A pre-signal based signal control and cav trajectory planning method

Abstract

The application discloses a signal control and CAV trajectory planning method based on a pre-signal lamp, and comprises the following steps: acquiring intersection phase information and first information according to a pre-signal lamp arranged in a signal area; acquiring second information in a regulation area and a special lane area; performing first model processing on the first information and the second information, predicting a vehicle motion trajectory, and acquiring a total number and a type of vehicles in the regulation area; performing second model and third model processing on the intersection phase information, the vehicle motion trajectory, the total number and the type of vehicles, and predicting a best phase duration and a best CAV trajectory. The application significantly improves the traffic capacity of the intersection, obviously reduces the average fuel consumption and the number of lane changes of the CAV, and obviously improves the traffic speed and delay of the heterogeneous traffic flow.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology. Specifically, it relates to a signal control and CAV trajectory planning method based on pre-signal lights. Background Technology

[0002] Pollution emissions at urban intersections are a serious problem, mainly due to the significant increase in fuel consumption caused by frequent acceleration and deceleration of vehicles and prolonged idling. Traffic management has been somewhat effective in addressing this issue.

[0003] From a signal control perspective, optimizing phase sequence and green light duration can reduce vehicle delays, thereby reducing energy consumption. Based on this, existing technologies have found optimal traffic light durations under different traffic conditions using derived average traffic congestion levels. From a vehicle control perspective, eco-driving strategies can effectively reduce fuel consumption and also effectively reduce pollution caused by driver operation. Therefore, existing technologies have established a model of driver acceptance of eco-driving strategies and pointed out the significant role of eco-driving in energy conservation. However, due to the rapid growth in traffic demand, traditional traffic management measures have limited effectiveness in addressing this issue.

[0004] With the rapid development of communication technology, new solutions have been proposed to reduce pollution emissions and travel delays, and new solutions have been provided for signal control and trajectory optimization. However, most existing research is based on pure CAV environments or connected human-driven vehicles (CHVs). In the foreseeable future, human driving behavior will remain an irreplaceable part of transportation, and CHVs are unlikely to become widespread in the short term. Therefore, the integration of CAVs and human-driven vehicles (HVs) urgently needs further research.

[0005] In heterogeneous traffic flows involving both controlled aerial vehicles (CAVs) and heavy vehicle vehicles (HVs), the cumulative effect of random factors caused by human drivers increases the overall chaos in the traffic system. Therefore, CAVs may fail to achieve their control objectives due to interference from HVs, and optimal timing may also be unattainable. The most serious issue is the safety risks posed by HVs. To address this problem, existing technologies propose dedicated CAV lanes to reduce the impact of HVs on the controlled system. However, most existing technologies assume, on the one hand, that HVs also participate in connectivity, and on the other hand, that HVs do not engage in lane-changing behavior—both assumptions are unrealistic. Furthermore, existing technologies suffer from poor vehicle throughput, long delays, and do not consider pollutant emissions. Therefore, there is an urgent need for a signal control and CAV trajectory planning method that improves the traffic system's efficiency and reduces vehicle pollutant emissions in heterogeneous traffic flow scenarios. Summary of the Invention

[0006] This invention is proposed based on the above-mentioned needs of the prior art. The technical problem to be solved by this invention is to provide a signal control and CAV trajectory planning method based on pre-signal lights to improve the communication capability of intersections, reduce vehicle delays, CAV fuel consumption and lane change frequency.

[0007] To solve the above problems, the present invention is implemented using the following technical solution:

[0008] A signal control and CAV trajectory planning method based on pre-signal lights is provided, comprising: acquiring intersection phase information and first information based on pre-signal lights set in the signaling zone, the first information including HV vehicle light semantic information and CAV steering demand information; acquiring second information within the control zone and dedicated lane zone, the second information including the position, speed, acceleration, and lane selection information of all vehicles; performing first model processing on the first information and the second information to predict vehicle motion trajectories, and acquiring the total number and vehicle type of vehicles in the control zone; the vehicle motion trajectory includes the trajectory of a vehicle following the preceding vehicle and the trajectory of a vehicle during lane changing; and planning the intersection... Phase information, vehicle trajectory, total number of vehicles, and vehicle type are processed by a second model and a third model to predict the optimal phase duration and optimal CAV trajectory. This includes: optimizing the phase duration using the second model based on the intersection phase information, total number of vehicles, and vehicle type; optimizing lane-changing strategies and vehicle acceleration using the third model based on the vehicle trajectory and optimized phase duration, with the goal of minimizing CAV travel delay time, CAV fuel consumption, and lane-changing costs; and determining the CAV position information at the next moment based on the optimized lane-changing strategy and vehicle acceleration. The third model then sends the CAV position information to the second model for the next optimization.

[0009] Optionally, the step of obtaining intersection phase information and first information based on the pre-signal lights set in the signaling zone includes: obtaining intersection phase information based on the pre-signal lights set in the signaling zone, wherein the intersection phase information includes phase sequence and phase duration; and obtaining HV vehicle light semantic information and CAV turning demand information in the signaling zone through a light signal detection device based on the intersection phase information.

[0010] Optionally, the first model includes a second-order car following model, the specific process of which includes: obtaining the current state information of the vehicle and the current state information of the corresponding preceding vehicle based on the first information and the second information; calculating the relative position and relative speed between the current vehicle and the preceding vehicle based on the current state information of the vehicle and the corresponding preceding vehicle; predicting the state information of the vehicle at the next moment based on the relative position and relative speed; repeating the second-order car following model processing process to determine the trajectory of the vehicle following the preceding vehicle.

