Intersection signal control optimization method based on travel time relationship function

Abstract

The application relates to an intersection signal control optimization method based on a travel time relation function, and relates to the technical field of signal control. The travel time calculation model and the queuing time calculation model of individual vehicles on a signal control section are constructed, the entire process of the individual vehicles from entering the research section, driving in the section, parking and queuing, and passing through the intersection is sequentially connected, the space-time distribution of the individual vehicles in the section can be accurately described, after the travel time relation function is constructed, the connection between the travel time and the intersection signal control is considered, the average travel time of all the individual vehicles passing through the intersection under a certain set signal control scheme is taken as an objective function, an intersection signal control optimization model is constructed for intersection signal control optimization, and the vehicle passing efficiency of the intersection is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal control, and more particularly to an intersection signal control optimization method based on travel time relationship function. BACKGROUND

[0002] The intersection refers to the intersection of two or more roads, which is the area where vehicles converge and diverge. The intersection is a key point of the urban road network, determines whether the road flow is smooth, and is also a key point for solving traffic congestion. Therefore, improving the management level of the intersection is an effective way to solve traffic problems, and the management of the intersection mainly reflects the design of the signal control scheme.

[0003] By using intelligent control algorithm to scientifically control regional traffic signals and reasonably induce and disperse traffic flow, traffic congestion is relieved, traffic safety and efficiency are improved, which is an important means of national development and application. The specific method is as follows: an optimization model is constructed with the minimum delay of the intersection as the target, the optimization model is solved by using genetic algorithm, and finally the optimization solving result is simulated and verified by a simulation platform to analyze the optimization effect. In addition, a single intersection traffic signal optimization control method is disclosed in the prior art. On the basis of establishing a single intersection traffic signal optimization control model, a chaotic genetic algorithm is proposed to solve the model by designing a fuzzy self-adaptive strategy, which can realize the optimization of the single intersection traffic signal timing under the premise of ensuring traffic safety. Finally, the effectiveness of the control method is verified through simulation analysis.

[0004] However, simulation and simulation through a simulation platform need to go through an important process, that is, by adjusting parameters, the simulation result of the traffic flow tends to be consistent with the operation law of the actual traffic flow. However, the actual traffic operation is a highly complex process, the individual differences of vehicles are large, and the traffic signal and traffic density easily affect the vehicle travel, which leads to the difficulty in obtaining a simulation result that is indistinguishable from the actual traffic flow. SUMMARY

[0005] In order to solve the problems of large individual differences of vehicles, low accuracy of vehicle travel time calculation, and poor signal control optimization effect of the intersection, the present application proposes an intersection signal control optimization method based on travel time relationship function, which considers the influence of traffic signal and traffic density on individual travel time. The travel time relationship function is applied to the intersection signal control optimization, the model solving and verification process are based on the travel time relationship function, and the solving does not need to be solved through the simulation platform, thereby improving the traffic efficiency of the intersection.

[0006] The present application aims to at least partially solve the above technical problems.

[0007] To solve the above technical problems, the technical scheme adopted by the present application is as follows:

[0008] An intersection signal control optimization method based on travel time relationship function, the method comprises the following steps:

[0009] S1: Construct a travel time calculation model and a queuing time calculation model of individual vehicles on the signal control section, based on the travel time calculation model and the queuing time calculation model, form a travel time relationship function of individual vehicles;

[0010] S2: According to the travel time relationship function of individual vehicles, solve the average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme;

[0011] S3: Taking the minimum average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme as the objective function, taking the longest and shortest green light time of each phase, the total green light time constraint and the intersection traffic congestion degree constraint as the constraint condition, construct the intersection signal control optimization model;

[0012] S4: Solve the intersection signal control optimization model, output the optimal intersection signal timing result.

[0013] Preferably, the travel time calculation model of step S1 is constructed by using a multiple linear regression model, and the process is:

[0014] S101: Considering the influence of road section length And traffic density K , a travel time calculation model is constructed, the expression is:

[0015]

[0016] Wherein, The road section length is L; The traffic density is D, The function relationship formula describing the relationship between traffic density and travel time is f; The coefficient of the travel time calculation model is a, i =1,2,3;

[0017] S102: Based on the historical travel data of individual vehicles on the signal control section, the coefficients of the travel time calculation model are fitted by using linear fitting .

[0018] According to the above technical scheme, the traffic density and road section length which mainly affect the travel time are taken as influencing factors, and a multiple linear regression model is used to construct the travel time calculation model, which fully describes the space-time distribution of individual vehicles on different sections.

[0019] Preferably, in step S1, the number of cycles traversed by an individual vehicle traveling from the upstream intersection to the downstream intersection is denoted as . n , >2 refers to the total number of downstream intersection signal light cycles crossed during the time an individual travels on the studied road segment before reaching the signalized intersection. The calculation formula is:

[0020]

[0021] in, The arrival time refers to the moment when a vehicle first stops or passes the stop line at the downstream intersection. The calculation expression is: The duration of the signal period;

[0022] Calculate the green light time for each vehicle during the first cycle. The expression is:

[0023]

[0024] in, The duration of the green light within a single cycle of the traffic light; The entry signal time refers to the duration after the green light of the traffic light at the downstream intersection corresponding to the vehicle's outgoing direction has started when the vehicle enters the studied road segment; if the traffic light at the downstream intersection is green when an individual vehicle enters the signal-controlled road segment, then... This represents the remaining green light duration within the current cycle. If a vehicle enters a signal-controlled section when the downstream intersection's traffic light is red, the remaining green light duration within that cycle is 0. =0;

[0025] Calculate the individual driving time of the vehicle during the first n Green time per cycle The expression is:

[0026] .

