A traffic problem diagnosis method for signal-controlled intersections on urban roads

Abstract

The present invention relates to a traffic problem diagnosis method for urban road signal-controlled intersections, including: starting from the two aspects of traffic status and control efficiency, determining the calculation results of various evaluation indicators; combining the traffic demand of the signal-controlled intersection, taking a single control period as the time granularity of problem diagnosis, taking the calculation results of the evaluation indicators as input, outputting a pre-optimization plan based on a signal control optimization model, and determining the reference intervals of various evaluable indicators, and conducting quantitative analysis on the operating status of various evaluable indicators; starting from the two aspects of unreasonable supply and unreasonable demand, constructing a traffic problem diagnosis model, constructing a corresponding relationship between problem diagnosis and operation evaluation based on the quantitative analysis results of the operating status of evaluable indicators, determining potential optimization targets, establishing problem-oriented control signal optimization, and realizing traffic problem diagnosis. Compared with the prior art, the present invention can accurately and autonomously diagnose the operating problems of signal-controlled intersections.

Description

Technical Field

[0001] The invention relates to the field of traffic information, and in particular to a traffic problem diagnosis method for a signal-controlled intersection on an urban road. Background Art

[0002] Traditional signal control optimization ideas are mostly based on specific evaluation indicators (maximum delay, queue length, number of stops, etc.). Experts analyze traffic problems and select corresponding control measures (signal control, traffic organization, channelization design, etc.) to optimize the system accordingly. The "problem diagnosis" link is mostly undertaken by control experience and expert knowledge. The existing traffic control system lacks state estimation to problem diagnosis, as well as problem-oriented control optimization, and cannot form an autonomous closed-loop mechanism.

[0003] In terms of traffic signal control problem diagnosis, domestic and foreign scholars have diagnosed the rationality of signal control parameters such as unreasonable coordination range, unreasonable coordination object, unreasonable phase difference setting, and unreasonable cycle length setting based on data directly collected by traffic detection equipment or operation evaluation parameters obtained by indirect calculation; and with the widespread application of new data sources such as Internet vehicle trajectory data and electronic police data, some scholars have directly conducted quantitative analysis on the operation quality of traffic flow and signal coordination effect in combination with actual application scenarios, so as to diagnose the problems existing in signal control and use it as the basis for signal control optimization; on this basis, visual diagnosis toolkits such as SPM system and PASS system have been further developed. In general, whether in academic research or engineering application, the diagnosis of traffic signal control problems has been initially explored, but there are still problems such as single type of diagnosis problems and lack of automation. Summary of the invention

[0004] The purpose of the present invention is to provide a traffic problem diagnosis method for signal-controlled intersections on urban roads. It improves the existing signal-controlled intersection management and control system, which lacks the functions from state estimation to problem diagnosis and problem-oriented control optimization, and cannot form an autonomous closed-loop mechanism. It enriches the types of diagnosed problems and can accurately and autonomously diagnose operating problems at signal-controlled intersections.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A traffic problem diagnosis method for a signal-controlled intersection on an urban road comprises the following steps:

[0007] Step 1) Starting from the traffic status and control benefits, an evaluation index system for urban road signal control intersections for traffic problem diagnosis is constructed, and cycle-level and lane-level calculations are performed based on the data collected by traffic detection equipment to obtain the calculation results of various evaluation indicators;

[0008] Step 2) In combination with the traffic demand of the analyzed signal-controlled intersection, a single control period is used as the time granularity of problem diagnosis, and the calculation result of the evaluation index is used as input, a pre-optimization scheme of the control signal is output based on the signal control optimization model, and a reference interval of various evaluable indicators is determined; based on the calculation result of the evaluation index and the reference interval, the operating status of various evaluable indicators is quantitatively analyzed, wherein the reference interval is a reasonable range where the indicator value should be when the traffic conditions are good, and the evaluable indicator is an evaluation indicator with a reference interval;

[0009] Step 3) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed. Based on the quantitative analysis results of the operating status of evaluable indicators, the correspondence between problem diagnosis and operation evaluation is constructed, potential optimization goals are determined, and problem-oriented control signal optimization is established to achieve traffic problem diagnosis.

[0010] The evaluation indicators related to traffic conditions include traffic volume, headway variation coefficient, maximum queue length, queue length when the green light is on, queue length when the red light is on, queue length growth rate, short lane queue overflow cycle ratio, and road section queue overflow cycle ratio.

[0011] The evaluation indicators related to control effectiveness include saturation, inter-phase demand imbalance index, intra-phase demand imbalance index, green light arrival ratio, average number of stops, proportion of vehicles queuing multiple times, average vehicle control delay, green light empty time ratio, green light utilization rate, queue clearing time ratio, phase supply and demand imbalance index, and intersection clearing remaining cycle ratio.