[0011] Optionally, the first model further includes an improved lane-changing model, the specific process of which includes: determining that the vehicle needs to change lanes based on the first information and the second information; calculating the guidance angle based on the vehicle's current position and the target position of the lane change; and determining the vehicle's trajectory during the lane-changing process using a path planning algorithm based on the guidance angle and the vehicle's current state information.

[0012] Optionally, the step of optimizing the phase duration using a second model based on the intersection phase information, the total number of vehicles, and the vehicle type includes: calculating the cycle duration using the Webster optimal cycle algorithm based on the intersection phase information, the total number of vehicles, and the vehicle type; and optimizing the effective green light time of each phase using the traffic flow ratio based on the cycle duration, wherein the effective green light time is between the shortest and longest effective green light times.

[0013] Optionally, the expression for the second model is: G e =C opt -L;L=n×(l+r e，i ); t p =g e，i g min ≤g e，i ≤g max Among them, C opt The optimal signal cycle is given by L, the total intersection loss time is given by Y, and the sum of the intersection traffic flow ratios is given by G. e Let n be the effective green light time for lane e, n be the number of signal phases, l be the signal loss time, and r be the signal loss time. e，i Let y be the red light duration for the i-th phase of lane e. i Let t be the traffic flow ratio of the critical lane in the i-th phase. p g e，i For lane e, the effective green light time for the i-th phase is g. min Indicates the shortest effective green light duration for a phase; g max The longest valid green light duration for a given phase.

[0014] Optionally, the expression for the objective function of minimizing CAV traffic delay time, CAV fuel consumption, and lane-changing cost is as follows: Where ω represents the identifier of CAV, The time when the CAV begins trajectory optimization in the control zone is indicated; h1 is the optimization time of the CAV in the control zone; h2 is the optimization time of the CAV in the dedicated lane zone; δ ω The identifier indicates whether the vehicle has entered the conflict zone; 0 indicates entering, and 1 indicates otherwise. ω This represents the longitudinal acceleration optimization curve. G represents the number of lane changes. ω Let θ1 represent the weighting coefficient of CAV travel delay time, θ2 represent the weighting coefficient of CAV fuel consumption, θ3 represent the weighting coefficient of lane-changing cost, and Δt represent the time interval.

[0015] Optionally, the objective function is constrained, including: constraining the lane-changing sequence according to a first constraint condition; the first constraint condition includes the duration of the continuous lane-changing interval of the CAV being greater than a preset minimum time interval, and the CAV being able to change lanes at most one time; constraining vehicle motion according to a second constraint condition; the second constraint condition includes the vehicle's acceleration being between a preset maximum acceleration and a preset maximum deceleration, and the vehicle's speed being between 0 and a preset vehicle speed limit; and constraining the longitudinal safe distance according to a third constraint condition; the third constraint condition includes the current vehicle maintaining a longitudinal safe distance from both the vehicle in front and the vehicle behind, the longitudinal safe distance being determined by the current vehicle's speed.

[0016] Optionally, the objective function is solved using an improved cuckoo algorithm based on an extreme value optimization algorithm, including: forming a nest from the randomly generated CAV spatial locations of each time period, wherein the spatial location includes lane selection information and vehicle acceleration; determining whether a preset number of searches has been reached, and if not, generating the next spatial location for each nest according to the Levy flight algorithm; optimizing each nest based on the next spatial location using an extreme value optimization algorithm and updating the spatial location of the nest; determining whether the maximum number of iterations has been reached, and if so, obtaining the optimal spatial location.

[0017] A computer-readable storage medium having a computer program stored thereon, the computer-readable storage medium storing a signal control and CAV trajectory planning program based on a pre-signal light, the signal control and CAV trajectory planning program based on the pre-signal light being executed by a processor to implement the signal control and CAV trajectory planning program based on the pre-signal light.

[0018] Compared with existing technologies, this invention proposes a signal control and CAV trajectory planning method based on pre-signal lights. It combines traditional signal timing schemes with CAV trajectory planning for dedicated and approach lanes to manage mixed traffic of HV and CAV vehicles. Based on discrete time, the control system constructs a two-layer optimization model for intersection signal phase duration and CAV trajectory planning, based on the collection of HV vehicle light semantic information, CAV turning demand information, and the position and speed information of all vehicles in the control zone and dedicated lane zone. This invention significantly improves the capacity of intersections, significantly reduces the average fuel consumption and lane change frequency of CAVs, and significantly improves the speed and delay of heterogeneous traffic flows. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings.

[0020] Figure 1 This is a flowchart of a signal control and CAV trajectory planning method based on pre-signal lights, provided by a specific embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of a model framework for a signal control and CAV trajectory planning method based on pre-signal lights, provided by a specific embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of a typical isolated crossroads with a dedicated CAV lane provided in a specific embodiment of the present invention;

[0023] Figures 4A-4D This is a schematic diagram showing the vehicle throughput, delay, and fuel economy of heterogeneous traffic flows under two different signal control strategies provided in this embodiment.

[0024] Figure 5 This is a schematic diagram of the vehicle's average fuel consumption provided in this embodiment;

[0025] Figure 6A This is a schematic diagram of the spatiotemporal trajectory of CAV under two different control strategies provided in this embodiment;

[0026] Figure 6B This is a schematic diagram showing the acceleration changes under two different control strategies provided in this embodiment;

[0027] Figure 7A , Figure 7B and Figure 7C This is a schematic diagram illustrating the average delay of heterogeneous traffic flow, CAV, and HV when the saturation levels are 0.5, 1.0, and 1.25, respectively, provided in this embodiment.

[0028] Figure 7D , Figure 7E and Figure 7F This embodiment illustrates the reduction in average delay for heterogeneous traffic flow, CAV, and HV under different demand levels.

