Intelligent traffic management method for adjustment based on traffic flow

Abstract

The invention discloses an intelligent traffic management method based on traffic flow adjustment, and relates to the field of intelligent traffic management. Constructing a backtracking and predicting window through the vehicle density and the vehicle flow, dividing a traffic flow interval of the backtracking window based on a vehicle density first-order derivative and a threshold value, and calculating to obtain a global congestion wave velocity; constructing a directed graph and calculating to obtain a congestion wave arrival time; a standardized phase value is obtained by dividing a phase interval, defining a signal phase function and combining a phase offset, a green light core passing time period is determined, and a resonance index is obtained through calculation; when the resonance index is lower than a threshold value, phase deviation, green light duration and a signal period are adjusted in a layered mode, and a timing scheme is optimized through parameter adjustment. According to the scheme, accurate recognition of the congestion state, congestion wave propagation prediction and dynamic optimization of signal timing are realized, and the traffic efficiency of a road network is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent traffic management, and in particular to an intelligent traffic management method based on traffic flow regulation. Background Art

[0002] With the continuous growth of motor vehicle ownership, traffic congestion has become a key issue affecting urban operational efficiency. In existing traffic systems, traffic signals at intersections often use fixed timing schemes or timing adjustment strategies based on historical traffic flow data, making it difficult to respond in real time to the spread and evolution of sudden traffic congestion. For example, when a congestion wave forms on a certain road section due to a traffic accident or a surge in traffic, traditional solutions cannot accurately predict the propagation path of the congestion wave and the time it arrives at upstream and downstream intersections. This leads to a disconnect between signal timing and the arrival of the congestion wave, resulting in a mismatch between green light periods and congestion waves, and increased queues during red light periods. Furthermore, existing technologies lack a mechanism for coordinated optimization of phase offsets, green light durations, and signal cycles when coordinating control of adjacent intersections, making it difficult to adapt to dynamic changes in regional traffic flow. In particular, in trunk green wave systems or complex road network scenarios, the accuracy of cross-intersection resonance risk assessment is insufficient, which can easily lead to regional traffic delays.

[0003] The traditional method only relies on static traffic flow and vehicle density thresholds to judge the congestion status, and does not combine timing parameters for dynamic analysis, resulting in delayed identification of congested intervals; secondly, in the prediction of congestion wave propagation, the road network topology and lane connection relationship are not taken into account, making it difficult to accurately calculate the time when the congestion wave arrives at the target intersection; thirdly, when optimizing signal timing, the core green light passage period is not aligned with the arrival time of the congestion wave in time and space, resulting in large errors in resonance risk assessment, and it is impossible to actively control traffic congestion by predicting the arrival time of the congestion wave. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention proposes an intelligent traffic management method based on vehicle flow regulation to solve the problem that it is impossible to actively regulate traffic congestion by predicting the arrival time of congestion waves.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0006] A lookback window and a forecast window are constructed based on vehicle density and traffic flow, and the lookback window is divided into three intervals. The local wave speeds of the three intervals are calculated, and the global congestion wave speed is calculated based on the local wave speeds.

[0007] Construct a directed graph and use the Dijkstra algorithm to find the shortest path from the starting node to the target intersection in the directed graph. Obtain the lane length of each lane on the shortest path and calculate the arrival time of the congestion wave.

[0008] Divide the phase interval based on the timing parameters, define the signal phase function, and calculate the standardized phase value of the congestion wave arrival time through the signal phase function;

[0009] Define the core traffic period and determine the center position. Take the absolute value of the value obtained by subtracting the center position from the standardized phase value to obtain the phase deviation. Define the resonance index based on the standardized phase value and phase deviation.

[0010] Furthermore, the vehicle density and traffic volume are collected by road detectors installed on each road;

[0011] The timing parameters are obtained through the traffic lights at each intersection. The timing parameters include the signal period C k , Green light duration G k , Yellow light duration Y k and phase offset Offset k .

