Vehicle scheduling method and device, vehicle and storage medium

Abstract

The invention relates to the technical field of vehicle control, and discloses a vehicle scheduling method and device, a vehicle and a storage medium, and the method comprises the steps: obtaining the driving state information of a vehicle before the vehicle is driven to an intersection without traffic lights, and obtaining the relative positions and predicted driving path information of other vehicles in a current intersection region; in each sampling period, determining a corresponding multi-objective optimization function based on the traffic interaction scene where the vehicle is located; wherein the traffic interaction scene is determined based on the driving state information of the vehicle and the relative positions and predicted driving path information of other vehicles; constructing a vehicle longitudinal dynamic model based on the driving state information of the vehicle, and constructing a model prediction control problem in combination with a multi-objective optimization function and a safety constraint condition; and solving the model prediction control problem to obtain a target control input sequence in a control time domain, and outputting the first control input quantity in the target control input sequence to an execution mechanism of the own vehicle to realize real-time regulation and control of the own vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a vehicle scheduling method, device, vehicle, and storage medium. Background Technology

[0002] Traffic intersections, as critical nodes in road networks where vehicles converge and diverge most frequently, are core elements of urban traffic flow organization. Due to the high convergence of traffic from multiple directions at these intersections, conflicts are easily triggered, leading to increased traffic congestion and delays, thus becoming bottlenecks in the entire traffic system, especially when there are no traffic lights or the traffic lights are malfunctioning. Simultaneously, the complex interactions at intersections also make them high-risk areas for traffic accidents, posing a serious challenge to driving safety. Therefore, achieving safe and orderly passage of vehicles at intersections without traffic lights is not only crucial for improving road operational efficiency but also a vital issue that must be addressed to advance the implementation of autonomous driving in urban areas. Summary of the Invention

[0003] In view of this, embodiments of this application provide a method, a vehicle dispatching device, a vehicle, and a computer-readable storage medium.

[0004] In a first aspect, embodiments of this application provide a method, including: Obtain the driving status information of the vehicle before it reaches the intersection without traffic lights, and obtain the relative position and predicted driving path information of other vehicles in the current intersection area; Within each sampling period, a corresponding multi-objective optimization function is determined based on the traffic interaction scenario in which the vehicle is located; wherein, the traffic interaction scenario is determined based on the driving state information of the vehicle and the relative positions and predicted driving path information of other vehicles; Based on the vehicle's driving state information, a longitudinal dynamics model of the vehicle is constructed, and a model predictive control problem is constructed by combining the multi-objective optimization function and safety constraints. Solve the model predictive control problem to obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

[0005] In an optional implementation, constructing a vehicle longitudinal dynamics model based on the vehicle's driving state information includes: Based on the nonlinear continuous dynamic equation of the vehicle's longitudinal motion, local linearization is performed at the initial state of the current sampling period to obtain a linearized state-space expression. The state-space expression is discretized to obtain a discrete-time state-space model; The ground rolling friction resistance term is incorporated into the control input variable to form an equivalent control input quantity. It is assumed that the system state transition matrix and the control input matrix remain unchanged in the prediction time domain, so as to construct the vehicle longitudinal dynamics model with displacement and velocity as state variables and the equivalent control input quantity as control variables.

[0006] In an optional implementation, a model predictive control problem is constructed based on the vehicle longitudinal dynamics model, combined with the multi-objective optimization function and the safety constraints, including: Based on the longitudinal dynamics model, the predicted state sequence of the vehicle in the prediction time domain is derived; The security constraints are transformed into a set of inequality constraints relating to the state prediction sequence and the control sequence; Based on the multi-objective optimization function and the set of inequality constraints, the model predictive control problem is constructed.

[0007] In optional implementations, the traffic interaction scenarios include conflict-free scenarios, linear conflict scenarios, point-based conflict scenarios, and composite conflict scenarios: The conflict-free scenario refers to a risk scenario in which there is no potential conflict between the vehicle and other vehicles. The linear conflict scenario refers to a linear conflict scenario in which the vehicle is at risk of rear-end collision with other vehicles in the same entrance lane. The point-like conflict scenario refers to a point-like conflict scenario in which the vehicle and other vehicles from different entrance lanes have intersecting paths within the intersection. The composite conflict scenario is a composite conflict scenario that simultaneously includes linear conflict risk and point-based conflict risk.

[0008] In an optional implementation, determining the corresponding multi-objective optimization function based on the traffic interaction scenario includes: When the vehicle is in the conflict-free scenario, a first-class multi-objective optimization function is adopted, with traffic efficiency index, energy economy index and ride comfort index as optimization sub-objectives. When the vehicle is in the linear conflict scenario, a second type of multi-objective optimization function is adopted, with the line safety performance index as the main optimization sub-objective and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives. When the vehicle is in the point-based conflict scenario, a third type of multi-objective optimization function is adopted, which takes the conflict point safety performance index as the main optimization sub-objective and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives. When the vehicle is in the complex conflict scenario, a fourth type of multi-objective optimization function is adopted, which takes the route safety performance index and the conflict point safety performance index as the main optimization sub-objectives, and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives.

[0009] In an optional implementation, the safety constraints include minimum safe time interval constraints at conflict points, minimum safe distance constraints between vehicles in the same lane, physical feasible region constraints for vehicle speed, physical feasible region constraints for acceleration, and physical feasible region constraints for jerk.

