A hierarchical speed control method for connected vehicles at intersections

Abstract

This invention provides a hierarchical speed control method for connected vehicles at intersections, comprising the following steps: Step S1, establishing a connected vehicle speed control model at the intersection; Step S2, utilizing V2X technology, the connected vehicle obtains in real-time information on its relative distance to the nearest intersection and the traffic light timing information of the nearest intersection; Step S3, constructing an upper-level controller; using a hybrid model predictive control algorithm to generate the optimal speed for the upper-level controller; Step S4, constructing a lower-level controller, building a following model, tracking the optimal speed of the upper-level controller, and generating a globally optimal speed based on safe speed constraints. This invention achieves energy conservation and emission reduction, improves fuel economy, and reduces carbon and nitrogen oxide emissions without braking.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a hierarchical speed control method for connected vehicles at intersections. Background Technology

[0002] With the improvement of productivity, the rapid increase in car ownership has brought about a series of traffic hazards and energy problems. The most complex driving behaviors in the traffic environment often occur at intersections. Intersections are characterized by complex information interference, requiring drivers to make complex judgments when navigating them. This leads to a high frequency of traffic accidents and poor energy efficiency at intersections. Studies have shown that the complex driving behaviors of vehicles at intersections increase fuel consumption and pollutant emissions. Therefore, utilizing intelligent transportation technologies to achieve energy conservation and emission reduction is of great significance.

[0003] Based on V2X (Vehicle to Everything) technology, connected vehicles can interact with surrounding vehicles and road infrastructure, providing real-time status information about the vehicles and their surroundings. Existing intersection speed control methods often only consider the simple operating conditions of a single intersection, or only consider combining traffic flow information from a fixed road segment to obtain a fixed timing strategy. Because this control method does not consider the influence of continuously signalized intersections and upstream queues, it is often only suitable for simple intersections with simple structures and low traffic flow density. Summary of the Invention

[0004] This invention provides a hierarchical speed control method for connected vehicles at intersections, which aims to solve the technical problem that existing speed control methods for intersections in the background art are only applicable to simple intersections with simple structures and low traffic flow density.

[0005] To achieve the above objectives, the present invention provides a hierarchical speed control method for connected vehicles at intersections, comprising the following steps:

[0006] Step S1: Establish a speed control model for connected vehicles at intersections; specifically including:

[0007] Step S11: Establish intersection queue model: Establish a queue model between connected vehicle speed, relative distance at intersection, and intersection traffic light timing;

[0008] Step S12: Establish a fuel consumption model: Establish a fuel consumption model between vehicle speed, acceleration and fuel consumption of connected vehicles;

[0009] Step S13: Establish a pollutant emission model: Establish a pollutant emission model between the speed of connected vehicles and the amount of pollutants emitted;

[0010] Step S2: Using V2X technology, the connected vehicle obtains in real time the relative distance information between itself and the nearest intersection ahead, as well as the traffic light timing information of the nearest intersection.

[0011] Step S3: Construct the upper-layer controller, specifically including:

[0012] Step S31: Using the hybrid model predictive control method, the relative distance between the vehicle itself and the nearest intersection traffic light is transformed into a finite dimension through the intersection queue model in step S11.

[0013] Step S32: Construct a numerical solution model for the optimal vehicle speed within each distance interval: Employ a multi-shot algorithm to calculate the optimal solution for the connected vehicle speed within each distance interval; specifically including:

[0014] Step S321: Construct a nonlinear programming problem based on distance intervals, and establish an objective function by considering fuel economy and driving speed through the fuel consumption model in step S12;

[0015] Step S322: Under the condition of satisfying the constraints, solve for the feasible vehicle speed v in each dimension interval. r Step S323: with v r To find the desired vehicle speed, the optimal feasible vehicle speed is solved by rolling optimization in the prediction time domain.

[0016] Step S324: Using the pollutant emission model in step S13, establish a function with fuel economy and vehicle speed stability as objectives;

[0017] Step S325: Under the condition of satisfying the constraints, solve for the optimal feasible vehicle speed v of the stable upper-level controller. dh .

[0018] Preferably, the method further includes the following steps:

[0019] Step S4: Construct the lower-level controller, specifically including:

[0020] Step S41: Construct a vehicle-following model with the goal of safe driving, and track the optimal feasible speed of the upper-level controller;

[0021] Step S42: Based on the safe speed constraint, considering the constraints of the connected vehicle's position, the position of the vehicle in front, and the safe distance, establish a linear programming model to calculate and generate the globally optimal speed v. d This enables vehicle speed planning.