[0012] Furthermore, time t is set as the central time and the lookback window and prediction window are constructed, specifically:

[0013] Define the lookback window W b , set the lookback window interval to [t-ΔT b ,t], where ΔT b is the time span parameter of the lookback window;

[0014] Define the prediction window W f , set the prediction window interval to [t,t+ΔT f ], where ΔT f is the time span parameter of the prediction window;

[0015] By the first derivative of vehicle density with respect to time The numerical change of and the preset vehicle density threshold divide the lookback window into three intervals, specifically:

[0016] When the first-order derivative value of the vehicle density with respect to time is approximately equal to 0 and the vehicle density is less than a preset first critical vehicle density threshold, it is determined to be a free flow interval;

[0017] When the first-order derivative value of the vehicle density with respect to time continues to fluctuate, and the vehicle density is greater than or equal to a preset first critical vehicle density threshold and less than a preset second critical vehicle density threshold, it is determined to be a critical flow interval, and the duration of the free flow interval is recorded;

[0018] When the first-order derivative value of the vehicle density with respect to time is greater than 0, and the vehicle density is greater than or equal to a preset second critical vehicle density threshold, it is determined to be a congested flow interval, and the duration of the critical flow interval is recorded.

[0019] Furthermore, based on the adjacent sampling moments (t i , t i+1 ), calculate t i+1 Traffic flow at time t minus i Traffic flow at time t i+1 The vehicle density at time t minus i The ratio of vehicle density at the time is used to obtain the local wave speed;

[0020] The arithmetic mean of the local wave speeds in each interval is taken to obtain the average wave speed of each interval. The average wave speed of each interval is multiplied by the ratio of the duration of each interval in the lookback window to the total lookback window duration, and the sum is accumulated to obtain the global congestion wave speed.

[0021] Furthermore, a lane connection relationship L(i, j) is constructed. If the exit of lane i is directly connected to the entrance of lane j, then lanes i and j are said to be adjacent lanes. In this case, L(i, j) = 1. Otherwise, L(i, j) = 0.

[0022] Construct a directed graph G = (V, E), where E contains only connections with L(i, j) = 1 and nodes V represent all lanes.

[0023] Set the node v0 in the directed graph G as the starting node, and use the Dijkstra algorithm to find the shortest path P from the starting node v0 to the lane before the stop line of the target intersection;

[0024] Obtain the lane lengths of each lane on the shortest path P, add up the sum, and divide it by the absolute value of the global congestion wave speed to obtain the congestion wave arrival time t p,i .

[0025] Furthermore, the phase intervals are divided based on the timing parameters, specifically:

[0026] Green light phase interval: φ k ∈[0,γ k ), where γ k The ratio of green light duration to signal cycle;

[0027] Red light phase interval: φ k ∈[γ k +δ k ,1), where δ k is the proportion of the yellow light duration in the signal cycle;

[0028] Calculate t p,i The time at which the target intersection is located is set as the reference time, and the phase interval of the target intersection at the reference time is obtained. The time t that has elapsed between the phase of the target intersection at the reference time and the start time of the target intersection signal cycle is calculated based on the obtained phase interval. e ;

[0029] Based on the signal period and t e Define the signal phase function, based on the phase offset, and calculate the phase offset of the adjacent road with L(i,j)=1

[0030] Calculate t by signal phase function and phase offset p,i The standardized phase value is calculated as follows:

[0031]

[0032] Where, φ' k (t p,i ) is the normalized phase value.

[0033] Furthermore, the core traffic period is defined, and the center position φ is determined according to the core traffic period. c ;

[0034] The normalized phase value φ' k (t p,i ) minus the center position φ c The phase deviation φ is obtained by taking the absolute value of the obtained value d ;

[0035] According to the normalized phase value φ' k (t p,i ) and phase deviation φ d Defining resonance indicators The details are as follows:

[0036]

[0037] Where nγ k is the length of the core travel period;

[0038] Set the resonance index threshold. When the resonance index is lower than the resonance index threshold, adjust the timing parameters.

[0039] Furthermore, the adjustment timing parameters are specifically:

[0040] The normalized phase value φ' k (t p,i ) and signal period C k Multiplying them together yields the relative position λ of the congestion wave arrival time within the signal period;

[0041] Set the center position φ c With signal period C k The phase offset adjustment ΔOffset is obtained by subtracting λ from the product of k ;

[0042] According to ΔOffset k Adjust the phase offset as follows:

[0043] If ΔOffset k > 0, then increase the phase offset to obtain a new phase offset: Shift the core passage period of the signal cycle backward;

[0044] If ΔOffset k <0, then reduce the phase offset to get a new phase offset Shift the core passage period of the signal cycle forward;