[0027] Remaining The green light duration is calculated as two green light cycles. -2 times, then the green light duration for an individual during driving... for:

[0028]

[0029] In the formula, The green light time for individual driving.

[0030] According to the above technical means, the individual vehicle is sequentially connected from the entire process of entering the research road section, driving regularly in the road section, parking in the queue, and passing through the intersection, and the space-time distribution of the individual vehicle in the road section is described.

[0031] Preferably, the number of vehicles in front of the individual vehicle is calculated when the individual vehicle arrives at the downstream intersection, and the process is as follows:

[0032] Based on the absolute overtaking amount data of each vehicle individual, the average value, median, overtaking probability and overtaken probability of multiple absolute overtaking amounts of the individual are calculated, the absolute overtaking amount refers to the difference between the number of overtaking vehicles in front of the individual driving in the lane and the number of overtaken vehicles by the rear vehicles in the unit road length, and the calculation expression is: , wherein, is the number of times of overtaking the vehicle in front of the individual during driving in the unit road length; is the number of times of being overtaken by the rear vehicle during driving of the individual in the unit road length;

[0033] Based on the BP neural network model, the overtaking probability, the overtaken probability, the average value of the absolute overtaking amount, the median and the traffic density are taken as the input characteristic variables of the BP neural network model, and the absolute overtaking amount of the vehicle individual is output P ;

[0034] Solving the number of vehicles in front of the individual when the individual arrives at the downstream intersection , the expression is:

[0035]

[0036] , wherein, q represents the number of vehicles in front of the individual when the individual arrives at the downstream intersection; is the number of vehicles in the same direction as the individual, which refers to the total number of vehicles in the same direction as the individual driving out of the road section when the individual enters the research road section; is the number of lanes in the same direction as the individual driving out of the road section; is the average interval time of vehicles passing the stop line during the green light period.

[0037] According to the above technical means, considering the large difference between individual vehicles, the absolute overtaking amount during driving of the individual is considered in the process of calculating the queue length, the absolute overtaking amount calculation model is constructed based on the BP neural network, the driving habit of the individual is calibrated through the overtaking habit of the individual history, the randomness and difference of the individual trip are effectively reduced, and the calculation accuracy of the travel time is improved.

[0038] Preferably, according to the number of vehicles in front of the individual vehicle when the individual vehicle arrives at the downstream intersection and the traffic capacity of the intersection, the number of parking times of the individual vehicle is calculated, and the calculation expression is:

[0039]

[0040] in, This indicates the number of times a vehicle stops.

[0041] Preferably, the expression for the queuing time calculation model described in step S1 is:

[0042]

[0043] in, Queuing time; Indicates the number of times an individual vehicle stops; This refers to the number of vehicles queuing ahead of an individual vehicle when it arrives at the downstream intersection. The arrival time of the signal; This represents the green light time required for an individual vehicle to wait for the queue ahead to disperse and pass through the stop line. The red light time represents the time during which a vehicle arrives at the downstream intersection when the light is on (i.e., the time during which the light is on). At that time, individual vehicles need to wait. A vehicle must complete a full red light to pass through the intersection; when an individual vehicle arrives at the intersection during the red light period, that is... At that time, the red light duration for that phase is still... Individual vehicles still need to wait. - You need to wait for one full red light to pass through the intersection.

[0044] Preferably, the expression for the individual vehicle travel time relationship function is:

[0045]

[0046] in, A function representing the time relationship of individual vehicle journeys.

[0047] Preferably, in step S2, a certain set signal control method is defined as follows: G The expression for the average travel time of all individual vehicles passing through the intersection under a certain signal control scheme is:

[0048]

[0049] in, The intersection signal control scheme is as follows: At that time, the first person passing through the intersection The travel time of each individual vehicle through the research road segment; V This indicates the total number of vehicles passing through the intersection.

[0050] Preferably, the intersection signal control optimization model described in step S3 is as follows:

[0051] Objective function:

[0052]

[0053] in, This represents the signal control scheme for an intersection, where the decision variables are the green light duration for each phase. = , This represents the number of phases at the intersection. For the first Green light duration for each phase;

[0054] Constraints:

[0055] 1) Longest and shortest green light time constraints:

[0056]

[0057] in, To determine the shortest possible green light duration allowed by the phase, and considering pedestrian safety, through Sure, The width of the pedestrian crossing The speed at which pedestrians walk. Yellow light duration The time is when all red is visible; The longest green light duration allowed for a given phase, ensuring that the traffic capacity is appropriate for the cycle length and other phases of traffic;

[0058] 2) Total green light duration constraint:

[0059]

[0060] in, This represents the total duration of the yellow light within the cycle. The cycle duration of the intersection signal control lights.