[0029] Figures 8A-8B This is a schematic diagram of HV travel trajectories under different control strategies when CAV penetration is 70%.

[0030] Figure 9 This is a schematic diagram illustrating the positive impact of CAV penetration rate on intersection capacity under different traffic demands, as provided in this embodiment. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] To facilitate understanding of the embodiments of the present invention, further explanations and descriptions will be provided below with reference to the accompanying drawings and specific embodiments. These embodiments do not constitute a limitation on the scope of protection of the present invention.

[0033] Example 1

[0034] This embodiment provides a signal control and CAV trajectory planning method based on pre-signal lights, such as Figures 1-2 As shown, it includes:

[0035] S1: Obtain the intersection phase information and first information based on the pre-signal lights set in the signaling area.

[0036] The intersection phase information includes phase sequence and phase duration; the first information includes HV vehicle headlight semantic information and CAV steering demand information.

[0037] In this embodiment, this step includes: obtaining intersection phase information based on the pre-signal lights set in the signaling zone; and obtaining HV vehicle light semantic information and CAV turning demand information in the signaling zone through a light signal detection device based on the intersection phase information.

[0038] like Figure 3The diagram illustrates the layout of an approach arm at a typical isolated crossroads with dedicated CAV lanes. Each arm accommodates vehicles traveling in three different directions: left turn, straight, and right turn. The dedicated CAV lane is for CAVs traveling straight and turning left. Right-turning CAVs and right-turning HVs are not signal-controlled and coexist in the right-turn lane, sharing the signal phase at the intersection. Each arm is divided into a signaling zone, a control zone, and a dedicated lane zone. A pre-signal is installed at the exit of the signaling zone. The control system announces the intersection phase information in advance at the pre-signal. Guided by the pre-signal, the HV provides turning information within the signaling zone via signal detection equipment before entering the control zone. The CAV, based on the pre-acquired intersection phase information, completes trajectory planning, including lane changing, within this zone before entering the dedicated lane zone. The dedicated lane zone is located between the control zone and the conflict zone. Lane changes are prohibited in this zone, and vehicles in the dedicated lane zone undergo further speed optimization to pass through the intersection without stopping. Based on vehicle information collected from the signal control and dispatch areas, traffic light phases and CAV (Caravan) trajectories are dynamically adjusted to reduce energy consumption and delays. To facilitate dynamic planning of CAV trajectories and traffic light phase durations, phase duration, CAV lateral lane-changing strategies, and longitudinal speed adjustments are optimized within a single framework. This collaborative optimization leverages the connectivity advantages of CAVs, improves signal control quality, and minimizes energy consumption.

[0039] S2: Obtain second information within the control zone and dedicated lane zone.

[0040] The second information includes the position, speed, acceleration, and lane selection information of all vehicles.

[0041] Because CAVs have network connectivity, a large number of sensing and communication devices are deployed in the control area, which can capture the position, speed, acceleration and lane selection information of all vehicles in the area in a timely manner.

[0042] S3: Perform first model processing on the first information and the second information to predict the vehicle trajectory and obtain the total number and type of vehicles in the control area.

[0043] The vehicle trajectory includes the trajectory of the vehicle following the vehicle in front and the trajectory of the vehicle changing lanes.

[0044] The first model includes a second-order car following model. The first information and the second information are processed using the second-order car following model. The specific process includes:

[0045] Step 1: Based on the first and second information, obtain the current status information of the vehicle and the corresponding current status information of the vehicle in front.

[0046] Step 2: Based on the current status information of the vehicle and the corresponding current status information of the vehicle in front, calculate the relative position and relative speed between the current vehicle and the vehicle in front.

[0047] Step 3: Based on the relative position and relative speed, predict the vehicle's state information for the next moment.

[0048] Step 4: Repeat the second-order car following model processing to determine the trajectory of the vehicle following the vehicle in front.

[0049] The first model also includes an improved lane-change model, which processes the first information and the second information using the lane-change model. The specific process includes:

[0050] Step 1: Based on the first and second information, determine whether the vehicle needs to change lanes.

[0051] Step 2: Calculate the guiding angle based on the vehicle's current position and the target position for lane changing.

[0052] Step 3: Based on the guidance angle and the vehicle's current state information, use a path planning algorithm to determine the vehicle's trajectory during the lane change process.

[0053] Vehicle trajectories are predicted using a second-order car following model and an improved lane-changing model. Considering the large number of sensors deployed in the control zone, the total number and type of vehicles within the control zone and dedicated lane zone can be statistically determined, directly impacting the phase duration. Because dedicated lanes are set up for CAVs within the dedicated lane zone, the CAVs directly receive phase duration information. The preceding vehicle will travel along the planned trajectory; therefore, knowing the trajectory information of the vehicle ahead, the CAV will generate a feasible trajectory.

[0054] S4: The intersection phase information, vehicle trajectory, total number of vehicles and vehicle type are processed by the second and third models to predict the optimal phase duration and optimal CAV trajectory.

[0055] The second model includes traffic light phase duration optimization, and the third model includes CAV trajectory planning.

[0056] The second and third models jointly optimize phase duration, CAV lane-changing strategy, and longitudinal acceleration curve. In the second model, the phase duration is optimized in real-time based on the actual total number and type of vehicles entering the control zone. In the third model, lane-changing strategy information and vehicle acceleration optimization curves are optimized to minimize CAV delay time, fuel consumption, and lane-changing costs. Simultaneously, the third model also promptly sends CAV position information to the second model to avoid wasted phase. The two models influence each other to predict the optimal phase duration and optimal CAV trajectory.

[0057] This step includes:

[0058] S400: Based on intersection phase information, total number of vehicles, and vehicle type, the phase duration is optimized using a second model.