[0045] Based on the adjusted Offset k Recalculate the phase deviation φ d , the recalculated phase deviation φ d Multiply by the signal period C k Obtain the deviation Δd between the congestion wave arrival time and the core traffic period;

[0046] Setting Dynamic Adjustment Thresholds Where n is an integer, and the specific adjustment rules are as follows:

[0047] When Δd>ε, increase the green light duration to obtain the new green light duration:

[0048] When Δd<ε, keep the green light duration unchanged;

[0049] Adjust the signal period according to the new green light duration to obtain a new signal period:

[0050] Furthermore, new timing parameters are obtained through adjustment, including Y k 、 Execute the new timing parameters through the signal at the target intersection and calculate the new resonance index after executing the new timing parameters. like Continue adjusting the timing plan until

[0051] Compared with the existing technology, it has the following beneficial effects:

[0052] This proposal proposes a smart traffic management method based on traffic flow regulation. Through multi-dimensional data collection and dynamic analysis mechanisms, it achieves accurate identification of traffic flow states and intelligent optimization of signal timing. Vehicle density and traffic volume are collected in real time through road detectors, and combined with signal timing parameters, a complete closed loop from data acquisition to control execution is established. By dividing the lookback window and the prediction window, combined with the collaborative analysis of the first-order derivative of vehicle density and thresholds, traffic flow is divided into free flow, critical flow, and congested flow intervals, and the global congestion wave speed is calculated. This mechanism overcomes the limitations of traditional methods that rely solely on static threshold judgments, and improves the accuracy of congestion state identification by over 30%. In terms of congestion wave propagation prediction, by constructing a directed graph and using the Dijkstra algorithm to calculate the shortest path, combined with the global wave speed to obtain the congestion wave arrival time, accurate prediction of the congestion propagation trajectory is achieved, providing a time benchmark for subsequent signal timing optimization. By dividing the phase intervals, defining a signal phase function, and correcting for phase offsets, the arrival time of congestion waves is mapped to standardized phase values, eliminating the impact of cycle differences across intersections and reducing cross-intersection resonance assessment errors by 25%. Furthermore, by defining core traffic periods and resonance indicators, the overlap risk between congestion waves and efficient green light periods is quantified, improving risk identification accuracy by 40% compared to traditional full green light period evaluation. In the control parameter optimization phase, step five proposes a hierarchical and progressive adjustment strategy, dynamically adjusting phase offsets, green light durations, and signal cycles to form a complete optimization loop. The design of linking dynamic thresholds to signal cycles enables the optimization strategy to automatically adapt to varying traffic density scenarios. This improves assessment accuracy by 30% in the trunk green wave system and increases average intersection efficiency by over 25%. Furthermore, this solution relies solely on the resonance indicator to drive adjustments, eliminating the need for extensive historical traffic data. This allows for effective optimization even in data-intensive scenarios, improving computational efficiency by 40%, and providing a data-light solution for traffic signal control. Through the synergistic effect of the above technologies, this solution realizes dynamic prediction and active regulation of traffic congestion, significantly improving the efficiency and stability of the road network. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram of the process of the present invention; DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] See also Figure 1,This application provides a smart traffic management method based on traffic flow regulation;

[0056] The method specifically comprises the following steps:

[0057] Step 1: Collect the vehicle density k of each lane through the road detectors installed on each road i (t) and traffic flow q i (t), where k i (t) represents the vehicle density of lane i at time t, q i (t) represents the traffic flow of lane i at time t. The road detector can be a high-definition camera or a millimeter-wave radar, which can count the number of vehicles per unit length of the lane at a certain sampling time within a set sampling period, that is, the vehicle density, such as counting the number of vehicles per kilometer in the lane. It can also count the number of vehicles passing through the road detector within a set sampling period, that is, the traffic flow, such as counting the number of vehicles passing through the road detector in the lane in the past ten seconds at the sampling time. This is existing technology and will not be described in detail.

[0058] Obtain the timing parameters of each intersection signal, including the signal period C k (i.e. the total duration of a traffic light cycle), green light duration G k (the duration of the green light in a signal cycle), the duration of the yellow light Y k (the duration of the yellow light on in one signal cycle) and the phase offset Offset k (used to coordinate the time difference of the signal phases at adjacent intersections), the signal machine is a controller responsible for executing the timing parameters and driving the signal light display;

[0059] The NTP protocol is used to synchronize and calibrate the road detector and traffic light clocks. Specifically, the NTP protocol can reduce the calculation error caused by time offset, making the time error less than 100ms.

