time delay compensation.

[0019] The fourth aspect of this application provides another lane-changing decision and delay compensation control device based on predictive information. The device is applied to a vehicle and includes: a second acquisition module, configured to acquire the vehicle's desired speed and lane-changing decision time from a cloud platform; wherein the cloud platform acquires traffic information collected by a roadside sensing unit, predicts traffic information within the sensing range for a future period based on the traffic information, and constructs a multi-objective optimization problem for the vehicle's future driving environment based on the predicted information, solving the multi-objective optimization problem to obtain the vehicle's optimal desired speed and lane-changing decision time; and a response module, configured to respond to the desired speed and lane-changing decision time using a lane-changing safety judgment algorithm, a lane-changing path planning algorithm, and a delay compensation control algorithm deployed in the vehicle, so as to adjust the speed in advance for lane-changing.

[0020] The fifth aspect of this application provides a cloud platform including a lane-changing decision and delay compensation control device based on predictive information as described in the above embodiments.

[0021] A sixth aspect of this application provides a vehicle including another lane-changing decision and delay compensation control device based on predictive information as described in the above embodiments.

[0022] Therefore, this application has at least the following beneficial effects:

[0023] This application embodiment can perform complex calculations based on a cloud platform, thus reducing the vehicle's computing resources and improving computing speed. Based on traffic vehicle prediction information within the perception range over a future period, this application embodiment can construct and solve for the target vehicle's optimal expected speed and lane-changing strategy. This allows the vehicle to anticipate the impact of adverse traffic conditions and adjust its speed and lane changes in advance, effectively optimizing driving efficiency and comfort. Furthermore, this application embodiment can utilize lane-changing safety algorithms, lane-changing path planning algorithms, and time-delay compensation control algorithms deployed in the vehicle to respond. This enables the vehicle to dynamically plan multiple lane-changing paths during driving, balancing comfort and efficiency during lane-changing, effectively eliminating the impact of time delay, and allowing the vehicle to accurately track reference control signals.

[0024] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0025] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0026] Figure 1 This is a flowchart of a lane-changing decision and delay compensation control method based on predictive information according to an embodiment of this application;

[0027] Figure 2 This is a schematic diagram illustrating a specific operating scenario of a vehicle traveling on a highway, according to one embodiment of this application.

[0028] Figure 3 This is a schematic diagram illustrating the technical framework and principles of the embodiments of this application;

[0029] Figure 4 This is a schematic diagram of a scenario where vehicles travel on a highway, according to an embodiment of this application.

[0030] Figure 5 This is a schematic diagram illustrating the calculation of the safe distance between the vehicle and the vehicle in front, according to an embodiment of this application.

[0031] Figure 6 This is a schematic diagram illustrating the calculation of the safe distance between the vehicle and a vehicle in the target lane, according to an embodiment of this application.

[0032] Figure 7 This is a schematic diagram of the path planning and switching strategy in an embodiment of this application;

[0033] Figure 8 This is a schematic diagram of the longitudinal control algorithm according to an embodiment of this application;

[0034] Figure 9 This is a control block diagram of the longitudinal control system according to an embodiment of this application;

[0035] Figure 10 This is a control block diagram for introducing a Smith predictor in an embodiment of this application;

[0036] Figure 11 This is a schematic diagram of the lateral control algorithm according to an embodiment of this application;

[0037] Figure 12 A flowchart of another lane-changing decision and delay compensation control method based on predictive information according to an embodiment of this application;

[0038] Figure 13 This is an example diagram of a lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application;

[0039] Figure 14 This is an example diagram of another lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application. Detailed Implementation

[0040] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0041] Lane changing is a fundamental driving maneuver, requiring consideration of lateral movement and the status of traffic in the target lane. Its complexity increases the likelihood of conflicts with other vehicles during lane changing, significantly impacting the traffic environment. Statistics show that approximately 15% to 20% of traffic accidents occur during lane changes, and lane-changing-related traffic congestion accounts for about 25% of all accidents. Various techniques have been proposed to address lane changing issues; for example:

[0042] Among related technologies, a "cloud control system" architecture can be used. This is a complex cyber-physical system that integrates the physical, information, and application layers of people, vehicles, roads, and the cloud using next-generation information and communication technologies. It achieves fusion perception, decision-making, and control, resulting in a comprehensive improvement in vehicle driving and traffic operation safety and efficiency. It is considered a new type of vehicle-road-cloud integrated collaborative system that addresses current development issues in the automotive and transportation industries and adapts to Chinese standards. Based on this research on intelligent connected vehicle applications using a cloud control system, a highway economic cruise algorithm based on the cloud control system can be designed. This algorithm obtains the road gradient from map data in the cloud control platform and then calculates the vehicle's economical driving speed when the terrain undulates. In a specific embodiment, this method achieved an average fuel saving rate of 4.68% in actual road tests.

[0043] Among related technologies, a cloud-based predictive cruise system for buses can also be used. This system can acquire traffic light phases and perceive traffic flow density and speed ahead based on traffic data in the cloud platform. It plans the vehicle's speed according to the traffic light phase information and the predicted queuing time at intersections, thereby reducing fuel consumption and red light stopping time. Compared with manually driven vehicles, the algorithm saves 44.94% to 56.74% of energy consumption and reduces stopping time by 26.8 seconds in a 250-second driving time, effectively optimizing vehicle driving performance.

[0044] Among related technologies, an autonomous driving application based on road traffic prediction information under a cloud control system can focus on cruise control technology. This technology can optimize vehicle speed based on traffic information such as road gradient and traffic light signal switching patterns, aiming to improve vehicle driving economy and reduce stopping time at traffic light intersections. However, limiting a vehicle's driving range to only one lane severely restricts the optimization space for vehicle driving. Furthermore, considering only static or regularly changing traffic information is insufficient to adapt to complex traffic situations. It is necessary to consider dynamic traffic prediction information to control the vehicle's lane-changing behavior, which can provide more possibilities for optimizing vehicle driving strategies, thereby further improving the vehicle's overall performance.

[0045] In related technologies, a lane-changing decision model can also be used. This model can simulate and model driver behavior to provide vehicles with lane-changing logic and decision-making processes when driving on urban roads and encountering obstacles ahead. This lane-changing decision model can propose three stages in the lane-changing process: the generation of lane-changing intention, the assessment of lane-changing feasibility, and the implementation of lane-changing behavior.

[0046] The research based on the above lane-changing decision model can be specifically as follows: Considering the impact of lane changing on surrounding traffic, a MOBIL (minimizing overall braking induced by lane change) lane-changing model can be further used to minimize the braking of surrounding traffic caused by lane changing. Furthermore, by analyzing lane-changing behavior, two factors influencing lane changing—speed difference coefficient and vehicle distance expectation—can be introduced, and lane-changing decisions can be modeled based on fuzzy logic theory. Further, the decision behavior database can be preprocessed using fuzzy logic before being input into a neural network for model training, and a gradient boosting tree algorithm is introduced to further optimize decision accuracy. In one embodiment of this method, the lane-changing decision accuracy of intelligent vehicles can reach 99%. Furthermore, the action selection and evaluation of lane changing can be decoupled, and a deep learning Q-network can be used for training to obtain a stable and accurate lane-changing decision model. In one embodiment of this method, it can improve the driving style of vehicles in high-traffic and environmental conditions. Even in more aggressive test scenarios, the success rate of lane-changing decisions remains above 92%. Furthermore, for highway merging areas, a competitive network decision-making model based on deep reinforcement learning and Markov decision models can be established to consider the impact of lane-changing behavior on other vehicles, improving the success rate of lane-changing decisions in different scenarios. Furthermore, the interaction behavior between vehicles can significantly influence lane-changing decisions, and game theory-based lane-changing decision-making methods can fully consider this condition. Further, research can be conducted on free lane-changing scenarios, considering the game strategies between vehicles and introducing a risk coefficient, applying game theory to model lane-changing behavior, improving lane utilization and driving safety. Finally, lane-changing safety distance and lane-changing duration can be incorporated into the game payoff of lane-changing decisions, improving the safety and comfort of lane-changing.