[0061] 3) Intersection traffic congestion constraints:

[0062]

[0063] in, This represents the maximum queuing time index. For a certain time interval, the first Queue time index at the entrance lane; For a certain time interval, the first i Queuing time at intersections for vehicles entering the lane; The signal control cycle duration is the maximum queuing time exponent. When the value range is [1.5, 2.1), the traffic congestion level of the intersection is moderate. If the traffic congestion level of the intersection is less than moderate congestion, the green light duration of the phase is constrained, and the expression is:

[0064] ( <1.5.

[0065] Based on the aforementioned technical means, the shortest travel time is taken as the optimization objective and applied to the optimization of intersection signal control, thereby improving the traffic efficiency of intersections from a technical perspective.

[0066] Preferably, in step S4, a genetic algorithm is used to solve the intersection signal control optimization model.

[0067] Compared with the prior art, the beneficial effects of the technical solution adopted in this invention are as follows:

[0068] This invention proposes an intersection signal control optimization method based on a travel time relationship function. By constructing a travel time calculation model and a queuing time calculation model for individual vehicles in a signal-controlled road segment, the entire process of an individual vehicle entering the study road segment, its driving pattern within the road segment, queuing, and passing through the intersection is sequentially linked, accurately depicting the spatiotemporal distribution of individual vehicles in the road segment. After constructing the travel time relationship function, the connection between travel time and intersection signal control is considered. The objective function is to minimize the average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme. An intersection signal control optimization model is constructed for intersection signal control optimization, thereby improving the vehicle traffic efficiency of the intersection. Attached Figure Description

[0069] Figure 1 A flowchart illustrating the intersection signal control optimization method based on travel-time relationship function proposed in this embodiment of the invention;

[0070] Figure 2 A schematic diagram illustrating the number of cycles n spanned during the individual vehicle's driving period as proposed in this embodiment of the invention;

[0071] Figure 3 This is a schematic diagram illustrating the green light time during driving as proposed in this embodiment of the invention.

[0072] Figure 4 This indicates the entry signal time proposed in the embodiments of the present invention. With arrival time Schematic diagram;

[0073] Figure 5 This indicates the arrival signal time proposed in the embodiments of the present invention. A schematic diagram;

[0074] Figure 6 This diagram illustrates the travel-time relationship function proposed in the embodiments of the present invention.

[0075] Figure 7 This is a schematic diagram illustrating the process of using a genetic algorithm for optimization in the embodiments of the present invention;

[0076] Figure 8 This is a frequency distribution diagram showing the calculation error of the absolute overtaking amount proposed in the embodiments of the present invention. Detailed Implementation

[0077] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0078] To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions;

[0079] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings;

[0080] The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0081] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0082] In this embodiment, for ease of understanding, the following concepts are explained first:

[0083] (1) Absolute overtaking volume Within a unit road segment, this is the difference between the number of vehicles an individual overtakes and the number of vehicles overtaken from behind while driving in a lane. The calculation formula is as follows:

[0084]

[0085] In the formula, This refers to the absolute amount of overtaking. The number of times an individual overtakes a vehicle in front during a given period of time within a given road length; This refers to the number of times an individual is overtaken by vehicles behind them during their journey within a given road length.

[0086] (2) Entering the signal time : The duration for which the green light of the traffic light at the downstream intersection corresponding to the direction the vehicle is exiting has started when the vehicle enters the study section.

[0087] (3) Arrival time The signal time when a vehicle first stops or passes the stop line at the downstream intersection.

[0088]

[0089] In the formula, This refers to the travel time.

[0090] (4) Number of cycles spanned during the journey n The total number of downstream intersection signal light cycles crossed by an individual during the time they travel on the studied road segment before reaching the signalized intersection.

[0091]

[0092] In the formula, The duration of the signal period; The rounding up symbol.

[0093] (5) The total number of vehicles on the road segment that are traveling in the same direction as the individual when the individual enters the study segment.

[0094] (6) Arrival time of signal The duration of the green light in the corresponding phase of the intersection traffic lights for the direction the vehicle is exiting, when a vehicle first stops or passes the stop line at the downstream intersection. The calculation formula is as follows:

[0095]

[0096] (7) Green light time during driving The total duration of the green light at the downstream intersection for the phase corresponding to the direction of vehicle departure, from the moment an individual enters the research section until the individual reaches the downstream intersection and first stops or passes the intersection without stopping.

[0097] Example 1 This example proposes an intersection signal control optimization method based on a travel-time relationship function. The flowchart of this method is as follows: Figure 1 As shown, see Figure 1 The method includes the following steps:

[0098] S1: Construct a calculation model for the travel time and queuing time of individual vehicles on signal-controlled road sections. Based on the travel time calculation model and queuing time calculation model, form a function relating the travel time of individual vehicles.