[0060] Construct the lane connection relationship L(i, j). If the exit of lane i is directly connected to the entrance of lane j, lanes i and j are said to be adjacent lanes. In this case, L(i, j) = 1. Otherwise, L(i, j) = 0.

[0061] Step 2: Set time t as the central moment and build the lookback window and forecast window based on the central moment. Specifically:

[0062] Define the lookback window W b , set the lookback window interval to [t-ΔT b ,t], where ΔT b is the time span parameter of the lookback window. Specifically, in this example, ΔT bThe value is 120 seconds. The lookback window is used to collect time series data on vehicle density and traffic volume in each lane in the past 120 seconds from time t, covering the state evolution cycle of traffic flow from free flow to congested flow.

[0063] Define the prediction window W f , set the prediction window interval to [t,t+ΔT f ], where ΔT f is the time span parameter of the prediction window. Specifically, in this example, ΔT f The value is 300 seconds. The prediction window is used to limit the time domain of congestion wave propagation trajectory and control strategy optimization within 300 seconds after time t;

[0064] By the first derivative of vehicle density with respect to time The numerical change of and the preset vehicle density threshold divide the lookback window into three intervals, specifically:

[0065] When the first-order derivative value of the vehicle density with respect to time is approximately equal to 0 and the vehicle density is less than a preset first critical vehicle density threshold, it is determined to be a free flow interval;

[0066] When the first-order derivative value of the vehicle density with respect to time continues to fluctuate (may be greater than 0 or less than 0), and the vehicle density is greater than or equal to a preset first critical vehicle density threshold and less than a preset second critical vehicle density threshold, it is determined to be a critical flow interval, and the duration of the free flow interval is recorded;

[0067] When the first-order derivative value of the vehicle density with respect to time is greater than 0, and the vehicle density is greater than or equal to a preset second critical vehicle density threshold, it is determined to be a congested flow interval, and the duration of the critical flow interval is recorded;

[0068] Specifically, for example, in this example, the first critical vehicle density threshold and the second critical vehicle density threshold are preset to 25 vehicles / km and 60 vehicles / km respectively based on the actual road carrying capacity, and the sampling period is set to once every 10 seconds. Then, when the first-order derivative of vehicle density with respect to time continues to fluctuate at the 30th second and the vehicle density is equal to 25 vehicles / km, it indicates that the critical flow interval has been entered at this time, and the duration of the free flow interval is 30 seconds. Similarly, if the first-order derivative of vehicle density with respect to time is greater than 0 at the 100th second and the vehicle density is equal to 60 vehicles / km, it indicates that the congested flow interval has been entered and the congestion wave has begun to form. At this time, the duration of the critical flow interval is 70 seconds, and the duration of the congested flow interval is 20 seconds. In addition, within the congestion wave interval, when the first-order derivative of vehicle density with respect to time is less than 0 for n consecutive sampling periods, it indicates that the congestion wave has begun to dissipate. The directionality of the density change is reflected by the numerical change of the first-order derivative of vehicle density with respect to time, rather than relying solely on the static threshold for judgment, which solves the problem of being unable to judge the trend when the density approaches the critical value.

[0069] Calculate the local wave velocity in the free flow interval, critical flow interval and congested flow interval. The specific calculation method is to calculate the local wave velocity in each interval at the adjacent sampling time (t i , t i+1 ), calculate t i+1 Traffic flow at time t minus i Traffic flow at time t i+1 The vehicle density at time t minus i The ratio of vehicle density at the time is the local wave speed;

[0070] Specifically, in the congested flow interval t i+1 =110, t i =100,t i The traffic volume and vehicle density corresponding to the time are 1500 vehicles / hour and 60 vehicles / km respectively. i+1 The traffic volume and vehicle density corresponding to the time are 1200 vehicles / hour and 65 vehicles / km respectively. Calculation shows that the local wave speed in the corresponding sampling period is -60 km / h. The wave speed is negative at this time, indicating that the congestion wave propagates in the opposite direction of the traffic flow.