[0010] In an optional implementation, the equivalent control input is the difference between the actual acceleration and the deceleration caused by rolling friction.

[0011] Secondly, embodiments of this application provide a vehicle dispatching device, comprising: The acquisition module is used to acquire the driving status information of the vehicle before it reaches the intersection without traffic lights, and to acquire the relative position and predicted driving path information of other vehicles in the current intersection area. The selection module is used to determine the corresponding multi-objective optimization function based on the traffic interaction scenario in which the vehicle is located within each sampling period; wherein, the traffic interaction scenario is determined based on the driving state information of the vehicle and the relative positions and predicted driving path information of other vehicles. The construction module is used to construct a longitudinal dynamics model of the vehicle based on the driving state information of the vehicle, and to construct a model predictive control problem by combining the multi-objective optimization function and safety constraints. The control module is used to solve the model predictive control problem, obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

[0012] Thirdly, embodiments of this application provide a vehicle, the vehicle including a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the vehicle scheduling method described above.

[0013] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed on a processor, implements the vehicle scheduling method described above.

[0014] The embodiments of this application have the following beneficial effects: This application obtains the driving status information of a vehicle before it reaches an intersection without traffic lights, and combines this with the relative positions and predicted driving path information of other vehicles in the current intersection area to identify the traffic interaction scenario in real time. Then, it dynamically matches the corresponding multi-objective optimization function based on the scenario characteristics. On this basis, based on the vehicle's longitudinal dynamics model, it constructs a model predictive control problem by integrating multi-objective optimization objectives and safety constraints, solves the optimal control input sequence in the control time domain, and outputs the first control quantity to the vehicle's actuator, achieving real-time closed-loop control of the vehicle. This method can accurately coordinate the passage behavior of multiple vehicles in complex intersection environments without relying on traditional traffic light control, effectively avoiding safety hazards such as rear-end collisions and lateral conflicts, while reducing unnecessary deceleration and stopping, thereby improving traffic efficiency and driving stability. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A first flowchart of a vehicle dispatching method according to an embodiment of this application is shown; Figure 2 A second flowchart of the vehicle dispatching method according to an embodiment of this application is shown; Figure 3 A third flowchart of the vehicle dispatching method according to an embodiment of this application is shown; Figure 4 A first schematic diagram of an intersection scenario according to an embodiment of this application is shown; Figure 5 A second schematic diagram of an intersection scenario according to an embodiment of this application is shown; Figure 6 A schematic diagram of a vehicle dispatching device according to an embodiment of this application is shown. Detailed Implementation

[0017] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0018] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0020] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0021] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0022] The vehicle dispatching method will be explained below with reference to some specific embodiments.

[0023] Figure 1 A schematic flowchart of a vehicle scheduling method according to an embodiment of this application is shown. Exemplarily, the vehicle scheduling method includes steps S110-S140: Step S110: Obtain the driving status information of the vehicle before it reaches the intersection without traffic lights, and obtain the relative position and predicted driving path information of other vehicles in the current intersection area.

[0024] In this embodiment, the vehicle is an intelligent connected vehicle with V2X (Vehicle-to-Everything) communication capability. The vehicle can work in collaboration with external communication devices through onboard sensors to complete information collection tasks.

[0025] In this step, the autonomous vehicle first acquires its own driving status information through its onboard perception system. This driving status information includes, but is not limited to, dynamic parameters such as the vehicle's current displacement, longitudinal velocity, acceleration, jerk, heading angle, and lane number. Displacement and velocity can be obtained through a high-precision GNSS / IMU combined positioning system; acceleration information can be directly measured by the onboard inertial measurement unit (IMU) or estimated through velocity differentiation; and lane information can be determined through high-precision map matching and lane line recognition algorithms. Additionally, the autonomous vehicle receives status data packets broadcast by other vehicles within the current intersection area via the V2X communication interface. This data originates from other autonomous vehicles within the intersection's control range or traditional networked vehicles equipped with OBUs (On-Board Units). Each other vehicle periodically sends messages containing its own driving status information and its predicted travel path for the next period.

[0026] Among them, the predicted driving path information refers to a set of future trajectory points generated by other vehicles based on their destination, current navigation route, lane keeping strategy and local obstacle avoidance logic. It is usually represented as a series of two-dimensional spatial coordinate points and their corresponding timestamps, which are used to characterize the spatial location that the vehicle is expected to pass through in the next few seconds.

[0027] In this embodiment, after receiving information from other vehicles, the vehicle first performs time synchronization processing, unifying the information of all vehicles under a local clock reference. Then, it performs coordinate transformation, converting the global geographic coordinate system to a local Cartesian coordinate system with the vehicle as the origin, facilitating subsequent relative position calculations. Through the above process, the vehicle can construct a complete local traffic situation map, covering its own precise status and the relative position distribution, direction of movement, and expected driving trajectory of other vehicles, serving as the basic input for subsequent scene recognition and collaborative decision-making.

[0028] Step S120: In each sampling period, determine the corresponding multi-objective optimization function based on the traffic interaction scenario in which the vehicle is located.

[0029] The traffic interaction scenario is determined based on the vehicle's driving status information, the relative positions of other vehicles, and predicted driving path information. It can be understood that the traffic interaction scenario in which the vehicle is located can be determined by the onboard collaborative decision-making module and updated in real time at the beginning of each sampling period to ensure that the control strategy can respond to the dynamically changing traffic environment.