[0059] Specifically, including:

[0060] Step 1: Calculate the cycle duration using Webster's optimal cycle algorithm based on the intersection phase information, total number of vehicles, and vehicle type.

[0061] Step 2: Based on the cycle duration, optimize the effective green light time for each phase using the traffic flow ratio, wherein the effective green light time is between the shortest and longest effective green light time.

[0062] The traffic flow is the ratio of actual traffic volume to saturated communication capacity.

[0063] The expression for the second model is:

[0064]

[0065] G e =C opt -L

[0066] L=n×(l+r e，i )

[0067]

[0068] t p =g e，i

[0069] g min ≤g e，i ≤g max

[0070] Among them, C opt The optimal signal cycle is given by L, the total intersection loss time is given by Y, and the sum of the intersection traffic flow ratios is given by G. eLet n be the effective green light time for lane e, n be the number of signal phases, l be the signal loss time, and r be the signal loss time. e，i Let y be the red light duration for the i-th phase of lane e. i Let t be the traffic flow ratio of the critical lane in the i-th phase. p g e，i For lane e, the effective green light time for the i-th phase is g. min Indicates the shortest effective green light duration for a phase; g max The longest valid green light duration for a given phase.

[0071] S410: Based on the vehicle trajectory and the optimized phase duration, the third model optimizes the lane-changing strategy and vehicle acceleration with the goal of minimizing CAV traffic delay time, CAV fuel consumption and lane-changing cost; based on the optimized lane-changing strategy and vehicle acceleration, the position information of the CAV at the next moment is determined, wherein the third model sends the CAV position information to the second model for the next optimization.

[0072] First, the objective function is constructed. The third model is based on the phase sustaining time t provided by the second model. p By selecting an appropriate lane-changing strategy g for CAV ω And formulate a reasonable longitudinal acceleration optimization curve a ω The goal is to minimize CAV travel time delays, reduce CAV fuel consumption, and lower lane-changing costs. Therefore, the objective function can be expressed as:

[0073]

[0074] Where ω represents the identifier of CAV, h1 represents the start time of trajectory optimization for the CAV in the control zone, h2 represents the optimization time of the CAV in the control zone, and h3 represents the optimization time of the CAV in the dedicated lane zone. This is the optimization time for CAV from the control region to the conflict region. δ ω The identification number indicates whether the vehicle has entered the conflict zone; entering δ ω If it is 0, otherwise δ ω It is 1. a ω This represents the longitudinal acceleration optimization curve. This indicates the number of lane changes. Within the control zone, if the CAV does not perform any lane changes, Recorded as 0; there was 1 lane change. The count is 1. g ω Let θ1 represent the weighting coefficient of CAV travel delay time, θ2 represent the weighting coefficient of CAV fuel consumption, θ3 represent the weighting coefficient of lane-changing cost, and Δt represent the time interval.

[0075] Vehicle delay typically refers to the difference between actual travel time and free travel time; therefore, reducing this gap is one of the key objectives of model control. The smoothness of vehicle acceleration also significantly impacts vehicle emissions. The number of lane changes by a CAV greatly affects lane-changing costs. Since this paper focuses on the traffic efficiency and energy conservation of CAVs, the weighting coefficients θ1, θ2, and θ3 for different control objectives are not the same, i.e., θ1 > θ2 >> θ3.

[0076] Then, constraints are applied to the objective function, including:

[0077] Step 1: Constrain the lane-changing sequence according to the first constraint condition; the first constraint condition includes that the duration of the continuous lane-changing interval of the CAV is greater than the preset minimum time interval, and that the CAV can only change lanes at most each time.

[0078] The expression for the first constraint is:

[0079]

[0080]

[0081] Where M represents a sufficiently large number, which can be considered as infinity, ∈ lc It is the minimum time interval between two consecutive lane change operations; k(g) ω (t) represents the lane number where vehicle ω is located at time t. The first formula in the expression constrains the time between two consecutive lane changes by CAV, requiring it to be greater than the minimum time interval; the second formula in the expression constrains that CAVω can only change lanes at most in one lane change.

[0082] Step 2: Constrain the vehicle motion according to the second constraint condition; the second constraint condition includes the vehicle's acceleration being between the preset maximum acceleration and the preset maximum deceleration, and the vehicle's speed being between 0 and the preset vehicle speed limit.

[0083] In addition to meeting the above constraints, vehicles must also meet the following constraints, including vehicle kinematic constraints and longitudinal safe distance constraints.

[0084] Before the CAV enters the intersection conflict zone, given the CAV's position at time t0... speed acceleration and lanes Given the information, the vehicle's motion conforms to a second-order vehicle kinematics model, that is:

[0085]

[0086]

[0087] The expression for the second constraint condition is:

[0088]

[0089]

[0090] Among them, v ω (t) represents the velocity ω of the vehicle at time t, v ω (t+1) represents the velocity ω of the vehicle at time t+1, x ω (t) represents the position of vehicle ω at time t, x ω (t+1) represents the position of vehicle ω at time t+1, a ω (t) represents the vehicle's acceleration ω at time t, and Δt represents the time interval from time t to time t+1. This represents the maximum acceleration ω of the vehicle. This represents the maximum deceleration of the vehicle, ω. It is the speed limit for vehicles on the road segment where vehicle ω is located.

[0091] Step 3: Constrain the longitudinal safe distance according to the third constraint condition; the third constraint condition includes the current vehicle maintaining a longitudinal safe distance from the vehicle in front and the vehicle behind, respectively, and the longitudinal safe distance is determined by the speed of the current vehicle.