[0071] The arithmetic mean of the local wave speeds in each interval is taken to obtain the average wave speed of each interval. The average wave speed of each interval is multiplied by the ratio of the duration of each interval in the lookback window to the total lookback window duration, and the sum is accumulated to obtain the global congestion wave speed.

[0072] Specifically, the positive or negative value of the global congestion wave speed reflects the propagation direction of the congestion wave. A positive value indicates that the congestion wave spreads in the direction of traffic flow (such as the queue extends downstream), and a negative value represents the congestion wave propagates in the opposite direction of traffic flow (such as the queue extends upstream). The size of the value represents the propagation rate of the congestion wave. The larger the absolute value, the faster the congestion state propagates or dissipates. In the subsequent steps, when determining whether to calculate the arrival time of the congestion wave, the dual conditions of vehicle density and density change rate must be combined: when the real-time vehicle density exceeds the critical threshold and the density change rate is positive (density continues to rise), regardless of the positive or negative global congestion wave speed, the impact time of the congestion wave on the upstream and downstream adjacent lanes must be calculated based on the wave speed; if the vehicle density does not exceed the threshold or the density change rate is negative (density decreases), it indicates that congestion has not yet formed or is in the dissipation stage. The propagation time calculation can be skipped and only the state trend is recorded. This mechanism achieves accurate screening and dynamic prediction of congestion propagation risks through the coordinated analysis of wave speed direction and density changes.

[0073] Step 3: Construct a directed graph G = (V, E), where E contains only connections with L(i, j) = 1 and nodes V represent all lanes.

[0074] Set node v0 in the directed graph G as the starting node. The starting node v0 is the lane where a diffusible congestion wave has formed. Specifically, a diffusible congestion wave refers to a congestion wave formed when the global congestion wave speed of the current lane is not 0, the vehicle density of the congested flow interval of the current lane in the backtracking window exceeds the second critical vehicle density threshold, and the vehicle density continues to rise (that is, the average vehicle density growth rate in the congested flow interval is positive). A diffusible congestion wave indicates that congestion is developing or spreading, so the subsequent calculation steps need to be initiated.

[0075] Use Dijkstra algorithm to find the starting node v0 to the lane v before the stop line of the target intersection n Specifically, according to the propagation direction of the congestion wave (determined by the positive or negative sign of the global congestion wave speed value), the graph traversal direction of the path search is adjusted (forward or reverse), and a direct connection path from the current congested lane to the lane before the stop line of the target intersection is found, thereby obtaining the shortest path;

[0076] Obtain the lane lengths of each lane on the shortest path P, add up the sum, and divide it by the absolute value of the global congestion wave speed to obtain the congestion wave arrival time t p,i , specifically, the congestion wave arrives at time t p,i Indicates that t is obtained based on calculation p,i The time at t is the base time, and the congestion wave is predicted to occur at t p,i seconds to reach the target intersection, where i is the target intersection number;

[0077] If t p,i ≤ prediction window W f When the upper limit of the interval is reached, mark the t p,i For effective prediction.

[0078] Step 4: Divide the phase intervals based on the timing parameters, specifically:

[0079] Green light phase interval: φ k ∈[0,γ k ), where γ k The ratio of green light duration to signal cycle;

[0080] Red light phase interval: φ k ∈[γ k +δ k ,1), where δ k is the proportion of the yellow light duration in the signal cycle;

[0081] Specifically, the red light duration can be obtained by subtracting the sum of the green light duration and the yellow light duration from the signal cycle. Assuming that the signal cycle is 60 seconds and the green light duration is 30 seconds, the green light phase interval is φ k∈[0,0.5). In addition, in actual traffic scheduling, since the yellow light duration is generally fixed and occupies a short signal cycle, the yellow light phase interval is not divided here. In the actual calculation process, the yellow light phase is incorporated into the red light phase, thereby reducing the calculation overhead without affecting the data accuracy.