[0099] The travel time between two adjacent intersections in a city can be divided into two parts: travel time and queuing time. The expression for the relationship function between individual vehicle travel time is:

[0100]

[0101] in, A function representing the time relationship of individual vehicle trips. For travel time, Queuing time; travel time Travel time refers to the time it takes for an individual to travel from an upstream intersection to a downstream intersection on a road segment, taking into account road length and traffic density. It is calculated from the moment the individual crosses the stop line at the upstream intersection until they reach the downstream intersection and either stop for the first time or cross the stop line without stopping. Queuing time is also included. Queuing time refers to the time it takes for an individual to queue after first stopping at the downstream intersection. The time elapsed from when an individual first stops at the downstream intersection until they cross the stop line is the queuing time.

[0102] S2: Based on the individual vehicle travel time relationship function, calculate the average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme;

[0103] S3: The objective function is to minimize the average travel time of all individual vehicles passing through the intersection under a certain signal control scheme. The constraints are the longest and shortest green light time of each phase, the total green light time, and the traffic congestion of the intersection. An intersection signal control optimization model is constructed.

[0104] S4: Solve the intersection signal control optimization model and output the optimal intersection signal timing result.

[0105] The phase of a traffic light determines the time allotted for traffic flows in different directions. Signal timing, or the configuration of traffic lights at an intersection, is used in this embodiment. Traffic density and road segment length, which significantly influence travel time, are considered as influencing factors. A multiple linear regression model is used to construct the travel time calculation model. First, the independent variables of the multiple linear regression model are determined. The characteristics of travel time variation under the influence of road segment length and traffic density are analyzed. The relationship between traffic density and travel time is constructed through curve fitting to determine the independent variables of the travel time calculation model. After determining the independent variables of the multiple linear regression model, the travel time calculation model is constructed. Finally, the parameters of the travel time calculation model are determined through linear fitting.

[0106] Without considering other influencing factors, the relationship between road segment length and travel time is a linear one. (The last sentence appears to be incomplete and lacks context.) As independent variables in a multiple linear regression model, based on the distribution trends of travel time and traffic density, the relationship between traffic density and travel time is initially determined to be several simple elementary functions, such as exponential functions, quadratic functions, and linear functions. Through fitting historical data, using correlation coefficients and significance parameters as indicators, the relationship with the strongest correlation and statistical significance is selected. If a quadratic relationship best describes the relationship between travel time and traffic density and is statistically significant, then the relationship between traffic density and travel time is a quadratic linear relationship. , As independent variables in the multiple linear regression model; simultaneously, considering the combined effect of traffic density and road segment length on travel time, the product of traffic density and road segment length is also considered. As an independent variable.

[0107] In this embodiment, considering the combined effect of traffic density and road segment length on travel time, the specific construction process is as follows:

[0108] S101: Considering road segment length and traffic density K Based on the influence of [the factors], a model for calculating travel time is constructed, expressed as:

[0109]

[0110] in, The length of the road segment; For traffic density, A functional expression describing the relationship between traffic density and travel time; The coefficients for the travel time calculation model are used. i =1,2,3;

[0111] S102: Based on the historical driving data of individual vehicles in signal-controlled road sections, the coefficients of the driving time calculation model are fitted using a linear fitting method. .

[0112] In this embodiment, the key points for calculating queuing time are queue length and number of stops. Queuing time not only considers the impact of the number of vehicles released at the downstream intersection and the absolute number of overtaking by individual vehicles on the queue length of individual vehicles during the individual vehicle's journey on the road segment, but also analyzes the changing patterns of queue length and queue delay during the individual queuing period by combining the two different signal light states of green light and red light blocking.

[0113] like Figure 2 As shown, let the number of cycles traversed by an individual vehicle traveling from the upstream intersection to the downstream intersection be... n , >2 refers to the total number of downstream intersection signal light cycles crossed during the time an individual travels on the studied road segment before reaching the signalized intersection. The calculation formula is:

[0114]

[0115] in, The arrival time refers to the moment when a vehicle first stops or passes the stop line at the downstream intersection. The calculation expression is: The duration of the signal period;

[0116] Green light time during driving This refers to the total green light duration of the traffic lights at the downstream intersection corresponding to the phase of the vehicle's outgoing direction, from the moment an individual enters the research section until the individual reaches the downstream intersection and first stops or passes the intersection without stopping.

[0117] For an illustration of the green light time for each cycle, please refer to [link / reference]. Figure 3 First, calculate the green light time for each vehicle during the first cycle. The expression is:

[0118]

[0119] in, The duration of the green light within a single cycle of the traffic light; The entry signal time refers to the duration after the green light of the traffic light at the downstream intersection corresponding to the vehicle's outgoing direction has started when the vehicle enters the studied road segment; if the traffic light at the downstream intersection is green when an individual vehicle enters the signal-controlled road segment, then... This represents the remaining green light duration within the current cycle. If a vehicle enters a signal-controlled section when the downstream intersection's traffic light is red, the remaining green light duration within that cycle is 0. =0;

[0120] Calculate the individual driving time of the vehicle during the first n Green time per cycle The expression is:

[0121] .

[0122] At this time, when an individual vehicle arrives at the downstream intersection, the traffic light is green. The duration of the green light in that cycle is the total green light duration. If a vehicle arrives at the downstream intersection when the traffic light is red, the duration of the green light in that cycle is the total green light duration. The duration of the green light on a traffic light.