[0022] Preferably, step S11 specifically includes:

[0023]

[0024] Where, d length Q is the length of the queue at the intersection, in meters; up P represents the traffic flow throughput upstream of the intersection, expressed in veh / h. up Upstream traffic flow density, in veh / km; Q down P represents the downstream traffic flow throughput at the intersection, expressed in veh / h. jam Blockage density, expressed in veh / km; s light x is the location of the traffic light, in meters (m); cav V represents the position of the controlled vehicle, in meters (m); v0 represents the initial speed of the controlled vehicle, in km / h; T red T represents the red light duration, measured in seconds (s). green The green light duration is in seconds (s); t represents the current time (s); P down Downstream traffic flow density, in veh / km;

[0025] Step S12 specifically involves:

[0026]

[0027] Where CostA is fuel consumption in g / km; α, β and γ are constant coefficients; m is vehicle mass in kg; v is instantaneous vehicle speed in m / s; a is instantaneous vehicle acceleration in m / s²; and a, b and c are constant coefficients.

[0028] Step S13 specifically involves:

[0029] CostB=δv+εCostA

[0030] Where CostB is the pollutant emission amount, in g / km; δ and ε are constant coefficients; and v is the instantaneous speed of the vehicle, in m / s.

[0031] Preferably, obtaining the relative distance information in step S2 specifically involves:

[0032] L l =s light -x cav -d length

[0033] Among them, L l The relative distance between the connected vehicle and the nearest intersection, in meters (m); s light The location of the nearest traffic light at the intersection is shown in meters (m); x cav Location of connected vehicles, in meters (m); d length This is the length of the waiting queue at the nearest intersection, in meters (m).

[0034] Preferably, step S3 specifically comprises:

[0035] Step S31: Using the hybrid model predictive control method, the relative distance between the vehicle itself and the nearest intersection signal light is transformed into a finite dimension by using the hybrid prediction of the n intersection queue models of the n intersections in step S11.

[0036] L l Transformed into n finite dimensions: [l0,l1],[l1,l2]...[l n-1 ,l n ], where 0 = l0 <l1<...<l n =L l ;

[0037] Step S32: Construct a numerical solution model for the optimal vehicle speed within each distance interval:

[0038]

[0039] Among them, a cav (l m ), u lm This is the system input for the m-th distance interval, specifically the connected vehicle acceleration in the m-th distance interval; x lm ,v cav (lm) represents the system state variable for the m-th distance interval, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval; f is the numerical solution model for the optimal solution. Let m be the system state quantity of the (m+1)th distance interval, that is, the instantaneous speed of the connected vehicle in the (m+1)th distance interval; m is the mth dimension among the n dimensions of Ll.

[0040] A multi-shot algorithm is used to calculate the optimal solution for the speed of the connected vehicle in each distance interval;

[0041] Step S321: Construct a nonlinear programming problem based on distance intervals, and establish an objective function considering fuel economy and driving speed; specifically:

[0042]

[0043] Where JM is the objective function considering fuel economy and driving speed; xl0 is the initial state quantity of the system in the first dimension, i.e., the initial speed of the connected vehicle, in m / s; xlm is the system state quantity in the m-th dimension, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval, in m / s; v is the instantaneous speed, in m / s; ω is the weighting coefficient; a, b, and c are constant coefficients. To perform distance integration on connected vehicles from the ml-th dimension to the m-th dimension, dl is the differential of the distance;

[0044] Step S322: Under the condition of satisfying the constraints, solve for the feasible vehicle speed vr in each dimension interval;

[0045] Step S323: with v r To find the desired vehicle speed, the optimal feasible vehicle speed is calculated by rolling optimization within the prediction time domain set in the calculation. Specifically, in the prediction time domain Np at time k, the optimal feasible vehicle speed in (k+1, k+2, ..., k+Np) is calculated by rolling optimization.