[0047] In related technologies, research on lane-changing decisions based on predictive information can specifically involve: predicting the trajectory of the vehicle in front based on a vehicle kinematics model and incorporating driving intentions, and planning the vehicle's movement in advance based on the predicted different driving behaviors of the vehicle in front, thereby effectively avoiding vehicle collisions; furthermore, a predictive overtaking strategy based on deep learning and reinforcement learning can be proposed, utilizing the NGSIM (Next Generation Simulation) dataset and employing LSTM (Long Short-Term Memory)... A long short-term memory (LSTM) neural network is used to predict the motion states of four recent vehicles in the environment. Based on the prediction results of the motion states of the vehicles in the environment, an improved Q-learning algorithm is established to set a reward for driving at the maximum speed. This algorithm can quickly converge and overtake the vehicles in the environment without collision. Furthermore, an improved Res-LSTM vehicle trajectory prediction model can be designed to predict the future trajectories of surrounding traffic vehicles. A dynamic drivable probability map of vehicles is established by combining current traffic information and traffic prediction information as the basis for vehicle lane-changing decisions. Furthermore, the trajectories of surrounding vehicles can also be predicted based on the LSTM network. An interactive game vehicle behavior prediction model is proposed by combining game theory. Finally, a stochastic game lane-changing decision model based on Nash Q-learning is designed to solve for the optimal sequence of vehicle actions.

[0048] In vehicle control technology, related methods can be used to improve the real-time performance of lane-change trajectory tracking by employing pre-aiming PID (proportional-integration-differentiation) control under low-speed, high-adhesion conditions. Furthermore, a dynamic sliding mode controller can be used for trajectory tracking, designing a sliding mode controller with vehicle speed and heading angle as outputs, and verifying the tracking effect under various conditions through simulation. Additionally, an optimal pre-aiming trajectory tracking control method can be proposed, designing an LQR (Linear Quadratic Regulator) to obtain the optimal steering wheel angle input, thereby achieving intelligent vehicle lane-change trajectory tracking.

[0049] In summary, lane-changing decision-making methods in related technologies can determine whether a lane change is necessary based on the current traffic environment faced by the vehicle. However, when these methods generate lane-changing decisions, the vehicle has often already lost some driving performance or is already in a less than ideal driving environment. When making lane-changing decisions, it is necessary to combine traffic vehicle prediction information, consider future traffic conditions, determine the best lane-changing time in advance, and optimize the vehicle's future driving performance by appropriately adjusting speed and changing lanes in a timely manner to avoid adverse driving environments.

[0050] However, using traffic vehicle prediction information to make lane-changing decisions requires acquiring a large range of vehicle information for prediction and decision-making calculations. The vehicle-mounted perception technology in related technologies is difficult to perceive the status of vehicles over a wider range. Furthermore, deploying all algorithms on the vehicle-mounted computer would not only significantly increase vehicle costs but also place a heavy computational load on the vehicle-mounted computer. The single-vehicle intelligence technology route is difficult to meet the requirements.

[0051] In addition, at the execution level of vehicle lane changing, the transmission of control commands and the execution of control commands by actuators require time, resulting in a general time delay in vehicle control systems. This time delay affects the vehicle control effect, causing vehicle vibration or even loss of control, which seriously affects driving safety.

[0052] To address the problems mentioned in the background section, this application provides a lane-changing decision-making and delay compensation control method based on predictive information. The following description, with reference to the accompanying drawings, outlines the lane-changing decision-making and delay compensation control method, apparatus, cloud platform, and vehicle based on predictive information of this application.

[0053] Specifically, Figure 1 This is a flowchart illustrating a lane-changing decision and delay compensation control method based on predictive information, provided in an embodiment of this application.

[0054] like Figure 1 As shown, this lane-changing decision-making and delay compensation control method based on predictive information is applied to a cloud platform. The method includes the following steps:

[0055] In step S101, traffic information collected by the roadside sensing unit is acquired.

[0056] Traffic information can include vehicle information, and the traffic information collected by the roadside sensing unit can be received by the cloud platform.

[0057] It is understandable that, such as Figure 2 As shown, in this embodiment of the application, traffic information collected by the roadside sensing unit can be received first on the cloud platform, so as to facilitate the prediction of the movement of traffic vehicles in the sensing range in subsequent embodiments.

[0058] In step S102, traffic information is used to predict the traffic vehicles within the perception range in the future. Based on the traffic information, a multi-objective optimization problem of the future driving environment of the target vehicle is constructed. The multi-objective optimization problem is solved to obtain the target vehicle's optimal expected speed and lane-changing decision time in the future.

[0059] In this application embodiment, at least one method can be used to predict traffic vehicles within the perception range for a future period of time, such as using a predictor; the perception range can be determined by the different needs of different vehicles, and no further limitations are imposed on it.

[0060] It is understood that, after obtaining road traffic information, the embodiments of this application can predict the predicted information of vehicles within the perception range in the future, then construct and solve a multi-objective optimization problem of the target vehicle's future driving environment, to obtain the target vehicle's optimal expected speed and lane-changing decision time in the future. Figure 3 As shown, embodiments of this application can establish the traffic vehicle state matrix, the micro-car-following model, and the discrete state-space equations, as detailed below:

[0061] In this embodiment of the application, predicting traffic information of vehicles within the perception range for a future period based on traffic information includes: establishing a traffic vehicle state matrix based on traffic information; inputting the information of the traffic vehicle state matrix into a micro-car-following model and using the micro-car-following model to calculate the predicted state transition amount of the traffic vehicles; combining the state transition amount with the discrete state space equation and iteratively deducing the predicted information of the traffic vehicles at each discrete moment in the future.

[0062] It is understandable that, such as Figure 3 As shown, in this embodiment of the application, a traffic vehicle state matrix can be established first, the predicted state transition of traffic vehicles can be calculated using a micro-car-following model, and then the state of traffic measurement at each discrete moment can be iteratively derived using discrete state-space equations. The specific process of deriving the predicted information of traffic vehicles at each discrete moment in this embodiment of the application can be as follows:

[0063] I. Establish a traffic vehicle state matrix:

[0064] The scenario of high-speed vehicle travel in the embodiments of this application can be as follows: Figure 4 As shown, besides the vehicle SV, there are n and m vehicles in the two lanes of the two-lane highway respectively; the vehicles in the left lane are numbered li from front to back, i∈(1,2,3…n), and the vehicles in the right lane are numbered ri from front to back, i∈(1,2,3…m).

[0065] Using set Z id (k) represents the state of each vehicle at time k, which can be expressed as:

[0066] Z id (k)={x id (k), v id (k), lane id (k)}, id∈{SV, r.1, r.2...rm, l.1, l.2...ln}#(1)

[0067] Where, x id v represents the longitudinal position of the vehicle at time k. idLane represents the vehicle speed at time k. id The lane number represents the current lane number of the vehicle at time k. id ={0, 1}, when the vehicle is in the left lane id =1, lane when in the right lane id =0.

[0068] Using set U id (k) represents the control variable that causes the state change of each vehicle at time k, U id (k) can be represented as:

[0069]

[0070] Among them, a id (k) represents the vehicle's acceleration at time k. This represents the vehicle's decision to change lanes to the left at time k; 0 indicates no lane change, 1 indicates left lane change; This represents the right lane-changing decision of the vehicle at time k, and its value is determined in the same way as the left lane-changing decision.

[0071] II. Establishing a micro-level car-following micro-model:

[0072] In highway driving scenarios, vehicle driving is relatively stable and lane changes are infrequent. Therefore, in traffic vehicle state prediction methods, it is assumed that vehicles will not change lanes during the prediction time domain. Thus, the lane change control variable for traffic vehicles is:

[0073]

[0074] Calculate the expected acceleration a of vehicles using a car-following model id (k) uses IDM (Intelligent Driver Model) to calculate the expected acceleration; the IDM model expression can be:

[0075]

[0076] Among them, a max v represents the maximum acceleration of the vehicle. e v represents the vehicle's desired speed. id (k) represents the velocity of any vehicle at time k, Δv id (k) represents the longitudinal relative velocity between any vehicle and the vehicle in front at time k, δ a The acceleration exponent is usually taken as δ. a =4, Δx id (k) represents the actual following distance of any vehicle at time k, s * (v id(k), Δv id (k) represents the expected following distance of any vehicle at time k, s0 represents the minimum vehicle spacing when stationary, THW r 'b' represents the desired headway, and 'b' represents the vehicle's set comfort deceleration.

[0077] In the prediction model, it is assumed that the first vehicle in the perception domain is traveling at a constant speed, and that the vehicle's driving behavior on the highway is usually relatively stable. Therefore, it can be assumed that the vehicle maintains a constant acceleration during each prediction time step.

[0078] III. Establishing Discrete State-Space Equations:

[0079] Based on the micro-car-following model, the current expected acceleration of a vehicle can be calculated. Using this expected acceleration, the predicted state of the vehicle at the next moment can be calculated. The state transition formula is as follows:

[0080] v id (k+1)=v id (k)+a id (k)ΔT#(5)

[0081]

[0082]

[0083] Where ΔT is the discrete time step.