[0012] The step 2) specifically comprises the following steps:

[0013] Step 2-1) Based on the actual traffic demand of the current signal-controlled intersection, a single control period is used as the time granularity for problem diagnosis, and a lightweight and implementable signal control optimization model is used. With traffic volume data as input, a signal timing pre-optimization plan for a single-point intersection is output;

[0014] Step 2-2) Based on the single-point evaluation index calculation method described in step 1) and the signal control optimization model described in step 2-1), the reference intervals of various evaluable indicators are calculated, wherein the reference interval determination method includes three types: a reference interval derivation method based on a practical limit of saturation, a reference interval determination method based on specifications / papers, and a reference interval derivation method based on traffic wave reconstruction;

[0015] Step 2-3) Based on the evaluation index calculation results and reference intervals, quantitative analysis is performed on the operating status of various evaluable indicators.

[0016] The implementation method of the signal control optimization model includes a WEBSTER method based on minimum delay and a HCM method based on saturation.

[0017] The step 3) specifically comprises the following steps:

[0018] Step 3-1) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed, wherein the traffic problem diagnosis model is a primary-secondary-tertiary architecture, wherein the primary diagnosis problem is divided into two categories: unreasonable supply and unreasonable demand, and such conclusions belong to macroscopic conclusions; the secondary diagnosis problem is subdivided from multiple aspects, directly corresponding to the optimization measures to be taken, wherein the multiple aspects include signal cycle, phase sequence, and channelization design; the tertiary diagnosis problem is used as the smallest unit for characterizing traffic system problems, and is used to establish a corresponding relationship between potential optimization targets and optimization parameters; at the same time, the severity of the diagnosis problem is divided;

[0019] Step 3-2) Based on the quantitative analysis results of the operating status of the evaluable indicators, a corresponding relationship between problem diagnosis and operation evaluation is established;

[0020] Step 3-3) Identify various third-level diagnostic problems, determine potential optimization targets based on the third-level diagnostic problems and their severity levels, and obtain corresponding optimization measures;

[0021] Step 3-4) Based on the identification of various third-level diagnostic problems, the diagnostic results of various third-level problems are comprehensively analyzed, and the first / secondary diagnostic problems are diagnosed to achieve macro control of the signal-controlled intersection.

[0022] The three-level diagnostic questions include:

[0023] During the TOD period, traffic volume fluctuates greatly, traffic intersection conflicts are serious, straight traffic is disturbed by turning traffic, the flow imbalance coefficient is large, the green light time utilization rate is low, the saturation is high, the queue length is long, the lane overflows, the number of parking times is high, the proportion of multiple queues is high, the signal control delay is high, the traffic dissipates slowly at the end of the phase, there are low-saturation intersections, congestion bottlenecks, road section overflows, low-speed convoys, and congestion spreads linearly. The diagnostic conclusions of each third-level diagnostic problem are divided into four levels according to the severity level, namely: no problem, slight problem, significant problem and serious problem.

[0024] The unreasonable supply refers to the traffic problems caused by unreasonable signal control, geometric design, and traffic organization during the control period under study, which are reflected in multiple aspects of the secondary diagnosis problem, including:

[0025] a1) Unreasonable TOD division: refers to traffic problems caused by unreasonable TOD division, which is reflected in the large fluctuation of traffic parameters within the TOD period. The time granularity of the corresponding problem diagnosis is the TOD period level, the spatial granularity is the intersection level, and the diagnostic indicator is the coefficient of variation of the specified traffic parameters within each TOD period;

[0026] a2) Unreasonable green-to-signal ratio: refers to traffic problems caused by unreasonable green-to-signal ratio parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: low green light time utilization, high saturation, low saturation intersections, high signal control delays, more parking times, high proportion of multiple queues, and large flow saturation imbalance coefficient;

[0027] a3) Unreasonable signal cycle: refers to traffic problems caused by unreasonable signal cycle parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: high signal control delay, high number of parking times, high proportion of multiple queues, long queue length, and short lane overflow;

[0028] a4) Unreasonable phase sequence: refers to traffic problems caused by unreasonable phase sequence parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: the overflow cycle ratio is too large and the queue length is long;

[0029] a5) Unreasonable green interval: refers to traffic problems caused by unreasonable green light interval time parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: severe traffic conflict, slow dissipation of traffic at the end of the phase;

[0030] a6) Unreasonable channelization design: refers to traffic problems caused by unreasonable intersection channelization design. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is intersection level. It is reflected in the following three-level diagnosis problems: through traffic is disturbed by turning traffic, green light time utilization is low, saturation is high, and the proportion of multiple queues is high;

[0031] a7) Unreasonable traffic organization: refers to traffic problems caused by unreasonable traffic organization at intersections. The time granularity of problem diagnosis is the TOD time period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnostic problems: high signal control delays, high number of stops, high proportion of multiple queues, large flow imbalance coefficient, long queue length, and serious traffic conflicts.