[0030] Based on the different types of potential conflicts between vehicles, this embodiment can classify traffic interaction scenarios into four categories: conflict-free scenarios, linear conflict scenarios, point-based conflict scenarios, and complex conflict scenarios. Each category corresponds to different safety risk levels and control priorities, and therefore the corresponding multi-objective optimization functions also differ. It is understood that the above classification of traffic interaction scenarios is merely exemplary; in some implementations, targeted classifications can be made as needed.

[0031] When optimizing multi-objective functions for traffic interaction scenarios, the most critical issue for multi-vehicle cooperative driving at intersections is ensuring the safe and smooth passage of all vehicles through the intersection. Secondary considerations include improving traffic efficiency, reducing fuel consumption, and enhancing the passenger experience. Based on a distributed control strategy, the problem of safe cooperative scheduling of multiple vehicles at intersections can be decomposed into a problem of safe cooperative control between pairs of vehicles. The specific states and positional relationships between any two vehicles within the intersection area are as follows: Figure 3 As shown. From a safety perspective, the state and positional relationship of two vehicles in an intersection area can be divided into whether a collision is possible. If no collision is likely, both vehicles pass through the intersection area in the most time-optimal manner; if a collision is possible, it is necessary to further clarify the location and area where the potential conflict occurs. The first scenario is when both vehicles enter the intersection area from the same entrance lane; in this case, it is necessary to control both vehicles within the control zone to avoid rear-end collisions. Additionally, if the two vehicles' travel paths in the intersection area are completely identical, it is necessary to control both vehicles to avoid conflict when passing through a series of obstacles within the intersection. The second scenario is when the two potentially conflicting vehicles enter the intersection area from different entrance lanes; preventing a conflict at a common point of conflict is a safety concern.

[0032] like Figure 4 (The painted vehicles are self-driving cars) As shown, the four types of scenarios defined in this way can be defined as follows: A conflict-free scenario is a scenario where there is no potential risk of conflict between the vehicle and other vehicles; such as Figure 4 As shown in (a), the vehicle and other vehicles are in two different lanes and are traveling in different directions, so they will not intersect in time and space.

[0033] A linear conflict scenario is a scenario where there is a risk of rear-end collision between the vehicle and other vehicles in the same entrance lane; such as... Figure 4 As shown in (b), if there are other vehicles in front of your vehicle in the same entrance lane, a rear-end collision may occur. This risk is mainly reflected in the longitudinal movement dimension, which is manifested as the distance between vehicles gradually decreasing or the relative speed exceeding the safety threshold. It is a continuous risk along the lane direction.

[0034] Point-based conflict scenarios refer to point-based conflict scenarios where a vehicle and other vehicles from different entrance lanes have intersecting paths within the intersection; such as... Figure 4 As shown in (c), the vehicle and other vehicles from different entrance lanes have a path intersection point inside the intersection, and their expected arrival times at the intersection point are similar, which may lead to the risk of them simultaneously occupying the same area in space. This type of conflict is concentrated at a specific geometric point and is therefore called a point conflict.

[0035] A complex conflict scenario is one that simultaneously includes both linear conflict risks and point-based conflict risks. For example... Figure 4 As shown in (d), the vehicle faces the risk of rear-end collision from the vehicle in front, and also needs to coordinate the passage order with vehicles coming from the side at the point of conflict.

[0036] After determining the traffic interaction scenario in which the vehicle is located, it is necessary to match the corresponding multi-objective optimization function.

[0037] Specifically, when the vehicle is in a conflict-free scenario, a first-class multi-objective optimization function is adopted, with traffic efficiency, energy economy, and ride comfort as optimization sub-objectives. It can be understood that in this situation, when the vehicle approaches and passes through an intersection area, even if other vehicles are traveling within the intersection area, they will not interfere with the vehicle's future driving state. Therefore, theoretically, the vehicle is safe regardless of how it travels within the intersection area within the constraints. In this case, the multi-objective optimization function does not need to use safety control as the primary optimization index of the controller; other optimization indices, including but not limited to traffic efficiency, energy economy, and ride comfort, serve as the main optimization sub-objectives.

[0038] Then the multi-objective optimization function can be expressed as: ;in, This indicates a traffic efficiency indicator. Indicates energy economic indicators, The values ​​represent ride comfort indicators; w1, w2, and w3 are weighting coefficients for different optimization objectives, which can be adjusted according to intersection scenarios, safety requirements, and vehicle performance.

[0039] When the vehicle is in a linear conflict scenario, a second-type multi-objective optimization function is adopted, with lane safety performance as the primary optimization sub-objective and traffic efficiency, energy economy, and ride comfort as secondary optimization sub-objectives. It can be understood that in this scenario, to prevent rear-end collisions with vehicles in front and behind in the same lane, ensuring a safe longitudinal distance must be prioritized. Therefore, the lane safety performance index is introduced as the primary optimization sub-objective in the second-type multi-objective optimization function. This index is typically constructed as a penalty term based on the deviation between the actual vehicle distance and the minimum safe vehicle distance, and given a high weight. The other three indices are retained as secondary optimization sub-objectives, but their weights are significantly lower than the safety term.

[0040] At this point, the multi-objective optimization function can be expressed as: ;in, Indicates the safety performance indicators of the line. As the weight of the line safety performance indicators, Greater than w1, w2, and w3.