[0092] To ensure safe vehicle operation, vehicles should maintain a longitudinal safety distance from vehicles in front and behind. This longitudinal safety distance varies at different vehicle speeds. For ease of calculation, this embodiment uses the Newell vehicle following model to determine the third constraint, including:

[0093]

[0094] x ω (t)≤l A +l D

[0095]

[0096]

[0097]

[0098] Where, x ω’ Indicates the position of vehicle ω'; τ cf d represents the time displacement in the Newell vehicle following model. cf To represent the spatial displacement in Newell's vehicle following model, in order to make For integers, Δt should be set appropriately; Vp (g ω (t) is the set of lead vehicles in front of the CAV, V f (g ω (t) is the set of vehicles following behind the CAV. v represents the maximum deceleration of the vehicle ω'. ω’ (t) represents the velocity of vehicle ω' at time t, l A Indicates the length of the search area, l D Indicates the length of the dedicated lane area.

[0099] The first formula in the third constraint condition represents the distance constraint between the vehicle ahead ω' and vehicle ω in the same lane, while the second formula in the third constraint condition represents the distance constraint between the following vehicle ω' and vehicle ω in the same lane.

[0100] Finally, since the second model in the two-layer model, i.e., the traffic light phase optimization, is relatively simple, while the lane-changing strategy g of the third model... ω (t) and acceleration a ω (t) Optimization is a complex nonlinear optimization problem with multiple constraints and variables, and it suffers from a large solution space. To address this, this embodiment, based on the calculation of the optimal phase duration using Webster's optimal period, designs an improved Cuckoo Algorithm solution model based on the EO algorithm, which has strong search performance, to optimize the objective function.

[0101] The cuckoo search algorithm is an algorithm based on the Lévy flight search mechanism, inspired by the biological phenomenon of cuckoos parasitizing and incubating their young. The nest location is updated according to the following search method:

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108] in, Let X represent the spatial location of the i-th bird's nest in generation t, n be the dimension of the optimization problem, m be the number of bird's nests per generation, α be the step size factor to control the range of random search, α0 be a constant, usually taken as 0.01, and X best This is the current optimal solution; symbol Levy(λ) represents point-to-point multiplication; it is a random search vector generated by a Rhine flight with parameter λ (1 < λ ≤ 3), and u represents an expected value of 0 and a variance of δ. u The normal distribution is given by v, where v represents a normal distribution with an expected value of 0 and a variance of 1, and Γ represents the integral operation.

[0109] Because the Cuckoo Algorithm is prone to getting trapped in extreme values ​​during local searches, leading to premature convergence, this embodiment introduces an extreme value dynamics optimization algorithm within the framework of the Cuckoo Algorithm to improve the accuracy and speed of model solving. For a minimization problem, the EO algorithm's process is as follows:

[0110] Step 1: For the bird's nest Let X be the optimal solution found so far. best Its objective function value is C(X) best ).

[0111] Step 2: [Regarding the current bird's nest] Perform the following operations: First, calculate the fitness λ of each component xij. ij j∈{1, 2, ..., n}. Next, sort the n fitness values ​​and find the element x with the smallest fitness. imin , that is, λ imin ≤λ ij If j = 1, 2, ..., n, then x imin That is, the worst-case element. Then, in the current Bird's Nest... Find a neighbor in the neighborhood. Force the worst element x imin Changes occur; and acceptance is unconditional. Finally, if the current objective function value C(X) best (greater than) but

[0112] Step 3: Repeat step 2 until the termination condition is met.

[0113] Step 4: Obtain the optimal solution X best and the optimal value of the objective function C(X) best ).

[0114] In conjunction with the above, the objective function is solved using an improved cuckoo algorithm based on extreme value optimization, including:

[0115] Step 1: Arrange the randomly generated CAV spatial locations for each time period into a bird's nest, whereby the spatial locations include lane selection information and vehicle acceleration.

[0116] Based on the planned and predicted trajectories of the vehicles ahead, and combined with relevant parameters from the Cuckoo algorithm, a nest is formed by randomly generating CAV spatial locations for each time period. For each nest... Its dimensions are N×T, where N is the number of controlled variables; in this embodiment, lane selection and acceleration are controlled variables. The fitness of each bird's nest is calculated, and the optimal location of the bird's nest X is determined. best .

[0117] Step 2: Determine if the preset number of searches has been reached. If the preset number of searches has not been reached, generate the next spatial location for each bird's nest according to Levi's flight path; otherwise, proceed directly to the next step.

[0118] In this embodiment, the preset number of attempts is set to 5.

[0119] Step 3: Based on the spatial location of the next step, optimize each bird's nest using an extreme value optimization algorithm, update the spatial location of the bird's nest, and obtain the optimal spatial location of the bird's nest and its corresponding fitness.

[0120] Bird's Nest x in i1 x i2 , ..., x in This is called the individual value. Updating the spatial location of a bird's nest requires updating the individual minimum and the global minimum: For each bird's nest, calculate its fitness value and compare it with the individual value. If it is better, replace the current individual minimum and the individual optimal position. Then compare the fitness of each bird's nest with the global optimum. If it is better than the global optimum, update the current global optimum and the position of the global optimal bird's nest.

[0121] Step 4: Determine if the maximum number of iterations has been reached. If the maximum number of iterations has been reached, obtain the optimal spatial location of the bird's nest; then determine the optimal spatial location of the bird's nest. Otherwise, return to Step 2.

[0122] Based on the above operations, the optimized lane-changing strategy and vehicle acceleration are obtained; based on the optimized lane-changing strategy and vehicle acceleration, the position information of the CAV at the next moment is determined; the third model sends the CAV position information to the second model for the next optimization.

[0123] The optimal phase duration is formed by combining the optimized phase duration of the CAV at each moment in the road; the optimal CAV trajectory is formed by combining the optimized CAV position information at each moment in the road.