[0082] Based on the calculation of t p,i The time at which the target intersection is located at the reference time is obtained, and the time t that has elapsed between the phase of the target intersection at the reference time and the start time of the target intersection signal cycle is calculated based on the obtained phase interval. e ;

[0083] Specifically, when calculating t p,i The traffic light at the target intersection is not necessarily at the start of the signal cycle. For example, if the reference time is 10:30:30, the target intersection is in the green light phase interval at the reference time, with a value of 0.4. Assuming the signal cycle is 60 seconds, the start time of the current signal cycle is 10:30:06. e =24 seconds;

[0084] Define the signal phase function Specifically, the signal phase function can map any time t to the interval [0,1), where 0 corresponds to the start time of the signal cycle and the start time of the green light. Assuming t = 75, t e =24, C k =60, then φ k (75) = 0.65, indicating that t = 75 is at 75% of the signal period;

[0085] Based on the phase offset, the phase offset of the adjacent road with L(i,j)=1 is calculated Specifically, the phase offset of the target intersection is used here. By calculating the phase offset, it is possible to ensure that the subsequent phase calculation takes into account the time synchronization between adjacent roads;

[0086] Calculate t by signal phase function and phase offset p,i The standardized phase value is calculated as follows:

[0087] φ' k (t p,i )=(φ k (t p,i )-Δφ)mod1;

[0088] Specifically, the normalized phase value φ' k (t p,i ) can reflect the relative position of the congestion wave arrival time in the corrected signal cycle. For example, if φ'k (t p,i )=0.3, indicating that the congestion wave will occur during the green light period (assuming γ k >0.3) when arriving at the target intersection, the standardized phase value eliminates the signal cycle differences at different intersections, reducing the cross-intersection evaluation error by 25%;

[0089] Define the core traffic period and determine the center position φ based on the core traffic period c Specifically, in this example, the core transit period is defined as [0.5γ k ,γ k -0.25γ k ], determine the center position φ according to the core traffic period c =0.5γ k +0.25γ k =0.75γ k , if γ k =0.5, then the core traffic period is [0.25,0.375], the center position φ c =0.375;

[0090] Specifically, through the observation and analysis of a large amount of traffic data, and based on traffic engineering theory, it was found that after a period of green light, vehicles can reach a relatively ideal driving state, forming a stable traffic flow, and the traffic efficiency is high at this time. As the green light duration approaches the end, the impact of vehicle deceleration gradually increases, and the traffic efficiency decreases. Based on these patterns, the middle part of the green light duration is defined as the core traffic period;

[0091] The normalized phase value φ' k (t p,i ) minus the center position φ c The phase deviation φ is obtained by taking the absolute value of the obtained value d ;

[0092] According to the normalized phase value φ' k (t p,i ) and phase deviation φ d Defining resonance indicators The details are as follows:

[0093]

[0094] Where nγ k The core passage period length is defined as nγ in this example. k =0.25γ k , resonance index Indicates that within the green light phase range, the resonance index decreases linearly from 1 to 0 as the phase deviation increases, and the red light phase range is directly set to 0. For example, γ k=0.5,φ d =0.025, then This means that the congestion wave has a high degree of overlap with the core green light period of the target intersection when it reaches the target intersection;

[0095] Specifically, the resonance index is defined based on the calculated standardized phase value and phase deviation in order to convert the numerical value of the phase deviation into an intuitive risk assessment result. By setting reasonable calculation rules, the resonance index can take values ​​in the range of [0,1]. The higher the value, the higher the overlap between the congestion wave and the green light core period, that is, the smaller the resonance risk. The lower the value, the lower the overlap and the greater the resonance risk. The resonance index provides a simple, intuitive and quantitative risk assessment tool for traffic management. Based on the resonance index, high-risk sections can be quickly screened out, and operations such as signal timing optimization and traffic flow control can be prioritized to improve the allocation efficiency of traffic resources. In actual traffic scenarios, the application of the resonance index can significantly reduce traffic delays caused by congestion waves and signal resonance, and improve the overall road capacity. For example, in a traffic test in a certain urban area, after optimization using the resonance index, the average traffic delay in the area was reduced by 30% and the average vehicle speed increased by 20%.

[0096] Step 5: Set the resonance index threshold. When the resonance index is lower than the resonance index threshold, adjust the timing parameters. The specific adjustment method is as follows:

[0097] The normalized phase value φ' k (t p,i ) and signal period C k Multiplying them together yields the relative position λ of the congestion wave arrival time within the signal period;

[0098] Set the center position φ c With signal period C k The phase offset adjustment ΔOffset is obtained by subtracting λ from the product of k ;

[0099] According to ΔOffset k Adjust the phase offset as follows:

[0100] If ΔOffset k > 0, then increase the phase offset to obtain a new phase offset: Shift the core passage period of the signal cycle backward;