[0084] Based on equations 5 to 7, embodiments of this application can transform the state transition equation into a discrete state-space equation, thereby enabling iterative calculation of the future state of the vehicle:

[0085] Z id (k+1)=AZ id (k)+BU id (k)#(8)

[0086] Where A and B are the coefficient matrices of the state-space equations, and the coefficient matrix form is as follows:

[0087]

[0088] In summary, in practical applications, since the error of the prediction result will increase with the growth of the prediction time domain, the embodiments of this application can use a rolling update prediction method to eliminate the prediction state error. By continuously updating the current state of the traffic vehicles received from the cloud, the prediction model can continuously update the predicted state of the traffic vehicles.

[0089] In this embodiment of the application, a multi-objective optimization problem for the future driving environment of the target vehicle is constructed based on the predicted information, including: obtaining the vehicle's own state information and road traffic speed limit information stored on the cloud platform; constructing a multi-objective optimization problem based on the predicted information, the vehicle's own state information, and the road traffic speed limit information, wherein the multi-objective optimization problem includes a driving cost function and constraints, and the cost function reflects the optimization objectives of vehicle driving efficiency and comfort, the requirements of car-following safety, and lane-changing safety.

[0090] It is understood that, in the embodiments of this application, after obtaining the vehicle status information of the target vehicle and the road traffic speed limit information stored on the cloud platform, a multi-objective optimization problem can be constructed based on the obtained information.

[0091] Specifically, this embodiment first utilizes predicted information, vehicle state information, and road traffic speed limit information stored on a cloud platform to construct a driving cost function and constraints. Then, the acceleration sequence of the vehicle is decoupled from the lane change decision sequence solution process, and lane subsequence discretization is performed, dividing the infinitely many possible lane change decision sequences in continuous time and space into a finite number of candidate subsequences. Next, this embodiment performs subsequence optimization to obtain the optimal acceleration sequence corresponding to each candidate subsequence. Finally, optimal sequence selection is performed, ranking and filtering the driving values ​​of the optimized candidate subsequences, and selecting the subsequence with the lowest cost as the optimal lane change decision sequence. The cost function reflects the optimization objectives of vehicle driving efficiency and comfort, as well as the requirements of car-following safety and lane-changing safety. The process of constructing a multi-objective optimization problem in this embodiment can be specifically described as follows:

[0092] The optimization objectives of the lane-changing decision algorithm based on traffic vehicle state prediction in this application embodiment mainly include driving efficiency, driving comfort, and the vehicle's following safety and lane-changing safety. Based on the driving optimization objectives and safety requirements, this application embodiment can design a vehicle driving cost function, wherein the cost function is expressed as:

[0093]

[0094] Among them, w v w a w f w 1c w1 and w1 represent the weighting coefficients for driving efficiency cost, driving comfort cost, car-following safety cost, lane-changing safety cost, and deviation from the desired lane cost, respectively. i v represents the vehicle's speed at the i-th predicted time. e v represents the desired speed of the vehicle. ifv Let a represent the speed of the vehicle in front of the current vehicle at the i-th predicted time. iLet f represent the acceleration of the vehicle at the i-th predicted time. i lc represents the car-following safety cost of this vehicle at the i-th predicted time. i p represents the lane-changing safety cost of this vehicle at the i-th predicted time. i This represents the cost of the vehicle deviating from the desired lane at the i-th prediction time.

[0095] Following the safety cost f i The calculation uses THW (Time Headway) and TTC (Time-To-Collision) as indicators. The calculation methods for these two indicators are as follows:

[0096]

[0097] Where x, x fv These represent the longitudinal positions of this vehicle and the vehicle in front, v and v', respectively. fv These represent the longitudinal speeds of the vehicle and the vehicle in front, respectively.

[0098] When following another vehicle in the current lane, two scenarios are primarily considered: the vehicle in front is traveling at a lower speed than this vehicle, and the vehicle in front is traveling at a higher speed than this vehicle. When the vehicle in front is traveling at a lower speed, not only is a safe following distance necessary, but the relative speed between vehicles must also be considered. Therefore, when calculating the car-following safety cost in this scenario, both Head-Wearing Distance (THW) and Time-To-Traffic (TTC) will be considered simultaneously. When the vehicle in front is traveling at a higher speed, there is no tendency for this vehicle to approach the vehicle in front, so the TTC indicator is not considered; only the Head-Wearing Distance (THW) indicator is considered to maintain a safe following distance. Therefore, the formula for calculating the car-following safety cost f is:

[0099]

[0100] Where a and b represent the weights in the calculation of the safety cost of following the car, respectively. r Indicates the expected headway, TTC r Indicates the expected collision time.

[0101] When there are no vehicles ahead, the vehicle is in a free-moving state. The safety cost of following another vehicle in this situation is f. i The value is 0; the formula for calculating the safety cost of lane changing for this vehicle can be:

[0102]

[0103] Among them, D real D e These represent the actual distance and the expected distance between the vehicle and the vehicle behind it in the target lane, respectively.

[0104] It should be noted that, to avoid frequent lane changes, even when there are no following vehicles in the target lane, a certain lane-changing safety cost will still be incurred. In this case, the embodiment of this application can set lc = 0.03. The lane-changing safety cost of this vehicle is only calculated in the prediction step where a lane change occurs; the lane-changing cost is 0 in the prediction step where no lane change occurs. p represents the vehicle's expected lane deviation cost. For example, the calculation formula when the expected lane is the right lane can be:

[0105]

[0106] Considering road speed limits and the vehicle's own dynamic performance, the vehicle's speed should be limited:

[0107]

[0108] in, These represent the road speed limit and the maximum speed a vehicle can reach, respectively.

[0109] The vehicle's acceleration is also limited by road conditions and the vehicle's powertrain, and must meet acceleration constraints:

[0110]

[0111] in, These represent the vehicle's maximum braking deceleration and maximum acceleration, respectively.

[0112] To prevent excessive changes in acceleration during vehicle operation, acceleration constraints need to be set.

[0113] jerk min ≤jerk(k)≤jerk max #(17)

[0114] Among them, jerk min jerk max Let represent the minimum and maximum allowable rates of change of acceleration at time k, respectively. The lane change command generated by this vehicle can only be a lane change or no lane change at each time point; a lane change to both the left and right is not allowed. Therefore, the lane change command constraint can be:

[0115]

[0116] In this embodiment of the application, solving a multi-objective optimization problem to obtain the target vehicle's optimal expected speed and lane-changing decision time includes: obtaining the target vehicle's own acceleration sequence; decoupling the solution process of the vehicle's own acceleration sequence from the lane-changing decision sequence to discretize the lane subsequence, discretizing the possible lane-changing decision sequences in continuous time and space into candidate subsequences; solving the optimal acceleration sequence corresponding to each candidate subsequence from the multi-objective optimization problem, filtering the optimal acceleration sequences, sorting and filtering the driving costs of the filtered candidate subsequences, and taking the subsequence corresponding to the lowest cost as the optimal lane-changing decision sequence, wherein the optimal lane-changing decision sequence includes the expected speed and lane-changing decision time.

[0117] It is understood that the process of solving the multi-objective optimization problem to obtain the desired speed and lane-changing decision time in the embodiments of this application may include: lane subsequence discretization, subsequence optimization solution, and optimal sequence selection. Specifically:

[0118] I. Discretization of Lane Subsequences:

[0119] The driving cost function in this embodiment can solve for the corresponding optimal driving speed control sequence after determining the lane-changing sequence of the vehicle. However, there are countless possibilities for the lane-changing sequence of the vehicle in continuous time, which makes the optimization problem unsolvable. Therefore, in order to simplify the computational complexity, it is necessary to discretize the time of lane-changing decision generation to be consistent with the prediction step size, that is, to solve whether to change lanes at each future prediction time. At this time, the output speed control sequence and lane-changing decision sequence of the predictive decision algorithm can be expressed as:

[0120]

[0121] Where, N p a represents the total number of prediction steps included in the prediction time domain. seq θ represents the vehicle's acceleration sequence in the prediction time domain. seq This represents the lane-changing sequence of vehicles in the prediction time domain; therefore, the lane-changing optimization problem can be divided into N parts based on the possible times of lane changes. p +1 definite lane-changing sequence subproblems, corresponding to not changing lanes (e.g., changing lanes in the 1st prediction step, changing lanes in the 2nd prediction step, and so on in the Nth prediction step). p (lane change at each prediction step), the set of lane change sequences can be represented as:

[0122]

[0123] Where, θ list Let LK represent the set of all lane-changing sequences, LC represent lane changes, and θ0 represent sequences without lane changes. i , i∈{1,2,3…Np} represents the lane-changing sequence where different lane-changing decisions occur at different times.