[0046] Step S324: Establish a function with fuel economy and vehicle speed stability as objectives; specifically:

[0047]

[0048] Among them, J H The objective function for rolling optimization; N p N represents the prediction time domain. c Represents the control time domain; k+i|k represents predicting the value at time k+i using the value at time k, v cav Represents instantaneous vehicle speed, v r The feasible vehicle speed v represents the range of each dimension. r ;

[0049] Step S325: Under the condition of satisfying the constraints, solve for the optimal feasible vehicle speed v of the stable upper-level controller. dh Specifically:

[0050]

[0051] Among them, v min and v max These are the minimum speed and the maximum speed, respectively; a min and a ma x represents the minimum and maximum accelerations, respectively; v dh The optimal feasible vehicle speed generated by the upper-level controller, v cav Represents instantaneous vehicle speed, a cav Represents instantaneous acceleration; u H These are the control variables of the upper-level controller system;

[0052] Preferably, step S323 specifically includes the following steps:

[0053] Step S3231: Construct a longitudinal dynamic model:

[0054]

[0055] Wherein, represents the relative distance and speed of the connected vehicle to the nearest traffic light ahead, in meters; Let k represent the relative distance between the connected vehicle and the nearest traffic light ahead. Let T be the instantaneous speed of the connected vehicle at time k. H Sampling time, in seconds; u H For system control input;

[0056] Step S3232: The scrolling optimization process is as follows:

[0057]

[0058] Among them, X H The upper-level controller state matrix is ​​defined by variable L. l With v cav Composition; Y H A is the state output matrix; H B H C H X is the coefficient matrix; H (k+1) represents the state variable value at time k+1 within the state matrix.

[0059] Preferably, step S4 specifically comprises:

[0060] Step S4: Construct the lower-level controller, specifically including:

[0061] Step S41: Construct a vehicle-following model with the goal of safe driving, and track the optimal feasible speed of the upper-level controller;

[0062] The following vehicle model is:

[0063]

[0064] Where, x cav (t) represents the location of the connected vehicle at time t, in meters (m); Ts represents the sampling time, in seconds (s); t+T represents the time t+T, in seconds; v cav (t) represents the speed of the connected vehicle at time t, in m / s; T D The delay between the upper-level controller and the lower-level controller; uL(t) is the system control input at time t;

[0065]

[0066] X L Y is the state matrix; L A is the state output matrix; L B L C L H and Z are coefficient matrices; v dl The optimal vehicle speed generated for the lower-level controller; S p This represents the drivable distance; v dlThe optimal vehicle speed, calculated for the lower-level controller, is expressed in km / h; L (k) represents the system control variable of the lower-level controller. The control quantity of the system at time k in the driving task is the optimal vehicle speed vdl of the lower-level controller; XL(k) is the state variable value at time k in the state matrix;

[0067] Step S42: Based on the safe speed constraint, generate the globally optimal speed v. d To achieve vehicle speed planning; specifically:

[0068]

[0069] Where, x cav and x pre These represent the positions of the connected vehicle and the vehicle ahead, respectively, in meters (m); TH represents the safe distance, in seconds (s); S f For a safe distance, S p v represents the drivable distance. dl For the optimal vehicle speed of the lower-level controller, v safe For safe driving speed.

[0070] This invention provides a hierarchical speed control method for connected vehicles at intersections, achieving the goal of reducing fuel consumption and pollutant emissions without braking. Vehicles acquire real-time timing information of traffic lights and intersection queue information via V2X technology, using these as constraints for the control method. The upper-level controller is a real-time speed planning layer that uses the acquired traffic light timing information and the vehicle's own state information to perform rolling optimization within various distance intervals using a hybrid model predictive control algorithm to calculate the optimal feasible speed to avoid braking. The lower-level controller is a speed planning layer based on safe distance constraints. Using model predictive control as a framework, within an established following model, it tracks the optimal feasible speed of the upper-level controller based on safe speed constraints to generate the optimal speed for the controlled vehicle. Attached Figure Description

[0071] Figure 1 This is a flowchart of a hierarchical speed control method for connected vehicles at intersections according to the present invention.

[0072] Figure 2 This is a control structure block diagram of a preferred embodiment of a hierarchical speed control method for connected vehicles at intersections according to the present invention.

[0073] Figure 3 This is a schematic diagram of a vehicle-following model of a preferred embodiment of a hierarchical speed control method for connected vehicles at intersections according to the present invention.

[0074] Figures 4(1) and 4(2) are simulation results of a preferred embodiment of the hierarchical speed control method for connected vehicles at intersections according to the present invention.

[0075] Figures 5(a) and 5(b) are real-vehicle test diagrams of a hierarchical speed control method for connected vehicles at intersections according to the present invention. Detailed Implementation

[0076] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0077] This invention addresses existing problems by providing a hierarchical speed control method for connected vehicles at intersections. Since traffic lights are fixed in location, V2X technology is used to obtain real-time information on the relative distance between the vehicle and the nearest intersection ahead, as well as the status of the traffic lights. Based on the wave model of the signalized intersection, the queue length is determined. The upper-level controller employs a hybrid model prediction method to generate the optimal feasible speed v based on the collected traffic information. dh This speed is then used as a reference speed input to the lower-level controller. The lower-level controller establishes a spatial following model and, based on the model prediction framework, generates the optimal speed and safe speed constraints, ultimately generating the globally optimal speed v. d This enables global vehicle speed planning and control.