[0123] Remaining The green light duration is calculated as two green light cycles. -2 times, then the green light duration for an individual during driving... for:

[0124]

[0125] In the formula, The green light time for individual driving.

[0126] Then, the number of vehicles queuing ahead of each individual vehicle when it arrives at the downstream intersection is calculated as follows:

[0127] Based on the absolute overtaking data of each individual vehicle, the average, median, overtaking probability, and probability of being overtaken are calculated for each individual vehicle. The absolute overtaking quantity refers to the difference between the number of vehicles overtaken by an individual and the number of vehicles overtaken by an individual while driving in a lane within a unit road segment. The calculation expression is as follows: ,in, The number of times an individual overtakes a vehicle in front during a given period of time within a given road length; This represents the number of times an individual is overtaken by vehicles behind it during its journey within a unit of road length. Since an individual vehicle may have multiple trips on the same road segment and travel records on different road segments, multiple absolute overtaking data points can be extracted for the same individual vehicle.

[0128] Based on a backpropagation (BP) neural network model, the overtaking probability, the probability of being overtaken, the average and median of the absolute overtaking amount, and traffic density are used as input feature variables to output the absolute overtaking amount P of each vehicle. The first four input feature variables consider individual driving habits from different perspectives, reflecting individual driving behavior: the higher the overtaking probability, the more aggressive the individual is, the more inclined to drive quickly, and the larger the absolute overtaking amount. The last feature variable considers whether the individual has the conditions to overtake: when the traffic density is very low, there are fewer vehicles that the individual can overtake, and the absolute overtaking amount is closer to 0; when the density is very high, the individual has no space to overtake, which also affects the absolute overtaking amount. By comprehensively considering driving habits and overtaking conditions, the accuracy of function calculation is improved.

[0129] Given that the green light time and the absolute number of overtaking vehicles during the journey are known, we need to determine the number of vehicles queuing ahead of the individual when it reaches the downstream intersection. The expression is:

[0130]

[0131] in, q This indicates the number of vehicles queuing ahead of an individual vehicle when it arrives at the downstream intersection. The number of vehicles in the network for a given direction refers to the total number of vehicles on the road segment that are traveling in the same direction as the individual when the individual enters the research road segment. The number of lanes whose lane function is consistent with the individual's exit direction; This is the average interval between vehicles crossing the stop line during a green light.

[0132] The number of times a vehicle stops is calculated based on the number of vehicles queuing ahead of it when it arrives at the downstream intersection and the intersection's capacity. The calculation expression is as follows:

[0133]

[0134] in, Indicates the number of times an individual vehicle stops. Queue length The average interval between vehicles crossing the stop line during a green light; For green light duration, The arrival time of the signal, the arrival time of the signal See the illustrative explanation. Figure 5 When an individual arrives at the downstream intersection during the green light period, that is... Since the green light had already been running for some time when the individual arrived, it would take some time after the green light ended for the vehicles in front of the individual to clear the road. If the green light time is 1 second, then the time required for an individual to cross the intersection is 1 second. The green light time is 1 second, and this part of the green light time needs to go through 1 second. A green light phase, therefore, individuals will queue up to park. Next. When an individual arrives at the intersection during the red light period, that is... At that time, the vehicles in front of each individual also need to be emptied. If the green light time is 1 second, then the time required for an individual to cross the intersection is 1 second. The green light time is 1 second, and this part of the green light time needs to go through 1 second. A green light phase, therefore, individual parking queues Second-rate.

[0135] Queue length Number of parkings It is known that the queuing time of an individual can be constructed. The model construction and queuing time calculation model expression is as follows:

[0136]

[0137] in, Queuing time; Indicates the number of times an individual vehicle stops; This refers to the number of vehicles queuing ahead of an individual vehicle when it arrives at the downstream intersection. The arrival time of the signal; This represents the green light time required for an individual vehicle to wait for the queue ahead to disperse and pass through the stop line. The red light time represents the time during which a vehicle arrives at the downstream intersection when the light is on (i.e., the time during which the light is on). At that time, individual vehicles need to wait. A vehicle must complete a full red light to pass through the intersection; when an individual vehicle arrives at the intersection during the red light period, that is... At that time, the red light duration for that phase is still... Individual vehicles still need to wait. - You need to wait for one full red light to pass through the intersection.

[0138] Finally, a schematic diagram of the travel-time relationship function can be found in [reference needed]. Figure 6 .

[0139] This embodiment considers the impact of traffic signals and traffic density on individual travel time and constructs a travel time relationship function. This function, through step-by-step reasoning and calculation, sequentially connects the entire process of an individual vehicle entering the study section, its driving patterns within the section, queuing, and passing through intersections, thus characterizing the spatiotemporal distribution of individual vehicles within the section. Furthermore, considering the significant differences between individuals, the absolute overtaking volume during the queue length calculation is taken into account. An absolute overtaking volume calculation model is constructed based on a BP neural network, and individual driving habits are calibrated using historical overtaking habits, effectively reducing the randomness and variability of individual travel.