[0124] II. Subsequence Optimization Solution:

[0125] The embodiments of this application can discretize N. p By substituting +1 lane-changing subsequences into the above cost function, we can obtain the optimal acceleration sequence for each given lane-changing sequence.

[0126]

[0127] III. Optimal Sequence Selection:

[0128] This application embodiment can select the subsequence with the minimum cost from all optimized lane-changing subsequences. The corresponding lane-changing subsequence and the optimal driving acceleration sequence under that sequence are the optimization solution results. The first control quantity is extracted from the acceleration sequence and lane-changing decision sequence and substituted into the state-space equation. According to Equation 8, the expected state of the vehicle at the next moment can be calculated. The expected state is output to the vehicle to control the vehicle. At each moment, the cloud will also continuously calculate the latest expected state based on the latest state uploaded by the vehicle and the update of prediction information and continuously send it to the vehicle.

[0129]

[0130] J e ={a e θ e}#(twenty three)

[0131] U SV ={a e (1), θ e (1)}#(24)

[0132] Among them, J e Let a represent the minimum cost among all lane-changing sequences. e θ e These represent the speed control sequence and lane-changing decision sequence corresponding to the minimum cost, respectively.

[0133] In step S103, the desired speed and lane-changing decision time are sent to the target vehicle. The target vehicle uses its own lane-changing safety judgment algorithm, lane-changing path planning algorithm and time delay compensation control algorithm to respond to the desired speed and lane-changing decision time, so as to adjust its speed in advance to change lanes.

[0134] It is understood that, in this embodiment of the application, the desired speed and lane-changing decision time can be sent to the target vehicle. After receiving the information, the target vehicle responds using the lane-changing safety judgment algorithm, the lane-changing path planning algorithm, and the delay compensation control algorithm, and performs corresponding control actions to adjust the vehicle's speed in advance for lane-changing. The response process of the lane-changing safety judgment algorithm, the lane-changing path planning algorithm, and the delay compensation control algorithm in this embodiment of the application can be specifically described as follows:

[0135] I. Lane Changing Safety Judgment Algorithm:

[0136] In this embodiment of the application, the lane change safety judgment algorithm includes: after receiving the lane change instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and the surrounding vehicles is greater than the safety distance, the lane change operation is performed; otherwise, the vehicle will perform a downgraded safety following operation and slow down to follow the vehicle in front.

[0137] It is understandable that, such as Figure 5 As shown, when the lower front corner of this vehicle is aligned with the upper rear corner of the vehicle in front, a minimum safe distance S0 must still be maintained between the two vehicles. Therefore, after receiving a lane-changing instruction, this embodiment of the application can compare the actual distance between this vehicle and the surrounding vehicles with the safe distance. When the actual distance is greater than the safe distance, i.e., when lane changing is safe, the lane-changing operation is performed; otherwise, the lane-changing is not performed, and a downgraded safe following operation is performed. Figure 5 The width of the vehicle is W, and the width of the vehicle is y. tc θ represents the lateral displacement of the vehicle at the critical collision point, and θ is the heading angle of the vehicle at the critical collision point.

[0138] After this car begins to change lanes, the two cars pass t c1 Later by Figure 5 The relative position on the left side reaches Figure 5 The longitudinal distance between the vehicles on the right is the minimum safe distance S0. The safe distance S between this vehicle and the vehicle in front is... safe1 The calculation formula is:

[0139] S SV -S FV ≤S safe1 -S0#(25)

[0140] Among them, S SV For this vehicle in t c1 The longitudinal displacement S during travel FV For the car in front at t c1 The longitudinal displacement during lane changing. Assuming all vehicles travel at a constant speed during the lane change, the above formula can be derived as:

[0141] vt c1 -v FV t c1 <S safe1 -S0#(26)

[0142] Among them, v, v FV These are the speeds of this vehicle and the vehicle in front, respectively.

[0143] Therefore, the safe distance from the vehicle in front is determined by the critical collision time t. c1 It is determined that the lane-changing path of this vehicle is represented by a fifth-order polynomial, and the vehicle travels at a constant speed during the lane-changing process. Therefore, the critical collision time t of this vehicle is... c1 The following relationship can be derived from the lateral displacement of the critical collision:

[0144] y tc =y(t) c )=b5*t c 5 +b4*t c 4 +b3*t c 3 +b2*t c 2 +b1*t c +b0#(27)

[0145] Where b0, b1…b5 are polynomial coefficients. When the vehicle reaches the ultimate collision position, the lateral displacement is:

[0146]

[0147] Therefore, the critical collision time t can be deduced from the above polynomial relationship in the embodiments of this application. c1 Further determine the safe distance.

[0148] It should be noted that, for vehicles in the target lane, this embodiment of the application requires ensuring that after the vehicle has fully entered the target lane and is driving smoothly, it maintains a certain distance from the vehicles in front and behind in the target lane. A schematic diagram for calculating the safe distance is shown below. Figure 6 As shown; where S is the safe distance S between this vehicle and the vehicle in front in the target lane. safe2 and the safe distance S from the vehicle behind in the target lane safe3 can be Figure 6 (a) and Figure 6 (b) The positional relationship is calculated as follows:

[0149] S SV -S TFV ≤S safe2 -S0#(29)

[0150] S TRV -S sV ≤S safe3 -S0#(30)

[0151] Figure 6This indicates a critical collision state between this vehicle and a vehicle in the target lane. At this point, the lateral displacement of this vehicle is:

[0152] y tc =W road #(31)

[0153] Among them, W road This represents the lane width. The corresponding critical collision time t can be calculated using Equation 27. c2 and t c3 Therefore, equations 29 and 30 can be written as:

[0154] vt c2 -v TFV t c2 ≤S safe2 -S0#(32)

[0155] v TRV t c3 -vt c3 ≤S safe3 -S0#(33)

[0156] Among them, v TFV v is the speed of the vehicle ahead in the target lane. TRV The speed of the vehicle behind in the target lane.

[0157] In summary, the safe distance model can be described as follows:

[0158]

[0159] It is understood that when the target vehicle in this application embodiment changes lanes, if all three safety distances of the vehicle are met, the lane change safety judgment is passed; if the safety judgment conditions are not met, the safe following strategy will overwrite the lane change decision control quantity with no lane change, and use the IDM model to calculate the expected acceleration based on the relative speed and relative position of the vehicle and the vehicle in front, and perform safe following control until the cloud recalculates the optimal lane change decision based on the updated traffic conditions and the safety judgment is passed, and then executes the instructions issued by the cloud.

[0160] II. Lane-changing path planning algorithm:

[0161] In this embodiment, the lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes based on the change in driving state; selecting the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, using the vehicle position receiving the lane-changing decision as the starting point coordinates of the lane change, setting a path optimization function based on vehicle comfort and lane-changing efficiency to calculate the optimal lane-changing time, setting the lane-changing process as a uniform speed lane change, and deriving the lane-changing endpoint coordinates; and obtaining the lane-changing path by fitting the starting point coordinates and the endpoint coordinates of the lane change using a polynomial curve.

[0162] It is understood that the embodiments of this application may first design a path planning and switching strategy, such as Figure 7 As shown, when the vehicle does not receive a lane-change decision, it maintains a straight-ahead state and plans a straight-ahead path. When a lane-change decision is received, the driving state switches to lane-change state, and the planning of a lane-change path begins. When the vehicle reaches the end of the lane-change path, the driving state switches back to straight-ahead state, and a straight-ahead path is planned. The path planning includes straight-ahead path planning and lane-change path planning. When planning the straight-ahead path, the center line of the road closest to the vehicle's center coordinates is selected as the reference path for straight-ahead driving in the current lane. When planning the lane-change path, a polynomial can be used to plan the vehicle's lane-change path.

[0163] Specifically, in this embodiment, the vehicle can be set to travel at a constant speed when changing lanes. When planning the lane-changing path, a fifth-order polynomial lane-changing trajectory can be planned first based on the time and space relationships. Then, by eliminating the time term, the lane-changing path function can be obtained. The fifth-order polynomial is used to describe the longitudinal and lateral trajectories respectively. In the fifth-order polynomial, six parameters need to be determined. The fifth-order polynomial function of the vehicle's lane-changing trajectory can be expressed as:

[0164]

[0165] Where t represents time, and a0, a1, a2, a3, a4, a5 and b0, b1, b2, b3, b4, b5 represent the coefficients to be determined for the lane-changing trajectory function. The initial state of this vehicle's lane change is... This vehicle has completed its lane change. According to the requirements of the vehicle's lane-changing trajectory, the vehicle must be centered on the lane line and have a heading angle of 0 at both the initial and final states of the lane change. The lateral acceleration at the initial and final moments of the lane change must also be 0. Therefore, the relationship between the initial and final states of the vehicle's lane change can be derived as follows:

[0166]

[0167]

[0168] x f =v·t f #(38)

[0169] y f =W road #(39)

[0170] Where t0 represents the initial time of the lane change, t f W represents the time when the lane change ends. road Indicates the lane width.