[0041] When the vehicle is in a point-based conflict scenario, a third type of multi-objective optimization function is adopted, with the conflict point safety performance index as the primary optimization sub-objective and traffic efficiency, energy economy, and ride comfort as secondary optimization sub-objectives. It can be understood that the core risk in this situation lies in the time conflict with other vehicles at the conflict point within the intersection. The third type of multi-objective optimization function uses the conflict point safety performance index as the primary optimization sub-objective. This index is constructed based on the time difference between the expected arrival times of each vehicle at the conflict point; if the time difference is less than the set minimum safe interval, a stronger penalty is applied. By optimizing the control sequence, the vehicle adjusts its passage timing to avoid arriving at the conflict point at the same time as other vehicles.

[0042] At this point, the multi-objective optimization function can be expressed as: ;in Indicates the safety performance indicators at the point of conflict. The weights of the safety performance indicators at the conflict points Greater than w1, w2, and w3.

[0043] When a vehicle is in a complex conflict scenario, a fourth type of multi-objective optimization function is adopted. This function primarily optimizes lane safety performance and conflict point safety performance as sub-objectives, while using traffic efficiency, energy economy, and ride comfort as auxiliary sub-objectives. It can be understood that the vehicle faces a dual threat, needing to simultaneously manage rear-end collision prevention within the lane and avoid conflicts in both time and space. Therefore, a fourth type of multi-objective optimization function is used, which includes both lane safety performance and conflict point safety performance as primary sub-objectives, jointly guiding the optimization direction, while the other three serve as auxiliary objectives for adjustment.

[0044] At this point, the multi-objective optimization function can be expressed as: ;in, Greater than w1, w2, and w3.

[0045] The following section explains each of the above optimization sub-objectives.

[0046] Traffic efficiency indicators This metric measures the operational efficiency of a vehicle navigating through an intersection, aiming to reduce unnecessary deceleration and waiting time. Efficiency is improved by minimizing the deviation between the vehicle's actual predicted speed and its expected speed within the prediction time domain. Specifically, it sets... Let k be the expected speed at step k. Given the predicted speed at the corresponding time, the traffic efficiency index can be defined as the sum of squared speed deviations, i.e. ;in, This is for predicting the time domain length. By minimizing this term, vehicles can approach the desired travel speed as closely as possible without violating safety constraints, thus improving overall traffic efficiency.

[0047] Energy economic indicators Used to assess the energy consumption level of a vehicle during control, primarily reflecting the energy costs associated with driving and braking operations. Since vehicle energy consumption is closely related to longitudinal acceleration (i.e., control input), abrupt acceleration and deceleration lead to higher power demands; therefore, the sum of the squares of the control inputs is used as an approximate measure of energy consumption. Specifically, let... If the control input for step k (i.e., the equivalent acceleration command) is given, then the energy economy index is defined as follows: Minimizing this helps to suppress frequent or excessive acceleration and deceleration, reduce the load on the powertrain, and improve fuel or electrical energy utilization efficiency.

[0048] Ride comfort index The core influencing factor for evaluating the subjective driving experience of occupants during vehicle operation is the rate of change of acceleration, i.e., jerk. Large jerks can cause vehicle vibration or sudden changes in posture, leading to discomfort. Therefore, this embodiment uses the smoothness of changes in control input as a quantifiable criterion for comfort. Let... and If the control input values ​​are two adjacent sampling times, then the ride comfort index is defined as follows: By minimizing this, the system tends to generate a smooth acceleration curve, effectively suppressing undesirable driving behaviors such as sudden braking and jerking, thereby improving passenger comfort.

[0049] Line safety performance indicators This metric measures a vehicle's ability to maintain a safe longitudinal distance from other vehicles in the same entry lane, preventing rear-end collisions. It is constructed based on the relative displacements between adjacent vehicles in the prediction time domain. Let d(k) be the predicted displacement of the vehicle at step k, and let d(k) be the corresponding predicted displacement of the vehicle in the same lane. The actual distance between the two is The preset minimum safe following distance is When the actual spacing is less than this threshold, the system determines that a safety hazard exists and suppresses it by constructing a penalty term. Specifically, the line safety performance index can be defined as... By introducing this term into the function and setting a high weight, the model predictive controller will prioritize adjusting the vehicle's acceleration and increasing the distance between vehicles, thereby avoiding the risk of rear-end collisions.

[0050] Safety performance indicators at conflict points This is used to measure the temporal separation of a vehicle from other vehicles arriving from different entry lanes at the intersection of their internal paths, preventing spatial conflicts caused by simultaneous arrivals. Let a geometric conflict point be the location where the trajectories of the two vehicles intersect, and let the expected arrival time of the vehicle at that point be... His car is expected to arrive at the following time: The preset minimum safe time interval is If the time difference between the arrival of the two vehicles is less than this threshold, a collision is considered possible. Therefore, the abrupt safety performance index can be expressed as: This metric iterates through all conflict points where paths intersect, accumulating the degree of violation over time. During optimization, the controller automatically adjusts the vehicle's acceleration and deceleration strategies to allow it to pass through conflict areas earlier or later, ensuring sufficient time intervals between it and other vehicles.

[0051] Step S130: Construct a longitudinal dynamics model of the vehicle based on the vehicle's driving state information, and construct a model predictive control problem by combining a multi-objective optimization function and safety constraints.