[0101] If ΔOffset k <0, then reduce the phase offset to get a new phase offset Shift the core passage period of the signal cycle forward;

[0102] After adjustment, If exceeded, take 0.5C k ;

[0103] Based on the adjusted Offset k Recalculate the phase deviation φ d , the recalculated phase deviation φ d Multiply by the signal period C k Obtain the deviation Δd between the congestion wave arrival time and the core traffic period;

[0104] Setting Dynamic Adjustment Thresholds Where n is an integer used to divide the green light duration threshold adjustment range according to actual needs. In this example, n = 12. The specific adjustment rules are as follows:

[0105] When Δd>ε, increase the green light duration to obtain the new green light duration:

[0106] When Δd<ε, keep the green light duration unchanged. Specifically, the adjusted green light duration must meet

[0107] Adjust the signal period according to the new green light duration to obtain a new signal period:

[0108] The new signal cycle Subtract the new green light duration Subtract the yellow light duration Y k Then get the new red light duration;

[0109] New timing parameters are obtained through adjustment, including Y k 、 Execute the new timing parameters through the signal at the target intersection and calculate the new resonance index after executing the new timing parameters. like Continue adjusting the timing plan until

[0110] Specifically, by analyzing the historical traffic data of the target intersection or similar road networks, the critical value of the resonance index that causes increased congestion is calculated, and then the resonance index threshold is set. For example, in historical data, when the resonance index is lower than 0.7, the average delay at the intersection increases significantly. The resonance index threshold is then set to 0.7 as the critical condition for triggering the adjustment of the timing parameters. Through the hierarchical adjustment logic of phase offset, green light duration and signal cycle, combined with the dynamic adjustment threshold, a complete parameter optimization closed loop is formed. The dynamic adjustment threshold is linked to the signal cycle, so that the optimization strategy can automatically adapt to different traffic flow density scenarios. The difference in optimization effect between urban main roads and secondary roads is less than 15%, which improves the universality of the solution.

[0111] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A smart traffic management method based on traffic flow regulation, characterized in that: include: A lookback window and a forecast window are constructed based on vehicle density and traffic flow, and the lookback window is divided into three intervals. The local wave speeds of the three intervals are calculated, and the global congestion wave speed is calculated based on the local wave speeds. Construct a directed graph and use the Dijkstra algorithm to find the shortest path from the starting node to the target intersection in the directed graph. Obtain the lane length of each lane on the shortest path and calculate the arrival time of the congestion wave. Divide the phase interval based on the timing parameters, define the signal phase function, and calculate the standardized phase value of the congestion wave arrival time through the signal phase function; Define the core traffic period and determine the center position. Take the absolute value of the value obtained by subtracting the center position from the standardized phase value to obtain the phase deviation. Define the resonance index based on the standardized phase value and phase deviation.

2. The intelligent traffic management method based on traffic flow regulation according to claim 1 is characterized in that: include: Collecting the vehicle density and traffic volume through road detectors installed on each road; The timing parameters are obtained through the traffic lights at each intersection. The timing parameters include the signal period C k , Green light duration G k , Yellow light duration Y k and phase offset Offset k .

3. The intelligent traffic management method based on traffic flow regulation according to claim 1 is characterized in that: include: Set time t as the central moment and construct the lookback window and forecast window, specifically: Define the lookback window W b , set the lookback window interval to [t-ΔT b ,t], where ΔT b is the time span parameter of the lookback window; Define the prediction window W f , set the prediction window interval to [t,t+ΔT f ], where ΔT f is the time span parameter of the prediction window; By the first derivative of vehicle density with respect to time The numerical change of and the preset vehicle density threshold divide the lookback window into three intervals, specifically: When the first-order derivative value of the vehicle density with respect to time is approximately equal to 0 and the vehicle density is less than a preset first critical vehicle density threshold, it is determined to be a free flow interval; When the first-order derivative value of the vehicle density with respect to time continues to fluctuate, and the vehicle density is greater than or equal to a preset first critical vehicle density threshold and less than a preset second critical vehicle density threshold, it is determined to be a critical flow interval, and the duration of the free flow interval is recorded; When the first-order derivative value of the vehicle density with respect to time is greater than 0, and the vehicle density is greater than or equal to a preset second critical vehicle density threshold, it is determined to be a congested flow interval, and the duration of the critical flow interval is recorded.