[0078] Examples of embodiments of the present invention Figure 1 As shown in Figures 5(a) and 5(b), the procedure includes the following steps:

[0079] Step S1: Establish a speed control model for connected vehicles at intersections; specifically including:

[0080] Step S11: Establish an intersection queue model: Establish the speed of connected vehicles and the relative distance to the intersection.

[0081] and the queue model between the timing of traffic lights at intersections; step S11 specifically includes:

[0082]

[0083] Where, d length Q is the length of the queue at the intersection, in meters; up P represents the traffic flow throughput upstream of the intersection, expressed in veh / h. up Upstream traffic flow density, in veh / km; Q down P represents the downstream traffic flow throughput at the intersection, expressed in veh / h. jam Blockage density, expressed in veh / km; s light x is the location of the traffic light, in meters (m); cavV represents the position of the controlled vehicle, in meters (m); v0 represents the initial speed of the controlled vehicle, in km / h; T red T represents the red light duration, measured in seconds (s). green The green light duration is in seconds (s); t represents the current time (s); P down Downstream traffic flow density, in veh / km;

[0084] Step S12: Establish a fuel consumption model: Establish a fuel consumption model relating vehicle speed, acceleration, and fuel consumption; Step S12 specifically involves:

[0085]

[0086] Where CostA is fuel consumption in g / km; α, β and γ are constant coefficients; m is vehicle mass in kg; v is instantaneous vehicle speed in m / s; a is instantaneous vehicle acceleration in m / s²; and a, b and c are constant coefficients.

[0087] Step S13: Establish a pollutant emission model: Establish a pollutant emission model between the speed of connected vehicles and the amount of pollutants emitted; Step S13 specifically includes:

[0088] CostB=δv+εCostA

[0089] Where CostB is the pollutant emission amount, in g / km; δ and ε are constant coefficients; and v is the instantaneous speed of the vehicle, in m / s.

[0090] Step S2: Using V2X technology, the connected vehicle obtains in real time the relative distance information between itself and the nearest intersection ahead, as well as the traffic light timing information of the nearest intersection; specifically, obtaining the relative distance information in step S2 involves:

[0091] L l =s light -x cav -d length

[0092] Among them, L l The relative distance between the connected vehicle and the nearest intersection, in meters (m); s light The location of the nearest traffic light at the intersection is shown in meters (m); x cav Location of connected vehicles, in meters (m); d length This is the length of the waiting queue at the nearest intersection, in meters (m).

[0093] Step S3: Construct the upper-layer controller, specifically including:

[0094] Step S31: Using the hybrid model predictive control method, the relative distance between the vehicle itself and the nearest intersection traffic light is transformed into a finite dimension through the intersection queue model in step S11.

[0095] The hybrid model predictive control method specifically involves using the hybrid prediction of n intersection queue models in step S11 to transform the relative distance between the vehicle itself and the nearest intersection traffic light, considering the intersection queue, into a finite dimension.

[0096] L l Transformed into n finite dimensions: [l0,l1],[l1,l2]...[l n-1 ,l n ], where 0 = l0 <l1<...<l n =L l ;

[0097] Step S32: Construct a numerical solution model for the optimal vehicle speed within each distance interval:

[0098]

[0099] This is a solution model, where u represents the input and x represents the state variables.

[0100] Among them, a cav (l m ), u lm This is the system input for the m-th distance interval, specifically the connected vehicle acceleration in the m-th distance interval; x lm ,v cav (lm) represents the system state variable for the m-th distance interval, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval; f is the numerical solution model for the optimal solution. Let m be the system state quantity of the (m+1)th distance interval, that is, the instantaneous speed of the connected vehicle in the (m+1)th distance interval; m is the mth dimension among the n dimensions of Ll.