[0052] The safety constraints include the minimum safe time interval constraint at the conflict point, the minimum safe distance constraint between vehicles in the same lane, the physical feasible region constraint of vehicle speed, the physical feasible region constraint of acceleration, and the physical feasible region constraint of jerk.

[0053] In some implementations, such as Figure 2 As shown, a longitudinal dynamics model of the vehicle is constructed based on the vehicle's driving state information, including steps S210-S230: Step S210: Based on the nonlinear continuous dynamic equation of the vehicle's longitudinal motion, local linearization is performed at the initial state of the current sampling period to obtain a linearized state-space expression.

[0054] Specifically, we first establish a longitudinal nonlinear continuous dynamic system for a vehicle traveling on a flat road, neglecting slope resistance. Let the system's state variables be... ,in, Indicates the actual displacement of the vehicle. The actual longitudinal speed of the vehicle is represented by the control input variable u(t), which represents the actual acceleration command generated by the drive or braking system.

[0055] According to Newton's second law, the longitudinal acceleration of a vehicle is affected by air resistance, rolling friction, and control input, and its dynamic relationship can be expressed as: ; In the formula, For a nonlinear continuous system, it is the state transition function. y(t) is the system's output function; y(t) is the system's q-dimensional output variable, which can be set to velocity, displacement, or other observables as needed; M is the vehicle's mass. The air drag coefficient, air density, Let g be the vehicle's frontal area, and g be the acceleration due to gravity. It is the coefficient of rolling friction between the ground and the tire.

[0056] To facilitate the design and online solution of the subsequent model predictive controller, this nonlinear system can be simplified. Specifically, the deceleration term caused by rolling friction can be simplified. Merge into the control input variable and define a new equivalent control input. ,Right now: The original system can then be rewritten as: ; This transformation converts the constant frictional influence, which originally existed as a disturbance term, into part of the control channel, thereby reducing the number of external disturbance parameters and facilitating the construction of subsequent optimization problems.

[0057] Since the model predictive controller needs to complete the solution within each sampling period, and the original nonlinear system is difficult to use directly for efficient numerical computation, it needs to be linearized. Based on the basic principles of model predictive control, within each sampling period, the current time is set to t, and the measured state variables at that time are used. and equivalent control input As the operating point, local linearization is performed on the nonlinear discrete system.

[0058] The Euler method is used to perform linear discretization of the nonlinear system, resulting in the state-space expression of the linear discrete system of vehicle longitudinal dynamics. ; ; ;in, and These represent the points of the nonlinear continuous system of vehicle longitudinal dynamics at points. , The 2 obtained after linear discretization 2D state transition matrix and 2 3D control input matrix, and q dimensional state output matrix and q The dimensional control output matrix, the specific form of which can be set according to the system output requirements, wherein... The transfer matrix vector is the linear discretization of the system.

[0059] Combining the rewritten formula and the state-space expression, we can obtain the specific... and expression.

[0060] in, ; If we assume that the vehicle is moving at approximately a constant speed, then , and , These parameters can be considered as constant values, further simplifying the calculation process.

[0061] Step S220: Discretize the state-space expression to obtain a discrete-time state-space model.

[0062] In this step, the rewritten continuous system will be transformed into a discrete system using the forward Euler method. ; ;in, This represents the feasible region of the system state. This is the feasible region for system control input.

[0063] This discretization process enables the system to be iteratively updated periodically on the digital controller, thereby meeting the requirements of the on-board real-time control platform for deterministic response.

[0064] Step S230: The ground rolling friction resistance term is merged into the control input variable to form an equivalent control input quantity. It is assumed that the system state transition matrix and the control input matrix remain unchanged in the prediction time domain, so as to construct a vehicle longitudinal dynamics model with displacement and velocity as state variables and equivalent control input quantity as control variables.

[0065] The equivalent control input is the difference between the actual acceleration and the deceleration caused by rolling friction.

[0066] As an example, to further reduce the burden of online optimization, an invariant control variable is applied to the nonlinear discrete system in the prediction time domain. This yields reference values ​​for the system state variables. ,but ; ; .

[0067] Based on this, in order to reduce the overhead of repeatedly calculating the Jacobian matrix in each prediction step, the following engineering simplification assumptions are made: throughout the prediction time domain Within this period, the system state transition matrix, control input matrix, and bias terms remain unchanged, i.e.: ; ; This assumption implies that within a prediction window, the vehicle's operating conditions change gradually, and the system matrix can be considered constant. This strategy reduces the computational complexity of the solver, ensuring that the controller can complete the optimization task within a finite time.

[0068] Therefore, the final vehicle longitudinal dynamics model is a simplified discrete model with displacement and velocity as state variables and equivalent control input as control variables.

[0069] Furthermore, in order to effectively embed the jerk physical feasible region constraint (i.e., the control input rate of change constraint) into the model predictive control problem and ensure that the optimization solution meets the vehicle actuator response capability requirements, this embodiment adopts control input increment. As a decision variable for the controller, rather than the original equivalent control input. Specifically, the control input increment is defined as the difference between two adjacent sampling times, i.e. .