[0140] Example 2: Let a certain signal control method be G. The expression for the average travel time of all individual vehicles passing through the intersection under a certain signal control scheme is:

[0141]

[0142] in, The intersection signal control scheme is as follows: At that time, the first person passing through the intersection The travel time of each individual vehicle through the research road segment; V This indicates the total number of vehicles passing through the intersection.

[0143] In this embodiment, the goal of constructing the intersection signal control optimization model is to adjust and optimize the green light duration of each phase under the premise that the intersection signal cycle duration and phase sequence remain unchanged, so as to minimize the average travel time of all vehicles exiting the intersection.

[0144] The travel time is one of the indicators for evaluating the quality of intersection signal control schemes. The shorter the travel time, the better the signal control scheme. Therefore, the travel time of each vehicle is calculated using a travel time calculation function, and the goal of optimizing the timing scheme is to minimize its average value.

[0145] In the process of constructing the travel time relationship function, the calculation of green light time, number of stops and queuing time during the journey fully reflects the effect of the intersection signal control scheme on travel time.

[0146] In this embodiment, the intersection signal control optimization model is as follows:

[0147] Objective function:

[0148]

[0149] in, This represents the signal control scheme for an intersection, where the decision variables are the green light duration for each phase. = , This represents the number of phases at the intersection. For the first Green light duration for each phase;

[0150] In this embodiment, the model's constraints include the longest and shortest green light times for each phase, the total green light time constraint, and the intersection traffic congestion constraint. The longest and shortest green light constraint ensures pedestrian safety in each phase; the total green light time constraint ensures the sum of the green and yellow light times for each phase equals the cycle length; and the intersection traffic congestion constraint prevents severe traffic congestion in a particular phase, leading to queue overflow. The constraints include:

[0151] 1) Longest and shortest green light time constraints:

[0152]

[0153] in, To determine the shortest possible green light duration allowed by the phase, and considering pedestrian safety, through Sure, The width of the pedestrian crossing The speed at which pedestrians walk. Yellow light duration The time is when all red is visible; The longest green light duration allowed for a given phase, ensuring that the traffic capacity is appropriate for the cycle length and other phases of traffic;

[0154] 2) Total green light duration constraint:

[0155]

[0156] in, This represents the total duration of the yellow light within the cycle. The cycle duration of the intersection signal control lights.

[0157] 3) Intersection traffic congestion constraints:

[0158]

[0159] in, This represents the maximum queuing time index. For a certain time interval, the first Queue time index at the entrance lane; For a certain time interval, the first i Queuing time at intersections for vehicles entering the lane; The signal control cycle duration is the maximum queuing time exponent. When the value range is [1.5, 2.1), the traffic congestion level of the intersection is moderate. If the traffic congestion level of the intersection is less than moderate congestion, the green light duration of the phase is constrained, and the expression is:

[0160] ( <1.5.

[0161] In this embodiment, a genetic algorithm is used to solve the intersection signal control optimization model. The solution process is described in [link to solution details]. Figure 7 The process is as follows:

[0162] Step 1: Genetic parameter initialization. Determine the number of generations, population size, crossover probability, and mutation probability.

[0163] Step 2: Initialize the population. This invention uses real number encoding, with each chromosome being a real number. Vector real number encoding does not require numerical conversion and can be directly applied to the genetic algorithm based on the phenotype of the solution. The original signal control scheme is used as one initial chromosome, and other chromosomes are randomly generated based on the longest and shortest green light constraints and the total green light duration constraints.

[0164] Step 3: Calculate the fitness of each chromosome. This invention finds the minimum value of a function and uses the reciprocal of the function value as the fitness value of an individual, as shown in the formula below. The smaller the function value, the larger the fitness value of the individual, and the better the individual.

[0165] F=

[0166] Step 4: Determine if the number of iterations has reached the required number of generations. If yes, proceed to Step 8; otherwise, proceed to Step 5.

[0167] Step 5: Selection Operation. The probability of an individual being selected is related to its fitness value; the higher the fitness value, the greater the probability of selection. The probability of individual i being selected is shown in the following formula. We will use a roulette wheel selection algorithm for the selection operation.

[0168]

[0169] In the formula, For individuals fitness value; The number of individuals in the population.

[0170] Step 6: Crossover Operation. Since individuals are encoded with real numbers, the crossover operation uses the real-number crossover method. Chromosomes and the Chromosomes The cross operation method is shown in the following formula.

[0171]

[0172] in, It is a random number in the interval [0,1].

[0173] Step 7: Mutation operation. The first individual One gene The procedure for performing mutations is shown in the following formula. Return to Step 3 after calculation.

[0174]

[0175] in, It's a gene. The upper bound; It's a gene. The lower bound; , It is a random number. It is the current iteration number. It is the maximum number of evolutions. It is a random number in the interval [0,1].

[0176] Step 8: Output the optimal signal timing scheme, and the algorithm terminates.

[0177] This embodiment associates the travel time relationship function with the intersection signal control scheme, applies the established travel time relationship function to the intersection signal control optimization, and proposes an intersection signal control optimization model based on the travel time relationship function. At the same time, the model solution and verification process are both based on the travel time relationship function, without the need for a simulation platform.