[0101] A multi-shot algorithm is used to calculate the optimal solution for the connected vehicle speed within each distance interval; specifically including:

[0102] Step S321: Construct a nonlinear programming problem based on distance intervals. Using the fuel consumption model from step S12, establish an objective function considering fuel economy and driving speed; specifically:

[0103]

[0104] Where JM is the objective function considering fuel economy and driving speed; xl0 is the initial state quantity of the system in the first dimension, i.e., the initial speed of the connected vehicle, in m / s; xlm is the system state quantity in the m-th dimension, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval, in m / s; v is the instantaneous speed, in m / s; ω is the weighting coefficient; a, b, and c are constant coefficients. To perform distance integration on connected vehicles from the ml-th dimension to the m-th dimension, dl is the differential of the distance;

[0105] Step S322: Under the condition of satisfying the constraints, solve for the feasible vehicle speed v in each dimension interval. r ;

[0106] Step S323: with v r To find the desired vehicle speed, a rolling optimization method is used in the prediction time domain to solve for the optimal feasible vehicle speed; specifically: using v r To determine the desired vehicle speed, the optimal feasible speed is solved using rolling optimization within the prediction time domain set in the computational settings; specifically, the prediction time domain N at time k. p In the process, rolling calculations are performed on (k+1,k+2,…,k+N). p The optimal feasible speed;

[0107] Step S323 specifically includes the following steps:

[0108] Step S3231: Construct a longitudinal dynamic model:

[0109]

[0110] Wherein, represents the relative distance and speed of the connected vehicle to the nearest traffic light ahead, in meters; Let k represent the relative distance between the connected vehicle and the nearest traffic light ahead. Let T be the instantaneous speed of the connected vehicle at time k. H Sampling time, in seconds; u H For system control input;

[0111] Step S3232: The scrolling optimization process is as follows:

[0112]

[0113] Among them, X H The upper-level controller state matrix is ​​defined by variable L. l With v cav Composition; Y H A is the state output matrix; H B H C H X is the coefficient matrix; H(k+1) represents the state variable value at time k+1 within the state matrix.

[0114] Step S324: Using the pollutant emission model from step S13, establish a function with fuel economy and vehicle speed stability as objectives; specifically:

[0115]

[0116] Among them, J H The objective function for rolling optimization; Np represents the prediction time domain, Nc represents the control time domain; k+i|k represents predicting the value at time k+i using the value at time k, v cav Represents instantaneous vehicle speed, v r The feasible vehicle speed v represents the range of each dimension. r ;

[0117] Step S325: Under the condition of satisfying the constraints, solve for the optimal feasible vehicle speed v of the stable upper-level controller. dh Specifically:

[0118]

[0119] Among them, v min and v max These are the minimum speed and the maximum speed, respectively; a min and a ma x represents the minimum and maximum accelerations, respectively; v dh The optimal feasible vehicle speed generated by the upper-level controller, v cav Represents instantaneous vehicle speed, a cav Represents instantaneous acceleration; u H These are the control variables of the upper-level controller system;

[0120] Unlike previous methods that relied solely on numerical solutions or model predictive control (MDC), the hybrid model predictive control (HMC) approach is a fusion method. First, a multi-shot algorithm is used to generate the optimal velocity trajectory within each distance interval, followed by rolling optimization within a model prediction framework. The hybrid MDC approach overcomes the computational burden of traditional MDC while retaining its predictive characteristics, resulting in calculations that better match the control objectives than traditional numerical solutions.

[0121] The constructed upper-level controller adopts a hybrid model predictive control method. Unlike previous methods that only use numerical solutions or model predictive control, the hybrid model predictive control method is a fusion approach. First, a multi-shot algorithm is used to generate the optimal velocity trajectory within each distance interval, and then rolling optimization is performed within the model prediction framework. The hybrid model predictive control method overcomes the disadvantage of large computational load in traditional model predictive control, while retaining the predictive characteristics of the model predictive control method. The calculation results are more consistent with the control objective than those of traditional numerical solutions, as shown in Figures 4(1) and 4(2).

[0122] Step S4: Construct the lower-level controller, specifically including:

[0123] Step S41: Construct a following model with the goal of safe driving, such as... Figure 3 As shown, the optimal feasible vehicle speed is tracked by the upper-level controller; the following vehicle model is:

[0124]

[0125] Where, x cav (t) represents the location of the connected vehicle at time t, in meters (m); Ts represents the sampling time, in seconds (s); t+T represents the time t+T, in seconds; v cav (t) represents the speed of the connected vehicle at time t, in m / s; T D The delay between the upper-level controller and the lower-level controller; uL(t) is the system control input at time t;

[0126] A model predictive control framework is used to track the globally optimal feasible vehicle speed of the upper-level controller:

[0127]

[0128] X L Y is the state matrix; L A is the state output matrix; L B L C L H and Z are coefficient matrices; v dl The optimal vehicle speed generated for the lower-level controller; S p This represents the drivable distance; v dl The optimal vehicle speed, calculated for the lower-level controller, is expressed in km / h; L (k) represents the system control variable of the lower-level controller. The control quantity of the system at time k in the driving task is the optimal vehicle speed vdl of the lower-level controller; XL(k) is the state variable value at time k in the state matrix.