[0070] Based on this, the original state variables are augmented to construct a new composite state vector: ;in, For the original two-dimensional state variables, This represents the equivalent control input for the previous time step. Accordingly, the augmented state transition matrix, control input matrix, and bias term are constructed as follows: , This leads to a new linear discrete state-space expression. = This augmented model incorporates historical values ​​of the control input into the state description, allowing the current control increment to directly influence the state evolution at the next moment, thus enabling the jerk constraint to be directly expressed as: This is then incorporated as a linear inequality constraint into the subsequent optimization problem. Furthermore, the system matrix is ​​assumed to remain constant throughout the prediction time domain. This approach effectively enhances the controller's adaptability to the dynamic characteristics of the actuators while avoiding the risk of violating other safety constraints due to post-processing corrections to the control sequence.

[0071] In some implementations, such as Figure 3 As shown, a model predictive control problem is constructed based on the vehicle's longitudinal dynamics model, combined with a multi-objective optimization function and safety constraints, including steps S310-S330: Step S310: Based on the longitudinal dynamics model, derive the predicted state sequence of the vehicle in the prediction time domain.

[0072] In this step, the vehicle longitudinal dynamics discrete model constructed in step S230 is used, with the initial state of the current sampling period known. Under the premise of predicting the future time domain The system state within the vehicle is recursively calculated to obtain the predicted state sequence of the vehicle.

[0073] Specifically, the following linear time-invariant discrete state-space model is used as the basis for prediction, namely: k=t, t+1···, t+ ;in, The state vector, which is the state vector starting at time t and predicting time k, includes displacement and velocity. For equivalent control input; , and These are the state transition matrix, control input matrix, and bias term, linearized and invariant at the current operating point, with a prediction time domain length of [missing information]. The sampling period can be set to 10–50, corresponding to a total prediction time of 1–5 seconds, to balance prediction capability and computational burden. Control time domain. (3 to 10 steps can be taken), and it is assumed that the control input remains unchanged after the control time domain, that is: ; through this equation of state from By progressively extrapolating, we can obtain state estimates for each future time step, thus forming a complete sequence of predicted states. This predicted state sequence reflects the expected trajectory of the vehicle over a period of time under the current control strategy. It serves as the basic input data for constructing various sub-objectives (such as traffic efficiency and ride comfort) and safety constraints (such as minimum distance between vehicles in the same lane and time interval between conflict points) in the multi-objective optimization function.

[0074] Since the prediction process is based on a linear model with fixed coefficients, the entire state prediction can be transformed into an affine function of the initial state and the control input sequence, which is convenient for efficient processing in standard optimization solvers such as quadratic programming.

[0075] Step S320: Transform the security constraints into a set of inequality constraints concerning the state prediction sequence and the control sequence.

[0076] The minimum safe time interval constraint at the conflict point means that the time interval between all vehicles passing the same conflict point must be greater than or equal to a certain set minimum safe time Δt. For example... Figure 5As shown, vehicles from different entrance lanes at an intersection will enter the intersection area along their respective planned paths, forming conflict points at certain geometric locations. For example, vehicle V13¹ from entrance 1 heading towards exit 3 and vehicle V32¹ from entrance 3 heading towards exit 2 form conflict point C13,23 at their path intersection. This conflict point indicates that the two vehicles will occupy the same space in space, so if the two vehicles arrive at this location simultaneously or too close together in time, a collision risk may occur. The geometric intersection points marked by the red dashed lines and red dots in Figure 5 are typical examples of conflict points. If any two vehicles' paths both cross this conflict point, they must adhere to the aforementioned time interval constraint to ensure that they do not simultaneously occupy the same physical space at the conflict point, thus guaranteeing the safety of multi-vehicle cooperative passage in the intersection area. This constraint can be expressed as: Where j represents the order in which vehicles arrive at the intersection, and ab and cd represent the points of conflict where the ab and cd entrance / exit routes intersect.

[0077] It is understandable that for any two vehicles passing through the same conflict point, their arrival order is not predetermined, but dynamically determined by the model predictive controller during the optimization process. That is, if vehicle j passes the conflict point first, then its arrival time with vehicle j+1 must satisfy... If vehicle j+1 passes the conflict point first, then its passing time with vehicle j must satisfy the following condition: If either of the above two inequalities holds true, it will ensure that the time interval between the passage of two vehicles at the same conflict point is not less than the set minimum safe time Δt, thereby ensuring that the two vehicles will not occupy the same physical space at the conflict point at the same time, and improving the safety of coordinated passage at the intersection.

[0078] The minimum safe following distance constraint in the same lane prevents rear-end collisions between vehicles entering the control zone in the entrance lane. Setting a minimum safe following distance ensures the safety of vehicles within the control zone, and can be expressed as: ;in, and Δd represents the displacement of the front and rear vehicles traveling in the same entrance lane. Δd is the minimum safe following distance between vehicles traveling in the same entrance lane.

[0079] The physical feasible region constraint for vehicle speed mainly considers the traffic speed restrictions in the intersection area. The constraints on the maximum and minimum longitudinal speeds can be defined as follows: and These are the minimum and maximum longitudinal speeds allowed for a vehicle to travel within the intersection area, respectively, while vcontrol is the speed at which the vehicle travels within the intersection area.

[0080] The physical feasible region constraint for acceleration can be expressed as: and These are the minimum and maximum permissible control inputs of the system, respectively, and u is the system control input.

[0081] The physical feasible region constraint for jerk acceleration, i.e., the rate of change of the control input should meet the requirements of actuator response capability and occupant comfort, can be expressed as: ; and These are the minimum and maximum permissible control increments of the system within a sampling period, respectively. This is the control increment for the system sampling period.