[0124] To verify the advancement of the optimization strategy proposed in this paper, we set up the following... Figure 3The simulation scenario involves a four-arm intersection, each arm with four lanes: a left-turn lane (HV), a dedicated lane (CAV), a straight-ahead lane (HV), and a dedicated right-turn lane. Communication is instantaneous. The signal control scheme prioritizes straight-ahead traffic before left-turn traffic. The signal zone is 100m long, the control zone is 500m long, and the dedicated lane zone is 30m long. Speed ​​limits are defined for the signal zone, control zone, and dedicated lane zone. The speed limit is 13.8 m / s, and the speed limit in the conflict zone is 8.3 m / s.

[0125] To better demonstrate the advantages of traffic light phase optimization and CAV trajectory planning, and to eliminate differences caused by driving behavior, this embodiment uses the same parameters for both CAV and HV in the driving model settings, namely, a vehicle length of 4m. The spatial displacement and temporal displacement in the car following model are 6m and 1s, respectively. The minimum time interval between consecutive lane change actions is ∈ lc The maximum acceleration of the vehicle is 5 seconds. and minimum acceleration The magnitudes are 2m / s 2 and 4m / s 2 In the objective function of the third model, the weighted parameters θ1, θ2, and θ3 are 1000s. -1 10m / s 2 And 1, the time step Δt is set to 1s. The vehicle enters following a Poisson distribution, and the initial velocity of the CAV vehicle entering the control zone is... and The values ​​are randomly generated between them; this embodiment sets them to be random. It is 3m / s. The speed is 13.8 m / s. The Wiedermann 74 algorithm in VISSIM 4.3 was used to capture the driving behavior of the vehicle (HV) in the simulation.

[0126] The average vehicle arrival rate per arm was 4200 vehicles per hour. The simulation included varying traffic demands, from undersaturated to oversaturated. The proportions of vehicles turning left, going straight, and turning right were 0.4, 0.4, and 0.2, respectively. The CAV penetration rate was 50%, except for the CAV penetration analysis. In the second model, a minimum green light time was set. Maximum green light time Interval time r e，i =2s. This paper uses Python 3.7.4 to write the simulation model, and all experiments were completed on a desktop computer with an Intel 2.6GHz CPU and 16GB of memory. To improve computational efficiency, five threads were used for parallel computation. Five random seeds were used in each simulation to account for the uncertainties of vehicle arrival and HV driving behavior. Each simulation ran for 1200 seconds, with a warm-up period of 150 seconds.

[0127] Analysis of the above simulation scenarios yielded the following results:

[0128] Because this embodiment sets up a signaling zone and pre-signal lights before the CAV control zone, the second model can promptly obtain the type and direction of travel of vehicles about to enter the control zone and dedicated lane zone. This will help optimize intersection signal lights and improve traffic capacity. To better analyze the advantages of the control model proposed in this embodiment, the VISSIM fixed-time signal control model, which does not have a signaling zone, is used as a control. The fixed-time signal model uses a signal plan cycle length of 60s, a green duration of 14s for both the straight-ahead and left-turn phases, a yellow light duration of 2s, and right-turning vehicles are not controlled by the signal lights. Figures 4A-4D Simulation analysis was conducted on the vehicle throughput, delay, and fuel economy of heterogeneous traffic flows under two different signal control strategies.

[0129] from Figure 4A It is clearly evident that when vehicles arrive at the intersection with a Poisson distribution, the cooperative timing strategy proposed in this embodiment significantly improves the intersection's capacity compared to a fixed timing strategy, increasing the throughput of all vehicles, CAVs, and HVs by 8%, 10%, and 5.1%, respectively. In contrast, the introduction of signal zones and pre-signal lights helps improve intersection throughput and has a greater impact on CAV throughput. This is mainly because the signal zones allow the intersection to obtain HV turning information in advance, aiding CAV trajectory planning and signal control, and enabling a more efficient allocation of signal control time.

[0130] from Figure 4B It becomes more apparent that, in addition to improving throughput, when vehicles arrive at the intersection with a Poisson distribution, the cooperative timing strategy proposed in this embodiment significantly improves the travel delay of heterogeneous traffic flows at the intersection compared to a fixed timing strategy. Specifically, the delays for all vehicles, CAVs, and HVs are reduced by 13.5%, 18.4%, and 9.26%, respectively. This further confirms that the introduction of signaling zones and pre-signal lights can effectively reduce vehicle delays. The reason for this change is that obtaining HV turning information in advance through the signaling zone helps the control system make signal adjustments, reducing conflicts caused by lane changes between heterogeneous traffic flows and increasing the average vehicle speed. This conclusion can also be confirmed by statistically analyzing the average vehicle speed and average number of lane changes between the adjustment zone and the conflict zone for heterogeneous traffic flows.

[0131] like Figure 4C and Figure 4DAs shown, compared to the fixed timing strategy, the cooperative timing strategy proposed in this embodiment increases the average speed of all vehicles, CAV and HV in heterogeneous traffic flow at the intersection by 8.54%, 11.6% and 5.9% respectively, and reduces the average number of lane changes by 16.38%, 22.15% and 10.6% respectively.

[0132] from Figure 5 It is clearly evident that, regarding average vehicle fuel economy, when vehicles arrive at intersections with a Poisson distribution, the cooperative timing strategy proposed in this embodiment significantly improves the average fuel economy of heterogeneous traffic flows compared to a fixed timing strategy, increasing the average fuel economy of all vehicles, CAVs, and HVs by 12.9%, 17.9%, and 8.8%, respectively. This further demonstrates that signal control strategies based on signal zones and pre-signal lights, along with vehicle trajectory planning, can improve the quality of heterogeneous traffic flow.