[0129] S p Shown Figure 2 In the middle, V is a variable. dl and uL (k) are essentially the same, u L The result of the algorithm calculation is a control variable. In a driving task, this is the optimal vehicle speed v of the lower-level controller. dl .

[0130] Step S42: Based on the safe speed constraint, considering the constraints of the connected vehicle's position, the position of the vehicle in front, and the safe distance, establish a linear programming model to calculate and generate the globally optimal speed v. d To achieve vehicle speed planning; specifically:

[0131]

[0132] Where, x cav and x pre These represent the positions of the connected vehicle and the vehicle ahead, respectively, in meters (m); TH represents the safe distance, in seconds (s); S f For a safe distance, S p v represents the drivable distance. dl For the optimal vehicle speed of the lower-level controller, v safe For safe driving speed.

[0133] By combining upper-level and lower-level controllers, the globally optimal vehicle speed can be obtained in real time to improve fuel efficiency, reduce pollutant emissions, and facilitate rapid passage, thus solving the problem of optimal vehicle speed control for connected vehicles when passing through continuous signalized intersections.

[0134] This invention provides a hierarchical speed control method for connected vehicles at intersections, achieving the goal of reducing fuel consumption and pollutant emissions without braking. Vehicles acquire real-time timing information of traffic lights and intersection queue information via V2X technology, using these as constraints for the control method. The upper-level controller is a real-time speed planning layer that uses the acquired traffic light timing information and the vehicle's own state information to perform rolling optimization within various distance intervals using a hybrid model predictive control algorithm to calculate the optimal feasible speed to avoid braking. The lower-level controller is a speed planning layer based on safe distance constraints. Using model predictive control as a framework, within an established following model, it tracks the optimal feasible speed of the upper-level controller based on safe speed constraints to generate the optimal speed for the controlled vehicle.

[0135] The technical principle of this invention is to construct an upper-level controller and a lower-level controller, and propose a hierarchical speed control method. The upper-level controller constructs a function with fuel economy, driving speed and vehicle speed stability as objectives, and transforms the control problem into a finite-dimensional problem in the distance domain through a hybrid model predictive control method. The lower-level controller constructs a following model with safe driving as the objective, and tracks the optimal feasible speed of the upper-level controller based on the safe vehicle speed to generate the globally optimal vehicle speed, so as to improve traffic efficiency and increase fuel utilization.

[0136] Distributing different driving tasks across two controllers reduces computation time. The upper and lower controllers can be configured to meet various application needs. For example, a centralized mode is used when connectivity is low, with the upper controller positioned on a central cloud server at the intersection, and vehicles receiving instructions from the cloud server. A standalone mode is used when connectivity is high, with the upper controller positioned in the central unit of each vehicle, allowing each connected car to independently achieve its control objectives. Furthermore, the hierarchical control method can also be configured with a hybrid central + standalone mode, providing a new approach for controlling connected vehicles at continuous signalized intersections.

[0137] The advantages of this invention are:

[0138] This method employs a hierarchical speed optimization design. The upper-level controller calculates the optimal feasible speed to avoid braking at a red light. The lower-level controller considers the influence of vehicles ahead and, combined with safety speed constraints, generates the optimal speed that satisfies a safe distance. This strategy achieves safe driving.

[0139] To address the complexity of solving problems under multiple constraints in real-world road networks, a speed control method for connected vehicles based on continuous signalized intersections is proposed for a connected vehicle environment. This method combines the multi-shot algorithm with traditional model predictive control (MDC). While retaining the characteristics of MDC, the multi-shot algorithm transforms the problem into a finite-dimensional solution, overcoming the computational burden of traditional MDC.

[0140] This invention fully considers driving needs under different driving conditions and is simple to implement. Based on different task objectives, the hierarchical control method provides a new approach for the control of intelligent connected vehicles at continuous signalized intersections.