[0178] Example 3

[0179] This embodiment is illustrated with reference to the actual experimental process. For example:

[0180] (1) Absolute overtaking amount based on BP neural network

[0181] In this embodiment, the BP neural network model uses a 3-layer network, with 5 input nodes in the input layer, 1 output node in the output layer, and 9 hidden nodes. The learning efficiency is initially set at 0.2 and adjusted experimentally. The sigmoid function is used as the activation function. The number of iterations is initially determined to be 1000, and adjusted experimentally based on the principle that the cost function value basically converges. In the calculation results of the absolute overtaking amount model based on the BP neural network for individual absolute overtaking amount, the MAE and RMSE are 5.93veh / km and 8.54veh / km, respectively. The frequency distribution diagram of the calculation error is shown in [Figure showing the error distribution]. Figure 8 .from Figure 8 It can be seen that the calculation error is within an acceptable range. Therefore, the calculation effect of the absolute overtaking amount calculation model based on BP neural network meets the calculation requirements.

[0182] (2) Travel time relationship function

[0183] The travel time calculation function of this invention is compared with the comparison function lv to analyze and compare the calculation effects of the two functions. In the process of modeling travel time, the issue of adjusting parameters arises, mainly adjusting the average interval time for individual vehicles to cross the stop line during the green light period. Meanwhile, for the travel time calculation function of the present invention, the following sections calculate the operation effects of three functions: one without considering overtaking, one considering overtaking, one considering overtaking, and one adjusting the ht value, and compare the functions.

[0184] The calculation results of each function are shown in Table 1.

[0185] Table 1

[0186]

[0187] The results show that the travel time calculation function proposed in this invention has a smaller error. Considering individual overtaking habits can effectively improve the function's accuracy. Compared with the calculation effect without considering individual overtaking behavior, the travel time calculation effect considering individual overtaking behavior is significantly improved. MAE, MAPE, and RSME are reduced from 12.61, 14.84%, and 28.06 to 10.44, 10.48%, and 25.98, respectively. After adjusting the value, the accuracy is further improved; the accuracy of the function in this invention is superior to the comparative function lv, and all indicators show greater advantages. (Lv function, function without overtaking consideration, function with overtaking consideration, function considering overtaking and adjustment) The evaluation indicators of the value function all show a decreasing trend, indicating that the proposed method is reasonable. Under these conditions, the travel time calculation results considering overtaking behavior are closer to the actual value.

[0188] (3) Signal control optimization model

[0189] This embodiment takes the intersection of Aofeng Road and Zhuangyuan Road in Xuancheng City as the optimization object, optimizes the signal control scheme based on the genetic algorithm, and verifies the optimization effect of the signal control scheme through simulation experiments. At the same time, the feasibility and superiority of the optimized signal control scheme are verified by comparing the original signal control scheme with the Webster scheme.

[0190] Table 2 shows the signal timing schemes for the original scheme, the Webster scheme, and the optimized scheme, as well as the average travel time of the intersection under each scheme.

[0191] Table 2

[0192]

[0193] The original solution outperformed the Webster solution, while the optimized solution showed the best application effect. Compared with the original solution, the optimized solution proposed in this application reduced the average travel time by 17.12% and the average queue length by a significant 23.25%, greatly reducing the delays caused by intersection traffic lights. Compared with the Webster solution, the average travel time was reduced by 24.11%. The optimized signal control solution can effectively reduce the total delay at the intersection.

[0194] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for intersection signal control optimization based on travel time relationship function, characterized in that, The method comprises the following steps: S1: constructing a travel time calculation model and a queuing time calculation model of a signal control section vehicle individual, and forming a vehicle individual travel time relationship function based on the travel time calculation model and the queuing time calculation model; S2: solving the average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme according to the vehicle individual travel time relationship function; S3: taking the minimum average travel time of all individual vehicles passing through the intersection under a certain set signal control scheme as an objective function, and taking the longest and shortest green light time of each phase, the total green light time constraint and the intersection traffic congestion degree constraint as constraint conditions, to construct an intersection signal control optimization model; The intersection signal control optimization model in step S3 is: Objective function: in, This represents the signal control scheme for an intersection, where the decision variables are the green light duration for each phase. = , This represents the number of phases at the intersection. For the first Green light duration for each phase; Constraint condition: 1) Longest and shortest green light time constraint: wherein, is the shortest green light duration allowed for the phase, from the perspective of pedestrian safety, through determination, is the width of the pedestrian crossing, is the speed of the pedestrian walking, is the yellow light duration, is the all-red time; is the longest green light duration allowed for the phase, which ensures moderate traffic capacity with the cycle duration and other phase traffic. 2) Total green light time constraint: wherein, is the total length of yellow light in a cycle; is the cycle length of the signal control light at the intersection; 3) Intersection traffic congestion degree constraint: wherein, is the maximum queue time index; is the number of vehicles in the queue at the import lane; is the queue time index of the import lane; is the number of vehicles in the queue at the import lane; i is the queue time index of the import lane; is the signal control cycle length, the maximum queue time index When the value interval of the maximum queue time index is [1.5, 2.1), the traffic congestion degree of the intersection is moderate congestion, and the traffic congestion degree of the intersection is less than moderate congestion, the phase green time is constrained, and the expression is: （ ）<1.5； S4: solving the intersection signal control optimization model to output the optimal intersection signal timing result.