[0171] Substituting equation 37 into equation 35, and taking the first and second derivatives of equation 35 respectively, we get:

[0172]

[0173] By combining equations 35, 36, and 40, the expression for the lane-changing trajectory can be derived as follows:

[0174]

[0175] Where t represents time. Solving the above two equations simultaneously to eliminate the time term yields the actual lane-changing path expression:

[0176]

[0177] Where, x f This indicates the longitudinal distance traveled by the vehicle when the lane change ends.

[0178] Select the maximum lateral acceleration a during the vehicle lane change process. ymax With maximum lateral jerk ymax As an indicator of lane change comfort, lane change time t f The lane-changing efficiency of this vehicle is measured, and the optimal lane-changing time t is ultimately determined. f Determine the optimal lane-changing path. Normalize each variable to obtain the optimization function, which can be expressed by Equation 43:

[0179]

[0180] Among them, a yres This indicates the maximum permissible lateral acceleration constraint for this vehicle, jerk. yres This indicates the maximum permissible lateral acceleration constraint for this vehicle, t. fmin t fmax ω1 and ω2 represent the lower and upper limits of the lane-changing duration of this vehicle, respectively. ω1 and ω2 represent the weight coefficients of each indicator, and the weight coefficients satisfy the relationship ω1 + ω2 = 1.

[0181] For the lane-changing path optimization function, this embodiment of the application can use GA (Genetic Algorithm) for optimization, with the population size set to 100 and the maximum number of generations to 1000. A floating-point encoding method is used to encode the initially randomly generated chromosome, i.e., generating a random floating-point number between [0, 1], and then mapping this number to the value range of the variable to obtain an initial feasible solution. The encoding method for chromosome T is as follows:

[0182] f=rand(0,1)×(max(t)-min(t))+min(t)#(44)

[0183] Where max(t) and min(t) represent the upper and lower limits of the variable, respectively.

[0184] After generating the initial population of individuals, the fitness of each initial chromosome can be calculated in this embodiment of the application. In the optimization problem of this embodiment of the application, the path switching optimization function is used as the fitness calculation function, and the value of the optimization function is used as the fitness.

[0185] min fitness = J(a y jerk y , t f )#(45)

[0186] In this embodiment, the roulette wheel operator can be selected as the selector. The probability of each chromosome being selected for crossover and mutation to form a new generation of chromosomes during iteration is determined by its fitness. If the optimization function needs to be minimized, the smaller the fitness value, the higher the probability of selection. Therefore, the roulette wheel operator selection probability calculation formula can be designed as follows:

[0187]

[0188] Where N represents the number of individuals in the population, f i This represents the fitness of the i-th chromosome.

[0189] III. Time Delay Compensation Control Algorithm:

[0190] In this embodiment, the time delay compensation control algorithm includes: building a longitudinal dynamics model of the vehicle, calculating the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and performing longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; building a lateral dynamics model of the vehicle, analyzing the lateral dynamics model using two degrees of freedom, calculating the lateral control law through feedforward control and a linear quadratic regulator, establishing a time delay lateral control model based on the time delay elements in the control system, and using the state augmentation method to transform the time delay system into a time delay-free system to achieve lateral time delay compensation.

[0191] It is understood that the embodiments of this application can use the vehicle dynamics model to calculate the longitudinal time delay compensation value and the lateral time delay compensation value of the vehicle. The solution process for the longitudinal time delay compensation value and the lateral time delay compensation value can be specifically as follows:

[0192] (1) Longitudinal control algorithm:

[0193] The schematic diagram of the longitudinal control algorithm in this application embodiment can be seen as follows: Figure 8As shown, speed control is achieved by inputting the vehicle's desired speed and actual speed, and outputting the vehicle's desired acceleration. In this embodiment, a longitudinal dynamics model can be established first. The longitudinal control algorithm aims to enable the vehicle to track the desired speed. The longitudinal dynamics model of the vehicle is established considering the communication delay and actuator hysteresis of the longitudinal control system.

[0194]

[0195] Where v(t) represents the vehicle speed, a(t) represents the actual vehicle acceleration, f represents the time constant of the actuator's inertial hysteresis, and a e (t) represents the desired acceleration output by the control algorithm, t d This indicates the communication delay within the control system.

[0196] Transform the longitudinal delay model into the form of a transfer function:

[0197]

[0198] Where s represents the Laplace variable, and v(s) and a e (s) are v(t) and a, respectively. e Laplace transform of (t).

[0199] For the longitudinal control law, the embodiments of this application can use a PID control algorithm to calculate the desired acceleration a. e The control law of the PID controller can be expressed as: (t),

[0200]

[0201] Among them, K p T represents the proportional coefficient. i T represents the integration time constant. d Let e(t) represent the differential time constant, and let e(t) represent the error between the state variable and the reference variable. This can be expressed as e(t) = v e (t)-v r (t), v e (t) represents the desired velocity, v r (t) represents the actual speed.

[0202] Regarding longitudinal control delay compensation, in the process of establishing the aforementioned longitudinal dynamics model, the embodiments of this application have already considered the communication delay when the control quantity is transmitted to the actuator and the hysteresis when the control quantity is executed. Therefore, the control system containing delay and hysteresis can be analyzed first, and its control block diagram can be as follows: Figure 9 As shown. The closed-loop transfer function of this embodiment is:

[0203]

[0204] Among them, v r (s) and v e (s) represents the Laplace transform between the actual velocity and the desired velocity, G c (s) represents the transfer function of the PID control algorithm, G(s)e -tds The transfer function represents the longitudinal model transfer function for time delay; the transfer function of the PID controller is:

[0205]

[0206] Introducing a Smith estimator into the system reduces the impact of latency. The control block diagram for introducing the Smith estimator can be shown as follows: Figure 10 As shown; assuming the established longitudinal dynamic model is accurate, the system transfer function is:

[0207]

[0208] Understandably, compared to a system with no delay, the transfer function is simply multiplied by e. -tds The closed-loop circuit does not contain any lag terms, and the lag elements do not affect the characteristic equation of the system, thus it can compensate for time delays.

[0209] (2) Lateral control method:

[0210] like Figure 11 As shown, in this embodiment, the planned reference path's lateral and longitudinal coordinates, reference heading angle, and path curvature information can be input. Combined with the vehicle's lateral position, speed, heading angle, heading angular velocity, and actual front wheel steering angle, the desired front wheel steering angle is calculated to control the vehicle's steering. This embodiment can use a two-degree-of-freedom vehicle model to establish a lateral dynamics model. First, the lateral forces on the entire vehicle model are analyzed. According to Newton's second law, the force relationships are:

[0211] ma y =F yr +F yf cosδ θ #(53)

[0212] Where m represents the vehicle mass, a y F represents the lateral acceleration of a vehicle. yr F represents the lateral force acting on the rear axle of the vehicle. yf δ represents the lateral force acting on the front axle of the vehicle. θ Indicates the steering angle of the front wheels.

[0213] Since the lateral motion of a vehicle is not only lateral translation, but also yaw motion during turning, Equation 53 can be written in the following form after considering the yaw motion of the vehicle:

[0214]

[0215]

[0216] in, v represents the acceleration of the vehicle along the y-direction. x This represents the vehicle's speed in the x-direction. The yaw rate of the vehicle is represented by the torque balance equation of the vehicle about the vertical axis (z-axis):

[0217]

[0218] Among them, I z This represents the yaw moment of inertia of the vehicle. The yaw acceleration of a vehicle, l a l b These represent the distances from the vehicle's front and rear axles to its center of gravity, respectively. Lateral force F yr With F yf Side slip angle α of the front and rear wheels of the vehicle f α r Related, among which,

[0219]

[0220]

[0221] in, Indicates the direction of the front wheel speed v f The angle between the vehicle's longitudinal axis x and the vehicle's longitudinal axis x. These represent the total lateral stiffness of the front and rear axles, respectively, and their values ​​are negative.

[0222] Based on the above formula, in order to derive the expressions for the magnitudes of the front and rear wheel slip angles, it is also necessary to calculate... The value of . Based on the vehicle velocity vector relationship, The expression is:

[0223]

[0224] Among them, v y It is the vehicle's lateral speed.