[0141] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A hierarchical speed control method for connected vehicles at intersections, characterized in that, Includes the following steps: Step S1: Establish a speed control model for connected vehicles at intersections; specifically including: Step S11: Establish intersection queue model: Establish a queue model between connected vehicle speed, relative distance at intersection, and intersection traffic light timing; Step S12: Establish a fuel consumption model: Establish a fuel consumption model between vehicle speed, acceleration and fuel consumption of connected vehicles; Step S13: Establish a pollutant emission model: Establish a pollutant emission model between the speed of connected vehicles and the amount of pollutants emitted; Step S2: Using V2X technology, the connected vehicle obtains in real time the relative distance information between itself and the nearest intersection ahead, as well as the traffic light timing information of the nearest intersection. Step S3: Construct the upper-layer controller, specifically including: Step S31: Using the hybrid model predictive control method, the relative distance between the vehicle itself and the nearest intersection traffic light is transformed into a finite dimension through the intersection queue model in step S11. Step S32: Construct a numerical solution model for the optimal vehicle speed within each distance interval: Employ a multi-shot algorithm to calculate the optimal solution for the connected vehicle speed within each distance interval; specifically including: Step S321: Construct a nonlinear programming problem based on distance intervals, and establish an objective function by considering fuel economy and driving speed through the fuel consumption model in step S12; Step S322: Solve the feasible vehicle speed v of each dimension interval under the condition of meeting the constraints r ; Step S323: Rolling optimization is solved for the optimal feasible vehicle speed in the prediction horizon with v r as the desired vehicle speed. Step S324: Using the pollutant emission model in step S13, establish a function with fuel economy and vehicle speed stability as objectives; Step S325: under the condition of meeting the constraints, the stable upper controller optimal feasible vehicle speed v is obtained by solving dh .

2. The hierarchical speed control method for connected vehicles at an intersection according to claim 1, characterized in that, It also includes the following steps: Step S4: Construct the lower-level controller, specifically including: Step S41: Construct a vehicle-following model with the goal of safe driving, and track the optimal feasible speed of the upper-level controller; Step S42: Based on the safe speed constraint, considering the constraints of the connected vehicle's position, the position of the vehicle in front, and the safe distance, establish a linear programming model to calculate and generate the globally optimal speed v. d This enables vehicle speed planning.

3. The hierarchical speed control method for connected vehicles at an intersection according to claim 2, characterized in that, Step S11 specifically involves: Where, d length Q is the length of the queue at the intersection, in meters; up P represents the traffic flow throughput upstream of the intersection, expressed in veh / h. up Upstream traffic flow density, in veh / km; Q down P represents the downstream traffic flow throughput at the intersection, expressed in veh / h. jam Blockage density, expressed in veh / km; s light x is the location of the traffic light, in meters (m); cav V represents the position of the controlled vehicle, in meters (m); v0 represents the initial speed of the controlled vehicle, in km / h; T red T represents the red light duration, measured in seconds (s). green The green light duration is in seconds (s); t represents the current time (s); P down Downstream traffic flow density, in veh / km; Step S12 specifically involves: Where CostA is fuel consumption in g / km; α, β and γ are constant coefficients; m is vehicle mass in kg; v is instantaneous vehicle speed in m / s; a is instantaneous vehicle acceleration in m / s²; k, b and c are constant coefficients. Step S13 specifically involves: Where CostB is the pollutant emission amount, in g / km; δ and ε are constant coefficients; and v is the instantaneous speed of the vehicle, in m / s.

4. The hierarchical speed control method for connected vehicles at an intersection according to claim 3, characterized in that, The specific steps for obtaining the relative distance information in step S2 are as follows: Among them, L l The relative distance between the connected vehicle and the nearest intersection, in meters (m); s light The location of the nearest traffic light at the intersection is shown in meters (m); x cav Location of connected vehicles, in meters (m); d length This is the length of the waiting queue at the nearest intersection, in meters (m).