[0133] pass Figure 6A The spatiotemporal trajectory of the CAV shows that, due to the advance acquisition of intersection signal control information and HV steering information, the CAV can make more reasonable trajectory planning, avoiding vehicle pauses, thus reducing vehicle starts at the intersection, increasing vehicle speed, and making full use of the green light time. To further demonstrate this point, we randomly selected one CAV and analyzed the acceleration changes under two different control strategies during this time period, such as... Figure 6B As shown, it is evident that, compared to the fixed timing strategy, the cooperative timing strategy proposed in this embodiment reduces the speed fluctuation of CAV.

[0134] As analyzed above, the optimized control model in this embodiment has a positive impact on the passage of heterogeneous traffic flows at intersections, helping to reduce traffic delays and improve average fuel economy while increasing intersection capacity. To further analyze the benefits of the proposed model, this embodiment will explore the impact of the control model on heterogeneous traffic flows under different traffic demands and varying CAV penetration rates.

[0135] Figure 7A , Figure 7B and Figure 7C The average delays of heterogeneous traffic flow, CAV, and HV are shown when the saturation levels are 0.5, 1.0, and 1.25. With increasing traffic demand, the average delays of heterogeneous traffic flow, CAV, and HV all increase significantly. With increasing CAV penetration, the average delays of heterogeneous traffic flow and HV at each demand level decrease significantly. This indicates that when there are more CAVs in mixed traffic, heterogeneous traffic flow signal control based on pre-signal lights and dedicated CAV lanes, along with CAV trajectory optimization, can reduce the delays of heterogeneous mixed traffic flow and HV, with the most significant reduction in HV delay. Figure 7D , Figure 7E and Figure 7F This study demonstrates the reduction in average delays for heterogeneous traffic flows, CAVs, and HVs under different demand levels, compared to a baseline case without CAV trajectory planning and traffic light coordination optimization. It is evident that while CAV trajectory planning is performed in a decentralized manner, it significantly improves CAV passage delays. Furthermore, when CAV penetration is high, for example, 70%, HV delay reductions also become significant, especially under conditions of excessive traffic saturation (V / C = 1.25), where delay reductions exceed 7 seconds. This is because, under more reasonable phase timing conditions, the introduction of pre-signal lights and dedicated CAV lanes provides more spatiotemporal resources for HV passage. HVs can adjust in advance, traversing intersections at higher speeds, avoiding sudden braking and acceleration, and preventing stops at stop barriers.

[0136] like Figures 8A-8B As shown, taking the HV traffic trajectory under different control strategies when the CAV penetration rate is 70% as an example, the trajectory segment in lane 3, i.e. the through lane, is marked with a dark color, while the trajectory segments in other lanes are marked with a light color. Figure 8A The HVs in the data stopped at the red light and started again after the light turned green, and their trajectories showed significant fluctuations. As a result, these HVs lost their startup time. In contrast, Figure 8B HVs can pass through the stop bars smoothly and quickly, avoiding stopping at the stop barriers and effectively utilizing the green light time. This demonstrates that pre-signal lights and dedicated CAV lanes effectively reduce HV delays.

[0137] In addition, for similar reasons, the average capacity of intersections will also be improved. Figure 9 The positive impact of CAV penetration rate on intersection capacity under different traffic demands was compared and analyzed. CAV penetration rates were tested at 0%, 25%, 50%, 75%, 85%, and 100%. The maximum capacity is reached when the traffic volume line flattens out. Figure 9 The results show that when the CAV penetration rate increases from 0% to 50%, the capacity improves significantly by 9.16%. A significant improvement of 13.36% is observed when the CAV penetration rate reaches 100%. These observations indicate that CAV trajectory planning can significantly improve intersection capacity, but requires a high CAV penetration rate, and the establishment of dedicated CAV lanes is particularly helpful in improving the capacity of heterogeneous traffic flow.

[0138] In summary, this embodiment proposes a signal control and CAV trajectory planning method based on pre-signal lights. It combines traditional signal timing schemes with CAV trajectory planning for dedicated CAV lanes and approach lanes to manage mixed traffic of HV and CAV. Based on discrete time, the control system constructs a two-layer optimization model for intersection signal phase duration and CAV trajectory planning, based on the collection of HV vehicle light semantic information, CAV turning demand information, and the position and speed information of all vehicles in the control zone and dedicated lane zone. The second model optimizes the phase duration in real time based on the actual total number and type of vehicles entering the control zone, while the third model optimizes the CAV lane-changing strategy and vehicle acceleration optimization curve based on the phase duration optimized by the second model. Based on the calculation of the optimal phase duration using Webster's optimal period formula, an improved Cuckoo algorithm solution model combining the EO algorithm with strong search performance is designed. Numerical studies verify the advantages of the proposed heterogeneous traffic flow signal control and CAV trajectory optimization based on pre-signal lights and dedicated CAV lanes. The capacity of the intersection is significantly improved, the average fuel consumption and lane change frequency of CAVs can be significantly reduced, and the speed and delay of heterogeneous traffic flows are significantly improved.

[0139] A computer-readable storage medium having a computer program stored thereon, the computer-readable storage medium storing a signal control and CAV trajectory planning program based on a pre-signal light, the signal control and CAV trajectory planning program based on the pre-signal light being executed by a processor to implement the signal control and CAV trajectory planning program based on the pre-signal light.