[0225] According to the principle of small-angle approximation, it can be approximated at this time as follows: The size is:

[0226]

[0227] Side slip angle α f α r Angle with velocity vector These are usually very small angles, so the expression for the deflection angle can be written as:

[0228]

[0229]

[0230] Therefore, the lateral force F can be calculated. yr With F yf The size is:

[0231]

[0232]

[0233] In the embodiments of this application, the front wheel steering angle generated by the vehicle is relatively small, therefore cosδ can be approximated. θ ≈1, Substituting equations 63 and 64 into equations 55 and 56, the resulting transverse dynamic equation is:

[0234]

[0235]

[0236] Using the vehicle's lateral position y and lateral velocity yaw angle and yaw rate As the state variable x of the system and the front wheel steering angle as the control variable u, the state-space equation of the lateral dynamics model of the vehicle can be derived as follows:

[0237]

[0238] in, Let A represent the rate of change of the state variables, and B be the coefficient matrices. The symbols in the state-space equations have the following meanings:

[0239]

[0240]

[0241] The lateral control rate in this embodiment can be calculated using the feedforward LQR control algorithm. The control rate calculated by the LQR control algorithm needs to satisfy the objective function:

[0242]

[0243] Where e(k) represents the state error, u(k) represents the control input, Q is the semi-positive definite state error weighting matrix, and R is the positive definite control input weighting matrix.

[0244] The optimal control law U obtained by the LQR controller is a linear function of the state error e(k):

[0245] U=-Ke(k=-[(R+B T PB) -1 B T PA]e(k)#(69)

[0246] Where A and B represent the coefficient matrices of the state-space equations of the error system; P can be represented as:

[0247] P = A T PA-A r PB(R+B r PB) -1 B T PA+Q#(70)

[0248] Through multiple iterations of the above formula, the solution for P can be converged. Then, an error system model is established, and the heading error is expressed as the difference between the vehicle's actual heading angle and the reference heading angle:

[0249] e θ =θ-θ r #(71)

[0250] Where θ represents the vehicle's actual heading angle, θ r This indicates the vehicle's reference heading angle.

[0251] Lateral position error can be represented by lateral acceleration error. The expected lateral acceleration when the vehicle travels along the reference path is:

[0252]

[0253] Among them, a yr v represents the lateral acceleration along the reference path. x R represents the longitudinal speed of the vehicle. r This represents the radius of curvature corresponding to the reference path point.

[0254] Considering the yaw motion of the vehicle, the embodiments of this application can assume that the vehicle's heading angle is approximately equal to the yaw angle, from which it can be deduced that:

[0255]

[0256]

[0257] Further derivation of the state-space equations for the error-based system is as follows:

[0258]

[0259] in, e represents the rate of change of the error of the vehicle state quantity. rr This represents the error in the vehicle's state variables. The coefficient matrices are as follows:

[0260]

[0261]

[0262]

[0263] Based on equations 69 and 75, the relationship between state error and the rate of change of error can be derived:

[0264]

[0265] According to Equation 76, it can be found that... and e rr =0 is not a solution to this differential equation. No matter what value K is obtained, the state error cannot be stabilized at 0, meaning the system has a steady-state error, which will lead to suboptimal control performance. To eliminate the steady-state error, this embodiment employs a feedforward control method. By inputting a feedforward control, the steady-state error of the system is compensated, making the system state as stable as possible, close to the reference state. The control rate with the feedforward control is:

[0266] U = -Ke rr (k)+u f #(77)

[0267] Where, the controller gain matrix K = [k1, k2, k3, k4] corresponds to the dimension of the state error matrix, u f This represents the feedforward control variable. When the system is stable... The steady-state error at this time is:

[0268]

[0269] Solve u f make e rr If the value of is as close to 0 as possible, then:

[0270]

[0271] Where, k r Indicates the curvature of the reference path.

[0272] The lateral delay compensation in this application embodiment can employ a state augmentation method for an ideal system without delay: First, a state-space equation for the lateral control system containing delay terms can be established. Considering the communication delay and actuator hysteresis of the control system, the communication delay is pure time delay, and the actuator hysteresis is inertial hysteresis. Combining the inertial hysteresis characteristics of the actuator with the original error model Equation 75, the state-space equation of the system can be obtained as follows:

[0273]

[0274] Where, δ r Indicates the actual front wheel steering angle of the vehicle, δ e τ represents the desired front wheel steering angle, and t represents the inertial hysteresis time constant of the steering actuator. d This represents the communication delay time of the control signal. Equation 80 can be converted into a discrete turntable space equation as follows:

[0275]

[0276] in, Represents the augmented state matrix of inertial hysteresis. These are the augmented coefficient matrices, where N represents the time delay t. d It is discrete into N sampling time periods.

[0277] Based on the existing optimal control theory of discrete time-delay systems, by controlling the system's past time delay (t)... d By incorporating the N front wheel steering angle control quantities calculated over time into the system's state matrix, the time-delayed system in 81 can be transformed into a new time-delay-free system:

[0278]

[0279] Equation 82 can be simplified to:

[0280]

[0281] Substituting Equation 83 into Equations 69 and 70 of the LQR control algorithm, the transverse feedback control law based on the augmented matrix and the compensation for time delay can be calculated:

[0282]

[0283] Therefore, the embodiments of this application can use the vehicle dynamics model to calculate the longitudinal and lateral time delay compensation values ​​of the vehicle.

[0284] In summary, the lane-changing decision-making and delay compensation control method based on predictive information proposed in this application can perform complex calculations on a cloud platform, thus reducing the vehicle's computing resources and improving calculation speed. This application can construct and solve the target vehicle's optimal expected speed and lane-changing strategy based on traffic vehicle prediction information within the perception range over a future period. This allows the vehicle to anticipate the impact of adverse traffic conditions and adjust its speed and lane changes in advance, effectively optimizing driving efficiency and comfort. Furthermore, this application can utilize lane-changing safety algorithms, lane-changing path planning algorithms, and delay compensation control algorithms deployed on the vehicle to respond, enabling the vehicle to dynamically plan multiple lane-changing paths during driving, balancing comfort and efficiency in the lane-changing process, effectively eliminating the impact of delay, and allowing the vehicle to accurately track reference control signals.

[0285] Next, with reference to the accompanying drawings, the lane-changing decision-making and delay compensation control method based on predictive information proposed in the embodiments of this application is described. Figure 12 This is a flowchart illustrating another lane-changing decision and delay compensation control method based on predictive information provided in an embodiment of this application.

[0286] like Figure 12 As shown, the lane-changing decision and delay compensation control method based on predictive information is applied to a vehicle. The method includes the following steps:

[0287] In step S201, the expected speed and lane-changing decision time of the vehicle are obtained from the cloud platform. The cloud platform obtains traffic information collected by the roadside perception unit, predicts the traffic information of vehicles in the perception range in the future based on the traffic information, and constructs a multi-objective optimization problem of the future driving environment of the vehicle based on the prediction information. The multi-objective optimization problem is solved to obtain the optimal expected speed and lane-changing decision time of the vehicle in the future.

[0288] In step S202, the lane-changing safety judgment algorithm, lane-changing path planning algorithm and time delay compensation control algorithm deployed in this vehicle are used to respond to the expected speed and lane-changing decision time, so as to adjust the speed in advance for lane changing.

[0289] It should be noted that the vehicle used in the lane-changing decision and delay compensation control method based on predictive information in this application embodiment is a vehicle controlled by the cloud.

[0290] The lane-changing decision-making and delay compensation control method based on predictive information proposed in this application can perform complex calculations on a cloud platform, thus reducing the vehicle's computing resources and improving calculation speed. This application embodiment can construct and solve the target vehicle's optimal expected speed and lane-changing strategy based on traffic vehicle prediction information within the perception range over a future period. This allows the vehicle to anticipate the impact of adverse traffic conditions and adjust its speed and lane changes in advance, effectively optimizing driving efficiency and comfort. Furthermore, this application embodiment can utilize lane-changing safety algorithms, lane-changing path planning algorithms, and delay compensation control algorithms deployed on the vehicle to respond. This enables the vehicle to dynamically plan multiple lane-changing paths during driving, balancing comfort and efficiency in the lane-changing process, effectively eliminating the impact of delay, and allowing the vehicle to accurately track reference control signals.

[0291] The following description, with reference to the accompanying drawings, describes a lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application. Figure 13 This is a block diagram of a lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application.

[0292] like Figure 13 As shown, the lane-changing decision and delay compensation control device 10 based on predictive information is applied to a cloud platform. The device 30 includes: a first acquisition module 310, a prediction module 320, and a distribution module 330.