5. A hierarchical speed control method for connected vehicles at an intersection according to claim 4, characterized in that, Step S3 specifically involves: Step S31: Using the hybrid model predictive control method, the relative distance between the vehicle itself and the nearest intersection signal light is transformed into a finite dimension by using the hybrid prediction of the n intersection queue models of the n intersections in step S11. L l Transformed into n finite dimensions: , ... Where, 0 = l0 <l1<...<l n = L l ; Step S32: Construct a numerical solution model for the optimal vehicle speed within each distance interval: Among them, a cav (lm), u lm This is the system input for the m-th distance interval, specifically the connected vehicle acceleration in the m-th distance interval; x lm ,v cav (lm) represents the system state variable for the m-th distance interval, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval; f is the numerical solution model for the optimal solution; Let m be the system state quantity of the (m+1)th distance interval, that is, the instantaneous speed of the connected vehicle in the (m+1)th distance interval; m is the mth dimension among the n dimensions of Ll. A multi-shot algorithm is used to calculate the optimal solution for the speed of the connected vehicle in each distance interval; Step S321: Construct a nonlinear programming problem based on distance intervals, and establish an objective function considering fuel economy and driving speed; specifically: Where JM is the objective function considering fuel economy and driving speed; x l0 x represents the initial state quantity of the system in the first dimension, i.e., the initial speed of the connected vehicle, in m / s; lm ω represents the system state quantity in the m-th dimension, i.e., the instantaneous speed of the connected vehicle in the m-th distance interval, in m / s; v represents the instantaneous speed, in m / s; ω represents the weighting coefficient; a, b, and c are constant coefficients. To perform distance integration on connected vehicles from the ml-th dimension to the m-th dimension, dl is the differential of the distance; Step S322: Under the condition of satisfying the constraints, solve for the feasible vehicle speed v in each dimension interval. r ; Step S323: with v r To find the desired vehicle speed, the optimal feasible vehicle speed is calculated by rolling optimization within the prediction time domain set in the calculation. Specifically, in the prediction time domain Np at time k, the optimal feasible vehicle speed in (k+1, k+2, ..., k+Np) is calculated by rolling optimization. Step S324: Establish a function with fuel economy and vehicle speed stability as objectives; specifically: Among them, J H The objective function for rolling optimization; Np represents the prediction time domain, Nc represents the control time domain; k+i|k represents predicting the value at time k+i using the value at time k, v cav Represents instantaneous vehicle speed, v r The feasible vehicle speed v represents the range of each dimension. r ; Step S325: Under the condition of satisfying the constraints, solve for the optimal feasible vehicle speed v of the stable upper-level controller. dh Specifically: Among them, v min and v max These are the minimum speed and the maximum speed, respectively; a min and a ma x represents the minimum and maximum accelerations, respectively; v dh The optimal feasible vehicle speed generated by the upper-level controller, v cav Represents instantaneous vehicle speed, a cav Represents instantaneous acceleration; u H These are the control variables of the upper-level controller system.

6. The hierarchical speed control method for connected vehicles at an intersection according to claim 5, characterized in that, Step S323 specifically includes the following steps: Step S3231: Construct a longitudinal dynamic model: Wherein, represents the relative distance and speed of the connected vehicle to the nearest traffic light ahead, in meters; Let k represent the relative distance between the connected vehicle and the nearest traffic light ahead. Let T be the instantaneous speed of the connected vehicle at time k. H Sampling time, in seconds; u H For system control input; Step S3232: The scrolling optimization process is as follows: Among them, X H The upper-level controller state matrix is ​​defined by variable L. l With v cav Composition; Y H A is the state output matrix; H B H C H X is the coefficient matrix; H (k+1) represents the state variable value at time k+1 within the state matrix.

7. A hierarchical speed control method for connected vehicles at an intersection according to claim 6, characterized in that, Step S4 specifically involves: Step S4: Construct the lower-level controller, specifically including: Step S41: Construct a vehicle-following model with the goal of safe driving, and track the optimal feasible speed of the upper-level controller; The following vehicle model is: Where, x cav (t) represents the location of the connected vehicle at time t, in meters (m); Ts represents the sampling time, in seconds (s); t+T represents the time t+T, in seconds; v cav (t) represents the speed of the connected vehicle at time t, in m / s; T D The delay between the upper-level controller and the lower-level controller; uL(t) is the system control input at time t; A model predictive control framework is used to track the globally optimal feasible vehicle speed of the upper-level controller: X L Y is the state matrix; L A is the state output matrix; L B L C L H and Z are coefficient matrices; v dl The optimal vehicle speed generated for the lower-level controller; S p This represents the drivable distance; v dl The optimal vehicle speed, calculated for the lower-level controller, is expressed in km / h; L (k) represents the system control variable of the lower-level controller. The control quantity of the system at time k in the driving task is the optimal vehicle speed vdl of the lower-level controller; XL(k) is the state variable value at time k in the state matrix; Step S42: Based on the safe speed constraint, generate the globally optimal speed v. d To achieve vehicle speed planning; specifically: Where, x cav and x pre These represent the positions of the connected vehicle and the vehicle ahead, respectively, in meters (m); TH represents the safe distance, in seconds (s); S f For a safe distance, S p v represents the drivable distance. dl For the optimal vehicle speed of the lower-level controller, v safe For safe driving speed.

Images