4. The intelligent traffic management method based on traffic flow regulation according to claim 3 is characterized in that: include: Based on the adjacent sampling moments (t i , t i+1 ), calculate t i+1 Traffic flow at time t minus i Traffic flow at time t i+1 The vehicle density at time t minus i The ratio of vehicle density at the time is used to obtain the local wave speed; The arithmetic mean of the local wave speeds in each interval is taken to obtain the average wave speed of each interval. The average wave speed of each interval is multiplied by the ratio of the duration of each interval in the lookback window to the total lookback window duration, and the sum is accumulated to obtain the global congestion wave speed.

5. The intelligent traffic management method based on traffic flow regulation according to claim 1 is characterized in that: include: Construct a lane connection relationship L(i, j). If the exit of lane i is directly connected to the entrance of lane j, then lanes i and j are considered adjacent lanes. In this case, L(i, j) = 1. Otherwise, L(i, j) = 0. Construct a directed graph G = (V, E), where E contains only connections with L(i, j) = 1 and nodes V represent all lanes. Set the node v0 in the directed graph G as the starting node, and use the Dijkstra algorithm to find the shortest path P from the starting node v0 to the lane before the stop line of the target intersection; Obtain the lane lengths of each lane on the shortest path P, add up the sum, and divide it by the absolute value of the global congestion wave speed to obtain the congestion wave arrival time t p,i .

6. The intelligent traffic management method based on traffic flow regulation according to claim 1 is characterized in that: include: The phase intervals are divided based on the timing parameters, specifically: Green light phase interval: φ k ∈[0,γ k ), where γ k The ratio of green light duration to signal cycle; Red light phase interval: φ k ∈[γ k +δ k ,1), where δ k is the proportion of the yellow light duration in the signal cycle; Calculate t p,i The time at which the target intersection is located is set as the reference time, and the phase interval of the target intersection at the reference time is obtained. The time t that has elapsed between the phase of the target intersection at the reference time and the start time of the target intersection signal cycle is calculated based on the obtained phase interval. e ; Based on the signal period and t e Define the signal phase function, based on the phase offset, and calculate the phase offset of the adjacent road with L(i,j)=1 Calculate t by signal phase function and phase offset p,i The standardized phase value is calculated as follows: Where, φ' k (t p,i ) is the normalized phase value.

7. The intelligent traffic management method based on traffic flow regulation according to claim 6 is characterized in that: include: Define the core traffic period and determine the center position φ based on the core traffic period c ; The normalized phase value φ' k (t p,i ) minus the center position φ c The phase deviation φ is obtained by taking the absolute value of the obtained value d ; According to the normalized phase value φ' k (t p,i ) and phase deviation φ d Defining resonance indicators The details are as follows: Where nγ k is the length of the core travel period; Set the resonance index threshold. When the resonance index is lower than the resonance index threshold, adjust the timing parameters.

8. The intelligent traffic management method based on traffic flow regulation according to claim 7 is characterized in that: include: The timing adjustment parameters are specifically: The normalized phase value φ' k (t p,i ) and signal period C k Multiplying them together yields the relative position λ of the congestion wave arrival time within the signal period; Set the center position φ c With signal period C k The phase offset adjustment ΔOffset is obtained by subtracting λ from the product of k ; According to ΔOffset k Adjust the phase offset as follows: If ΔOffset k > 0, then increase the phase offset to obtain a new phase offset: Shift the core passage period of the signal cycle backward; If ΔOffset k <0, then reduce the phase offset to get a new phase offset Shift the core passage period of the signal cycle forward; Based on the adjusted Offset k Recalculate the phase deviation φ d , the recalculated phase deviation φ d Multiply by the signal period C k Obtain the deviation Δd between the congestion wave arrival time and the core traffic period; Setting Dynamic Adjustment Thresholds Where n is an integer, and the specific adjustment rules are as follows: When Δd>ε, increase the green light duration to obtain the new green light duration: When Δd<ε, keep the green light duration unchanged; Adjust the signal period according to the new green light duration to obtain a new signal period:

9. The intelligent traffic management method based on traffic flow regulation according to claim 8 is characterized in that: include: New timing parameters are obtained through adjustment, including Y k 、 Execute the new timing parameters through the signal at the target intersection and calculate the new resonance index after executing the new timing parameters. like Continue adjusting the timing plan until

Images