2. The method for intersection signal control optimization based on travel time relationship function according to claim 1, characterized in that, The travel time calculation model in step S1 is constructed by using a multiple linear regression model, and the process is: S101: Consider the length of the road segment and the influence of traffic density K , build a travel time calculation model, the expression is: wherein, is a length of a road section; is a traffic density, is a functional relationship describing a relationship between the traffic density and the travel time; is a coefficient of the travel time calculation model, i = 1, 2, 3; S102: Based on the individual historical driving data of vehicles on the signal-controlled section, the coefficients of the driving time calculation model are fitted in a linear fitting manner .

3. The method for intersection signal control optimization based on travel time relationship function according to claim 1, wherein, In step S1, let the number of cycles crossed by a vehicle during individual travel from an upstream intersection to a downstream intersection be n , >2, which refers to the number of downstream intersection signal cycles crossed in total during the time of travel on the study section before the individual reaches the signal intersection, and the calculation formula is as follows: wherein, is the time of arrival, which is the time when the vehicle reaches the first stop line of the downstream intersection or passes through the stop line without stopping, and the calculation expression is: is the signal cycle length; calculating the green time for the first cycle during which the individual vehicle is driving , the expression being wherein, is the green light duration in a unit cycle of the signal light; is the entering signal time, which refers to the time duration when the vehicle enters the study section and the green light of the signal light at the downstream intersection of the section corresponding to the vehicle's driving direction has already started; when the vehicle individual enters the signal control section and the signal light at the downstream intersection is in the green light state, then is the remaining green light duration in the cycle; when the vehicle individual enters the signal control section and the signal light at the downstream intersection is in the red light state, the remaining green light duration in the cycle is 0, then is 0. The green light time of the first n cycle during the individual driving of the vehicle is calculated , and the expression is: ； Remaining -2 cycles of green light time, i.e. 2 times the duration of the green light of the traffic light -2, the duration of the green light during the journey of the individual is: In the formula, is the green light time during individual travel.

4. The method for intersection signal control optimization based on travel time relationship function according to claim 3, characterized in that, The number of vehicles queuing in front of the downstream intersection when the vehicle individual arrives is calculated, and the process is: Based on absolute overtaking data of each individual vehicle, average value, median, overtaking probability and overtaken probability of multiple absolute overtaking of the individual are calculated, the absolute overtaking data refers to the difference between the number of overtaking the front vehicle and the number of being overtaken by the rear vehicle during the individual driving in the lane within the unit road length, the calculation expression is: wherein, is the number of overtaking the front vehicle during the individual driving within the unit road length; is the number of being overtaken by the rear vehicle during the individual driving within the unit road length; Based on the BP neural network model, the probability of overtaking, the probability of being overtaken, the average value of the absolute overtaking amount, the median and the traffic density are used as the input characteristic variables of the BP neural network model, and the absolute overtaking amount of the vehicle individual is output P ; Solving the number of queued vehicles in front of a downstream intersection when an individual arrives , the expression is: wherein, q represents the number of vehicles queuing in front of the vehicle when it arrives at the downstream intersection; represents the number of vehicles in the same direction as the individual when the individual enters the study link; represents the number of lanes in the same direction as the individual when the individual exits the link; represents the average time interval between vehicles passing the stop line during the green light.

5. The method for intersection signal control optimization based on travel time relationship function according to claim 4, characterized in that, The number of vehicle individual parking times is calculated according to the number of vehicles queuing in front of the downstream intersection when the vehicle individual arrives and the traffic capacity of the intersection, and the calculation expression is: wherein, represents the number of individual stops of the vehicle.

6. The method for intersection signal control optimization based on travel time relationship function according to claim 5, wherein, The expression of the queuing time calculation model in step S1 is: wherein, is the queue time; represents the number of stops for the vehicle individual; is the number of vehicles in front of the vehicle individual when it arrives at the downstream intersection; is the arrival time of the signal; represents the green time needed for the vehicle individual to wait for the front queue to disperse and pass the stop line, represents the red time per cycle, when the vehicle individual arrives at the downstream intersection during the green time, i.e. , the vehicle individual needs to wait for a complete red light to pass the intersection; when the vehicle individual arrives at the intersection during the red time, i.e. , the red time of the phase is still , the vehicle individual still needs to wait for -1 complete red light to pass the intersection.

7. The method of intersection signal control optimization based on travel time relationship function according to claim 6, wherein, The expression of the vehicle individual travel time relationship function is: wherein, denotes the vehicle individual travel time relationship function.

8. The method for intersection signal control optimization based on travel time relationship function according to claim 7, wherein, In step S2, let a certain set signal control method be G The expression of the average travel time of all individual vehicles at the intersection under a certain set signal control scheme is wherein, represents the intersection signal control scheme is the first vehicle individual through studying the travel time of the road section; V represents the total number of vehicles passing through the intersection.

9. The method for intersection signal control optimization based on travel time relationship function of claim 1, wherein, In step S4, the genetic algorithm is used to solve the intersection signal control optimization model.

Images