[0082] All of the above constraints are transformed into a series of linear or nonlinear inequalities, using the predicted state sequence and control input sequence as variables, thus forming a complete set of constraints.

[0083] Step S330: Based on the multi-objective optimization function and the set of inequality constraints, construct the model predictive control problem.

[0084] As an example, using the predicted state sequence and control input sequence established in step S310 as decision variables, the multi-objective optimization function selected in step S120 is used as the objective function. Its form is adaptively selected based on the current traffic interaction scenario, and the priority relationship between different sub-objectives is reflected through weight coefficients. This objective function has been parameterized as a quadratic expression about the predicted state sequence and control input sequence, suitable for processing by standard numerical optimization solvers.

[0085] Simultaneously, the set of inequality constraints generated in step S320, including the minimum safe time interval constraint for conflict points, the minimum safe distance constraint for vehicles in the same lane, speed constraints, acceleration constraints, and jerk constraints, are uniformly organized into a set of linear or piecewise linear inequalities concerning the predicted state sequence and the control input sequence, forming the constraint boundary of the optimization problem.

[0086] Furthermore, the discrete model of vehicle longitudinal dynamics established in step S230 is used as the system dynamic equation and embedded in the optimization problem as an equality constraint (i.e., This dynamic coupling ensures that all predicted states satisfy the vehicle's kinematic laws. Ultimately, the objective function, the set of inequality constraints, and the system dynamic equations together constitute a structured finite-time optimal control problem. This problem has a clear mathematical expression and can be solved online by an onboard real-time optimization solver in each sampling period, outputting the optimal control sequence to achieve closed-loop control of the vehicle's driving behavior.

[0087] Step S140: Solve the model predictive control problem to obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

[0088] Using the optimization solver on the vehicle-mounted embedded computing platform, the finite-time optimal control problem composed of the received current state information, prediction model structure, objective function and constraints is solved in real time numerically.

[0089] The solution process aims to minimize the multi-objective cost function, searching for the control input sequence that optimizes the overall performance while satisfying all safety constraints. Since the model predictive control problem has been linearized and discretized, and employs the fixed system matrix assumption, its mathematical form can be transformed into a standard quadratic programming (QP) problem or a convex optimization problem, exhibiting good convergence and solution efficiency.

[0090] After the solution is completed, the control time domain is obtained. An optimal equivalent control input sequence within: Based on the rolling time-domain optimization mechanism of model predictive control, only the first control input in the sequence is taken. This is then converted into usable vehicle control commands. Specifically, by combining the vehicle's dynamic characteristics with the actuator interface protocol, the equivalent acceleration command is converted into throttle opening, braking pressure, or torque request signals, which are then sent to the vehicle's drive or braking actuators via the CAN bus.

[0091] Subsequently, at the end of the current sampling period, the control command is applied to the vehicle's power system to achieve real-time control of the vehicle's longitudinal movement. When the next sampling period arrives, the updated vehicle status and other vehicle information are acquired again, and steps S110 to S140 are repeated to form a closed-loop rolling optimization process.

[0092] Through the above mechanism, this embodiment can continuously respond to new interactive situations in dynamic traffic environments, adjust driving strategies in a timely manner, and ensure the safe and efficient passage of the vehicle in areas without traffic lights.

[0093] This application constructs an accurate discretized prediction model based on the longitudinal dynamics of the vehicle, and combines real-time state information of the vehicle and surrounding vehicles to dynamically identify traffic interaction scenarios and adaptively switch multi-objective optimization functions. While ensuring that rear-end collisions and side collisions are effectively avoided, it also considers vehicle traffic efficiency, energy economy, and passenger comfort. By introducing control input increments as optimization variables and constructing an augmented state-space model, jerk constraints can be directly embedded into the optimization problem, ensuring smooth and continuous generation of control commands that conform to the physical response capabilities of the onboard actuators. The controller adopts a rolling time-domain optimization mechanism, resolving the optimal control sequence based on the latest perception information in each sampling period and applying the initial control quantity to the vehicle actuators, thereby achieving rapid response and closed-loop feedback adjustment to sudden traffic conditions. The entire scheduling process does not rely on a central coordination unit; each intelligent driving vehicle can independently make decisions based on local information. This application enables multiple vehicles to efficiently, orderly, and safely pass through complex intersections without the need for traditional traffic lights, effectively alleviating traffic congestion.

[0094] Figure 6 A schematic diagram of a vehicle dispatching device according to an embodiment of this application is shown. Exemplarily, the vehicle dispatching device includes: The acquisition module 100 is used to acquire the driving status information of the vehicle before it reaches the intersection without traffic lights, and to acquire the relative position and predicted driving path information of other vehicles in the current intersection area.

[0095] The selection module 200 is used to determine the corresponding multi-objective optimization function based on the traffic interaction scenario in which the vehicle is located within each sampling period; wherein, the traffic interaction scenario is determined based on the vehicle's driving status information and the relative positions and predicted driving path information of other vehicles.

[0096] Module 300 is used to construct a longitudinal dynamics model of the vehicle based on the vehicle's driving state information, and to construct a model predictive control problem by combining a multi-objective optimization function and safety constraints.