[0293] The system includes a first acquisition module 310, which acquires traffic information collected by the roadside sensing unit; a prediction module 320, which predicts traffic information of vehicles within the sensing range in the future based on the traffic information, and constructs a multi-objective optimization problem of the future driving environment of the target vehicle based on the prediction information, and solves the multi-objective optimization problem to obtain the target vehicle's optimal expected speed and lane-changing decision time; and a transmission module 330, which transmits the expected speed and lane-changing decision time to the target vehicle, wherein the target vehicle uses its own lane-changing safety judgment algorithm, lane-changing path planning algorithm and time delay compensation control algorithm to respond to the expected speed and lane-changing decision time, so as to adjust its speed in advance to change lanes.

[0294] In this embodiment, the prediction module 320 is further configured to: establish a traffic vehicle state matrix based on traffic information; input the information of the traffic vehicle state matrix into a micro-car-following model, and use the micro-car-following model to calculate the predicted state transition amount of the traffic vehicle; combine the state transition amount with the discrete state space equation to iteratively deduce the predicted information of the traffic vehicle at each discrete moment in the future.

[0295] In this embodiment, the prediction module 320 is further used to: obtain the vehicle status information of the target vehicle and the road traffic speed limit information stored on the cloud platform; and construct a multi-objective optimization problem based on the prediction information, the vehicle status information and the road traffic speed limit information. The multi-objective optimization problem includes a driving cost function and constraints. The cost function reflects the optimization objectives of vehicle driving efficiency and comfort, as well as the requirements of car-following safety and lane-changing safety.

[0296] In this embodiment, the prediction module 320 is further configured to: obtain the acceleration sequence of the target vehicle; decouple the acceleration sequence of the vehicle from the lane change decision sequence solution process to discretize the lane subsequence, and discretize the possible lane change decision sequences in continuous time and space into candidate subsequences; solve the optimal acceleration sequence corresponding to each candidate subsequence from the multi-objective optimization problem, and filter the optimal acceleration sequence; sort and filter the driving costs of the filtered candidate subsequences, and take the subsequence corresponding to the lowest cost as the optimal lane change decision sequence, wherein the optimal lane change decision sequence includes the expected speed and the lane change decision time.

[0297] In this embodiment of the application, the issuing module 330 is further configured to: after receiving the lane change instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and the surrounding vehicles is greater than the safety distance, then the lane change operation is performed; otherwise, the vehicle will perform a downgraded safety following operation and slow down to follow the vehicle in front.

[0298] In this embodiment, the sending module 330 is further configured to: switch between lane-changing and straight-line planning modes based on the change of driving state; select the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, use the vehicle position receiving the lane-changing decision as the starting point coordinate of the lane-changing, set a path optimization function based on vehicle comfort and lane-changing efficiency to calculate the optimal lane-changing time, set the lane-changing process as a uniform speed lane-changing, and derive the lane-changing endpoint coordinates; and obtain the lane-changing path by fitting the starting point coordinates and the ending point coordinates of the lane-changing through a polynomial curve.

[0299] In this embodiment, the distribution module 330 is further configured to: build a longitudinal dynamics model of the vehicle, calculate the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and perform longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; build a lateral dynamics model of the vehicle, analyze the lateral dynamics model using two degrees of freedom, calculate the lateral control law through feedforward control and a linear quadratic regulator, establish a time-delayed lateral control model based on the time delay elements in the control system, and use the state augmentation method to transform the time-delayed system into a time-delay-free system to achieve lateral time delay compensation.

[0300] It should be noted that the foregoing explanation of the lane-switching decision and delay compensation control method based on predictive information also applies to the lane-switching decision and delay compensation control device based on predictive information in this embodiment, and will not be repeated here.

[0301] The lane-changing decision-making and delay compensation control device based on predictive information proposed in this application can perform complex calculations on a cloud platform, thus reducing the vehicle's computing resources and improving calculation speed. This application embodiment can construct and solve the target vehicle's optimal expected speed and lane-changing strategy based on traffic vehicle prediction information within the perception range over a future period. This allows the vehicle to anticipate the impact of adverse traffic conditions and adjust its speed and lane changes in advance, effectively optimizing driving efficiency and comfort. Furthermore, this application embodiment can utilize lane-changing safety algorithms, lane-changing path planning algorithms, and delay compensation control algorithms deployed in the vehicle to respond. This enables the vehicle to dynamically plan multiple lane-changing paths during driving, balancing comfort and efficiency in the lane-changing process, effectively eliminating the impact of delay, and allowing the vehicle to accurately track reference control signals.

[0302] The following description, with reference to the accompanying drawings, describes a lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application. Figure 14 This is a block diagram of a lane-changing decision and delay compensation control device based on predictive information according to an embodiment of this application.

[0303] like Figure 14 As shown, the lane-changing decision and delay compensation control device 20 based on predictive information is applied to a vehicle. The device 40 includes a second acquisition module 410 and a response module 420.

[0304] The second acquisition module 410 is used to acquire the vehicle's expected speed and lane-changing decision time from the cloud platform. The cloud platform acquires traffic information collected by the roadside perception unit, predicts the traffic vehicles in the perception range for a future period based on the traffic information, and constructs a multi-objective optimization problem of the vehicle's future driving environment based on the prediction information. The multi-objective optimization problem is solved to obtain the vehicle's optimal expected speed and lane-changing decision time in the future. The response module 420 is used to respond to the expected speed and lane-changing decision time using the lane-changing safety judgment algorithm, lane-changing path planning algorithm and time delay compensation control algorithm deployed in the vehicle, so as to adjust the speed and change lanes in advance.

[0305] It should be noted that the foregoing explanation of another embodiment of the lane-switching decision and delay compensation control method based on predictive information also applies to the lane-switching decision and delay compensation control device based on predictive information in this embodiment, and will not be repeated here.

[0306] The lane-changing decision-making and delay compensation control device based on predictive information proposed in this application can perform complex calculations on a cloud platform, thus reducing the vehicle's computing resources and improving calculation speed. This application embodiment can construct and solve the target vehicle's optimal expected speed and lane-changing strategy based on traffic vehicle prediction information within the perception range over a future period. This allows the vehicle to anticipate the impact of adverse traffic conditions and adjust its speed and lane changes in advance, effectively optimizing driving efficiency and comfort. Furthermore, this application embodiment can utilize lane-changing safety algorithms, lane-changing path planning algorithms, and delay compensation control algorithms deployed in the vehicle to respond. This enables the vehicle to dynamically plan multiple lane-changing paths during driving, balancing comfort and efficiency in the lane-changing process, effectively eliminating the impact of delay, and allowing the vehicle to accurately track reference control signals.

[0307] This application also provides a cloud platform, including a lane-changing decision and delay compensation control device based on predictive information as described in the above embodiments.

[0308] This application also provides a vehicle including another lane-changing decision and delay compensation control device based on predictive information as described in the above embodiments.

[0309] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0310] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0311] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0312] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, 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 (FPGAs), field-programmable gate arrays (FPGAs), etc.

[0313] Those skilled in the art will understand that all or part of the steps of the methods 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] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for lane change decision and time delay compensation control based on prediction information, characterized in that, The method is applied to a cloud platform, and includes the following steps: Acquire traffic information collected by roadside sensing units; Based on the traffic information, predict the traffic information of vehicles within the perception range in the future, and construct a multi-objective optimization problem of the future driving environment of the target vehicle based on the predicted information. Solve the multi-objective optimization problem to obtain the target vehicle's optimal expected speed and lane-changing decision time in the future. The desired speed and lane-changing decision time are sent to the target vehicle. The target vehicle responds to the desired speed and lane-changing decision time using its own lane-changing safety judgment algorithm, lane-changing path planning algorithm, and time-delay compensation control algorithm to adjust its speed and change lanes in advance. The lane-changing safety judgment algorithm includes: after receiving the lane-changing instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, a lane-changing operation is performed; otherwise, a downgraded safety following operation is performed, slowing down to follow the vehicle in front. The lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes based on the change in driving state; the straight-line selection uses the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, with the position of the vehicle receiving the lane-changing decision as the starting point for the lane change. The system calculates the optimal lane-changing time using a path optimization function based on vehicle comfort and lane-changing efficiency, and sets the lane-changing process as a uniform speed lane change, deriving the lane-changing endpoint coordinates. The lane-changing path is obtained by fitting the lane-changing starting point coordinates and the lane-changing endpoint coordinates using a polynomial curve. The time-delay compensation control algorithm includes: building a longitudinal dynamics model of the vehicle, calculating the longitudinal control law based on the input state error and proportional-integral-derivative control algorithm, and performing longitudinal time-delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; building a lateral dynamics model of the vehicle, analyzing the lateral dynamics model using two degrees of freedom, calculating the lateral control law through feedforward control and a linear quadratic regulator, establishing a time-delay lateral control model based on the time-delay elements in the control system, and using the state augmentation method to transform the time-delay system into a time-delay-free system to achieve lateral time-delay compensation.