[0140] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A signal control and CAV trajectory planning method based on pre-signal lights, characterized in that, include: Based on the pre-signal lights set in the signaling area, obtain the intersection phase information and first information, the first information including HV vehicle light semantic information and CAV turning demand information; Acquire second information within the control zone and dedicated lane zone, the second information including the position, speed, acceleration, and lane selection information of all vehicles; The first information and the second information are processed by a first model to predict the vehicle trajectory and obtain the total number and type of vehicles in the control area; the vehicle trajectory includes the trajectory of the vehicle following the vehicle in front and the trajectory of the vehicle during lane change. The intersection phase information, vehicle trajectory, total number of vehicles, and vehicle type are processed by a second model and a third model to predict the optimal phase duration and optimal CAV trajectory. This includes: optimizing the phase duration using the second model based on the intersection phase information, total number of vehicles, and vehicle type; optimizing lane-changing strategies and vehicle acceleration using the third model based on the vehicle trajectory and optimized phase duration, with the goal of minimizing CAV travel delay time, CAV fuel consumption, and lane-changing costs; and determining the CAV position information at the next moment based on the optimized lane-changing strategy and vehicle acceleration. The third model then sends the CAV position information to the second model for the next optimization. The expression for the objective function of minimizing CAV traffic delay time, CAV fuel consumption, and lane-changing cost is as follows: in, The identifier number representing CAV. This indicates the moment when the CAV begins trajectory optimization within the control region. It is the optimization time of CAV within the control region. This refers to the optimized time of CAVs within dedicated lane zones. The indicator number signifies whether the vehicle has entered a conflict zone; 0 indicates entry, and 1 indicates otherwise. This represents the longitudinal acceleration optimization curve. Indicates the number of lane changes. Indicates a lane-changing strategy. The weighting coefficient represents the CAV passage delay duration. The weighting coefficient for CAV fuel consumption. The weighting coefficient represents the lane-changing cost. The time interval is represented; the objective function is constrained, including: constraining the lane-changing sequence according to the first constraint condition; the first constraint condition includes that the duration of the continuous lane-changing interval of the CAV is greater than the preset minimum time interval, and that the CAV can only change lanes at most each time it changes lanes. The vehicle motion is constrained according to the second constraint condition; the second constraint condition includes the vehicle's acceleration being between a preset maximum acceleration and a preset maximum deceleration, and the vehicle's speed being between 0 and a preset vehicle speed limit. The longitudinal safe distance is constrained according to the third constraint condition; the third constraint condition includes that the current vehicle maintains a longitudinal safe distance from the vehicle in front and the vehicle behind, respectively, and the longitudinal safe distance is determined by the speed of the current vehicle; The objective function is solved using an improved cuckoo algorithm based on extremum optimization, including: The randomly generated CAV spatial locations for each time period are combined to form a bird's nest, and the spatial locations include lane selection information and vehicle acceleration. Determine if the preset number of searches has been reached. If not, generate the next spatial location for each bird's nest according to Levi's flight path. Based on the next step of spatial location, the extreme value optimization algorithm is used to optimize each bird's nest and update the spatial location of the bird's nest; Determine if the maximum number of iterations has been reached. If the maximum number of iterations has been reached, obtain the optimal spatial position.

2. The signal control and CAV trajectory planning method based on pre-signal lights according to claim 1, characterized in that, The step of obtaining intersection phase information and first information based on the pre-signal lights set in the signaling area includes: Based on the pre-signal lights set in the signaling area, the intersection phase information is obtained, which includes the phase sequence and phase duration. Based on the intersection phase information, the HV vehicle light semantic information and CAV steering demand information in the signaling area are obtained through the light signal detection device.

3. The signal control and CAV trajectory planning method based on pre-signal lights according to claim 1, characterized in that, The first model includes a second-order car following model; the second-order car following model is used to process the first information and the second information, and the specific process includes: Based on the first and second information, obtain the current status information of the vehicle and the corresponding current status information of the vehicle in front; Based on the current status information of the vehicle and the current status information of the vehicle in front, calculate the relative position and relative speed between the current vehicle and the vehicle in front. Based on the relative position and relative speed, the vehicle's state information at the next moment can be predicted; Repeat the above steps to determine the trajectory of the vehicle following the vehicle in front.

4. The signal control and CAV trajectory planning method based on pre-signal lights according to claim 3, characterized in that, The first model also includes an improved lane-change model, which processes the first information and the second information using the lane-change model. The specific process includes: Based on the first and second pieces of information, it is determined that the vehicle needs to change lanes; Calculate the guiding angle based on the vehicle's current position and the target position for lane changing; Based on the guidance angle and the vehicle's current state information, a path planning algorithm is used to determine the vehicle's trajectory during the lane change process.

5. The signal control and CAV trajectory planning method based on pre-signal lights according to claim 1, characterized in that, The step of optimizing phase duration using a second model based on intersection phase information, total number of vehicles, and vehicle type includes: Based on the intersection phase information, the total number of vehicles and vehicle type, the cycle duration is calculated using the Webster optimal cycle algorithm. Based on the cycle duration, the effective green light time for each phase is optimized using the traffic flow ratio, and the effective green light time is between the shortest and longest effective green light times.

6. The signal control and CAV trajectory planning method based on pre-signal lights according to claim 5, characterized in that, The expression for the second model is: in, The optimal signal cycle is given by L, where L is the total intersection loss time and Y is the sum of the intersection traffic flow ratios. For lane e The effective green light time The number of phases of the signal. For the signal loss time, For lane e No. Red light duration for each phase For the first i The traffic flow ratio of the critical lanes in each phase, , For lane e No. i The effective green time for each phase Indicates the shortest effective green light duration for a phase; The longest valid green light duration for a given phase.

7. A computer-readable storage medium having a computer program stored thereon, the computer-readable storage medium storing a signal control and CAV trajectory planning program based on a pre-signal light, the signal control and CAV trajectory planning program based on the pre-signal light being executed by a processor to implement the signal control and CAV trajectory planning program based on any one of claims 1-6.

Images