[0097] The control module 400 is used to solve the model predictive control problem, obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

[0098] It is understood that the device in this embodiment corresponds to the vehicle scheduling method in the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0099] This application also provides a vehicle, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor, by running the computer program, causes the vehicle to perform the functions of the various modules in the above-described vehicle scheduling method or vehicle scheduling device.

[0100] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0101] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving execution instructions.

[0102] This application also provides a computer-readable storage medium for storing the computer program used in the aforementioned vehicle. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0104] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0105] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0106] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A vehicle dispatching method, characterized in that, include: Obtain the driving status information of the vehicle before it reaches the intersection without traffic lights, and obtain the relative position and predicted driving path information of other vehicles in the current intersection area; Within each sampling period, a corresponding multi-objective optimization function is determined based on the traffic interaction scenario in which the vehicle is located; wherein, the traffic interaction scenario is determined based on the driving state information of the vehicle and the relative positions and predicted driving path information of other vehicles; Based on the vehicle's driving state information, a longitudinal dynamics model of the vehicle is constructed, and a model predictive control problem is constructed by combining the multi-objective optimization function and safety constraints. Solve the model predictive control problem to obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

2. The vehicle dispatching method according to claim 1, characterized in that, The construction of the vehicle longitudinal dynamics model based on the vehicle's driving state information includes: Based on the nonlinear continuous dynamic equation of the vehicle's longitudinal motion, local linearization is performed at the initial state of the current sampling period to obtain a linearized state-space expression. The state-space expression is discretized to obtain a discrete-time state-space model; The ground rolling friction resistance term is incorporated into the control input variable to form an equivalent control input quantity. It is assumed that the system state transition matrix and the control input matrix remain unchanged in the prediction time domain, so as to construct the vehicle longitudinal dynamics model with displacement and velocity as state variables and the equivalent control input quantity as control variables.

3. The vehicle dispatching method according to claim 1, characterized in that, Based on the vehicle longitudinal dynamics model, and combined with the multi-objective optimization function and the safety constraints, a model predictive control problem is constructed, including: Based on the longitudinal dynamics model, the predicted state sequence of the vehicle in the prediction time domain is derived; The security constraints are transformed into a set of inequality constraints relating to the state prediction sequence and the control sequence; Based on the multi-objective optimization function and the set of inequality constraints, the model predictive control problem is constructed.

4. The vehicle dispatching method according to claim 1, characterized in that, The traffic interaction scenarios include conflict-free scenarios, linear conflict scenarios, point-based conflict scenarios, and complex conflict scenarios: The conflict-free scenario refers to a risk scenario in which there is no potential conflict between the vehicle and other vehicles. The linear conflict scenario refers to a linear conflict scenario in which the vehicle is at risk of rear-end collision with other vehicles in the same entrance lane. The point-like conflict scenario refers to a point-like conflict scenario in which the vehicle and other vehicles from different entrance lanes have intersecting paths within the intersection. The composite conflict scenario is a composite conflict scenario that simultaneously includes linear conflict risk and point-based conflict risk.

5. The vehicle dispatching method according to claim 4, characterized in that, The step of determining the corresponding multi-objective optimization function based on the traffic interaction scenario includes: When the vehicle is in the conflict-free scenario, a first-class multi-objective optimization function is adopted, with traffic efficiency index, energy economy index and ride comfort index as optimization sub-objectives. When the vehicle is in the linear conflict scenario, a second type of multi-objective optimization function is adopted, with the line safety performance index as the main optimization sub-objective and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives. When the vehicle is in the point-based conflict scenario, a third type of multi-objective optimization function is adopted, which takes the conflict point safety performance index as the main optimization sub-objective and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives. When the vehicle is in the complex conflict scenario, a fourth type of multi-objective optimization function is adopted, which takes the route safety performance index and the conflict point safety performance index as the main optimization sub-objectives, and the traffic efficiency index, the energy economy index and the ride comfort index as auxiliary optimization sub-objectives.

6. The vehicle dispatching method according to claim 1, characterized in that, The safety constraints include the minimum safe time interval constraint at the conflict point, the minimum safe distance constraint in the same lane, the physical feasible region constraint of vehicle speed, the physical feasible region constraint of acceleration, and the physical feasible region constraint of jerk.

7. The vehicle dispatching method according to claim 2, characterized in that, The equivalent control input is the difference between the actual acceleration and the deceleration caused by rolling friction.

8. A vehicle dispatching device, characterized in that, include: The acquisition module is used to acquire the driving status information of the vehicle before it reaches the intersection without traffic lights, and to acquire the relative position and predicted driving path information of other vehicles in the current intersection area. The selection module is used to determine the corresponding multi-objective optimization function based on the traffic interaction scenario in which the vehicle is located within each sampling period; wherein, the traffic interaction scenario is determined based on the driving state information of the vehicle and the relative positions and predicted driving path information of other vehicles. The construction module is used to construct a longitudinal dynamics model of the vehicle based on the driving state information of the vehicle, and to construct a model predictive control problem by combining the multi-objective optimization function and safety constraints. The control module is used to solve the model predictive control problem, obtain the target control input sequence in the control time domain, and output the first control input quantity in the target control input sequence to the actuator of the vehicle to realize real-time control of the vehicle.

9. A vehicle, characterized in that, The vehicle includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the vehicle scheduling method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed on a processor, implements the vehicle scheduling method according to any one of claims 1-7.

Images