2. The lane-changing decision-making and delay compensation control method based on predictive information according to claim 1, characterized in that, The prediction information of traffic vehicles within the sensing range for a future period based on the traffic information includes: A traffic vehicle status matrix is ​​established based on the traffic information; The information of the traffic vehicle state matrix is ​​input into the micro-car-following model, and the predicted state transition amount of the traffic vehicle is calculated using the micro-car-following model. By combining the state transition variables with the discrete state-space equations, the predictive information of traffic vehicles at each discrete moment in the future can be iteratively derived.

3. The lane-changing decision-making and delay compensation control method based on predictive information according to claim 1, characterized in that, The step of constructing a multi-objective optimization problem of the future driving environment of the target vehicle based on the predicted information includes: Obtain the vehicle status information of the target vehicle and the road traffic speed limit information stored on the cloud platform; The multi-objective optimization problem is constructed based on the predicted information, the vehicle status information, and the road traffic speed limit information. The multi-objective optimization problem includes a driving cost function and constraints. The cost function reflects the optimization objectives of vehicle driving efficiency and comfort, as well as the requirements of car-following safety and lane-changing safety.

4. The lane-changing decision-making and delay compensation control method based on predictive information according to claim 1 or 3, characterized in that, Solving the multi-objective optimization problem to obtain the target vehicle's optimal future expected speed and lane-changing decision time includes: Obtain the acceleration sequence of the target vehicle; The acceleration sequence of the vehicle is decoupled from the lane change decision sequence solution process to discretize the lane subsequence and discretize the possible lane change decision sequences in continuous time and space into candidate subsequences. The optimal acceleration sequence corresponding to each candidate subsequence is solved from the multi-objective optimization problem, and the optimal acceleration sequence is filtered. The driving costs of the filtered candidate subsequences are sorted and filtered, and the subsequence corresponding to the lowest cost is taken as the optimal lane-changing decision sequence. The optimal lane-changing decision sequence includes the expected speed and the lane-changing decision time.

5. A lane-changing decision-making and time-delay compensation control method based on predictive information, characterized in that, The method is applied to a vehicle, and the method includes the following steps: The system obtains the vehicle's expected speed and lane-changing decision time from the cloud platform. The cloud platform obtains traffic information collected by the roadside perception unit, predicts traffic information of vehicles within the perception range in the future based on the traffic information, and constructs a multi-objective optimization problem of the vehicle's future driving environment based on the predicted information. The system solves the multi-objective optimization problem to obtain the vehicle's optimal expected speed and lane-changing decision time in the future. The vehicle utilizes a lane-changing safety judgment algorithm, a lane-changing path planning algorithm, and a time-delay compensation control algorithm to respond to the desired speed and the lane-changing decision time, adjusting speed in advance for lane changes. The lane-changing safety judgment algorithm includes: upon receiving a lane-changing instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, a lane-changing operation is executed; otherwise, a downgraded safety following operation is performed, slowing down to follow the vehicle in front. The lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes based on changes in driving state; the straight-line selection uses the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, using the position of the vehicle receiving the lane-changing decision as the starting point coordinates for the lane change, and setting lane-changing parameters based on vehicle comfort and lane-changing efficiency. The optimal lane-changing time is calculated using a path optimization function, and the lane-changing process is set as a uniform speed lane change. The coordinates of the lane-changing endpoint are derived. The lane-changing path is obtained by fitting the coordinates of the lane-changing starting point and the lane-changing endpoint using a polynomial curve. The time delay compensation control algorithm includes: building a longitudinal dynamic model of the vehicle, calculating the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and performing longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; building a lateral dynamic model of the vehicle, analyzing the lateral dynamic model using two degrees of freedom, calculating the lateral control law through feedforward control and a linear quadratic regulator, establishing a time-delayed lateral control model based on the time delay elements in the control system, and using the state augmentation method to transform the time-delayed system into a time-delay-free system to achieve lateral time delay compensation.

6. A lane-changing decision-making and delay compensation control device based on predictive information, characterized in that, The device is used on a cloud platform, wherein the device includes: The first acquisition module is used to acquire traffic information collected by the roadside sensing unit; The prediction module is used to predict the traffic information of vehicles within the perception range in the future based on the traffic information, and to construct a multi-objective optimization problem of the future driving environment of the target vehicle based on the predicted information. The multi-objective optimization problem is solved to obtain the target vehicle's optimal expected speed and lane-changing decision time in the future. The distribution module is used to distribute the desired speed and the lane-changing decision time to the target vehicle. The target vehicle responds to the desired speed and the lane-changing decision time using its own lane-changing safety judgment algorithm, lane-changing path planning algorithm, and time-delay compensation control algorithm to adjust its speed and change lanes in advance. The lane-changing safety judgment algorithm includes: after receiving the lane-changing instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, a lane-changing operation is performed; otherwise, a downgraded safety following operation is performed, slowing down to follow the vehicle in front. The lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes based on the change in driving state; the straight-line selection uses the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, using the vehicle position receiving the lane-changing decision as... The starting coordinates of the lane change are determined. Based on vehicle comfort and lane change efficiency, a path optimization function is set to calculate the optimal lane change time. The lane change process is set as a uniform speed lane change, and the ending coordinates of the lane change are derived. The lane change path is obtained by fitting the starting and ending coordinates of the lane change using a polynomial curve. The time delay compensation control algorithm includes: building a longitudinal dynamics model of the vehicle; calculating the longitudinal control law based on the input state error and proportional-integral-derivative control algorithm; and performing longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system. A lateral dynamics model of the vehicle is built. The lateral dynamics model is analyzed using a two-degree-of-freedom approach. The lateral control law is calculated using feedforward control and a linear quadratic regulator. A time-delay lateral control model is established based on the time delay elements in the control system. The state augmentation method is used to transform the time-delayed system into a time-delay-free system, achieving lateral time delay compensation.

7. A lane-changing decision-making and delay compensation control device based on predictive information, characterized in that, The device is applied to a vehicle, wherein the device includes: The second acquisition module is used to acquire the vehicle's expected speed and lane-changing decision time from the cloud platform. The cloud platform acquires traffic information collected by the roadside perception unit, predicts traffic information of vehicles within the perception range in the future based on the traffic information, and constructs a multi-objective optimization problem of the vehicle's future driving environment based on the prediction information. The multi-objective optimization problem is solved to obtain the vehicle's optimal expected speed and lane-changing decision time in the future. The response module is used to respond to the desired speed and the lane-changing decision time using the lane-changing safety judgment algorithm, lane-changing path planning algorithm, and time delay compensation control algorithm deployed in the vehicle, so as to adjust the speed for lane changing in advance. The lane-changing safety judgment algorithm includes: after receiving the lane-changing instruction, the vehicle performs a safety distance judgment; if the actual distance between the vehicle and surrounding vehicles is greater than the safety distance, a lane-changing operation is performed; otherwise, a downgraded safety following operation is performed, slowing down to follow the vehicle in front. The lane-changing path planning algorithm includes: switching between lane-changing and straight-line planning modes according to the change in driving state; the straight-line selection uses the road centerline closest to the vehicle's center coordinates as the reference path for straight-line driving in the current lane, using the position of the vehicle receiving the lane-changing decision as the starting point coordinates of the lane-changing, and considering vehicle comfort and lane-changing... The efficiency setting path optimization function calculates the optimal lane-changing time and sets the lane-changing process as a uniform speed lane-changing, deriving the coordinates of the lane-changing endpoint; the lane-changing path is obtained by fitting the coordinates of the lane-changing starting point and the lane-changing endpoint using a polynomial curve; the time delay compensation control algorithm includes: building a longitudinal dynamic model of the vehicle, calculating the longitudinal control law based on the input state error and the proportional-integral-derivative control algorithm, and performing longitudinal time delay compensation based on the signal input time delay and actuator time delay in the longitudinal control system; building a lateral dynamic model of the vehicle, analyzing the lateral dynamic model using two degrees of freedom, calculating the lateral control law through feedforward control and a linear quadratic regulator, establishing a time-delay lateral control model based on the time delay elements in the control system, and using the state augmentation method to transform the time-delay system into a time-delay-free system to achieve lateral time delay compensation.

8. A cloud platform, characterized in that, Includes the lane-changing decision and delay compensation control device based on predictive information as described in claim 6.

9. A vehicle, characterized in that, Includes the lane-changing decision and delay compensation control device based on predictive information as described in claim 7.

Images