[0032] The unreasonable demand refers to the traffic problems caused by unreasonable traffic demand management during the control period under study, which is reflected in multiple aspects of the secondary diagnosis problem, including:

[0033] b1) Local overflow: refers to the traffic problem of local overflow caused by unreasonable demand. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is entrance road level / intersection level. It is reflected in the following three-level diagnosis problems: road overflow, congestion bottleneck, frequent parking, and low-speed convoy;

[0034] b2) Global overflow: refers to the traffic problem of global overflow caused by unreasonable demand. It is an extreme case of local overflow problem, that is, the overflow phenomenon further spreads to adjacent intersections. The time granularity of problem diagnosis is TOD time period level, and the spatial granularity is intersection level. It is reflected in the following three-level diagnostic problems: road section overflow, congestion bottleneck, multiple parking, low-speed convoy, and linear spread of congestion.

[0035] The steps 3-4) specifically include the following steps:

[0036] Step 3-4-1) Determine the correspondence between the first / second level diagnostic questions and the third level diagnostic questions;

[0037] Step 3-4-2) Based on the multi-level correspondence of diagnostic problems, on the basis of the scoring results of the third-level diagnostic problems, a comprehensive indicator evaluation method is used to calculate the scoring results of the first / secondary diagnostic problems, wherein the comprehensive indicator evaluation method is divided into two categories: subjective weighting method and objective weighting method. The subjective weighting method includes hierarchical analysis method and fuzzy comprehensive evaluation method, and the objective weighting method includes entropy method and grey correlation method.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] (1) Strong operability: The present invention takes into account the shortcomings of existing signal-controlled intersection management and control systems, such as the lack of state estimation to problem diagnosis and problem-oriented control optimization, and the difficulty in forming an autonomous closed-loop mechanism. It proposes a traffic problem diagnosis method for urban road signal-controlled intersections, enriches the types of problem diagnosis, establishes the connection between state estimation and problem diagnosis, and realizes problem-oriented optimization control. It can accurately and autonomously diagnose the operation problems of signal-controlled intersections, and promotes the development of traffic signal control towards object precision, plan refinement, and process autonomy.

[0040] (2) High scalability: Although the present invention is aimed at signal-controlled intersections on urban roads, the traffic problem diagnosis process constructed is also applicable to different scenarios such as trunk lines, road networks, and expressways, and has high scalability.

[0041] (3) High flexibility: The present invention can use different traffic detection data to replace the built-in algorithm module, which has high flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of the method of the present invention;

[0043] Figure 2 is a schematic diagram of traffic wave reconstruction in an embodiment of the present invention;

[0044] Figure 3 A schematic diagram of lane overflow in an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of road section overflow in an embodiment of the present invention;

[0046] Figure 5 Schematic diagram of a low-speed vehicle fleet in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0048] This embodiment provides a traffic problem diagnosis method for urban road signal-controlled intersections. Figure 1 As shown, the following steps are included:

[0049] Step 1) Starting from the traffic status and control efficiency, an urban road signal control intersection evaluation index system for traffic problem diagnosis is constructed, and cycle-level and lane-level calculations are performed based on the data collected by traffic detection equipment to obtain the calculation results of various evaluation indicators.

[0050] The "People's Republic of China Public Security Industry Standard GA / T 527.2-2016" points out that traffic status refers to general indicators that reflect traffic operation conditions under various signal control modes and specific indicators that reflect traffic operation conditions under specific signal control modes and time and space constraints, such as traffic flow, queue length, etc.; control efficiency refers to indicators that reflect the change range of characteristic indicators before and after signal control adjustment and indicators that reflect signal control effects, such as saturation, average signal control delay, etc. Therefore, an intersection evaluation index system can be constructed as shown in Table 1.

[0051] Table 1

[0052]

[0053]

[0054] Based on the evaluation index system shown in Table 1, some evaluation index results can be directly collected based on traffic detection equipment (for example, traffic volume data can be collected based on checkpoint electronic police equipment, etc.). For indicators that require data-driven calculations, the calculation steps vary depending on the data source, but they are all commonly used by technicians in this field, so they will not be repeated in this embodiment.

[0055] Step 2) In combination with the traffic demand of the analyzed signal-controlled intersection, a single control period is used as the time granularity of problem diagnosis, and the calculation results of the evaluation index are used as input. A pre-optimization plan for the control signal is output based on the signal control optimization model, and the reference intervals of various evaluable indicators are determined; based on the calculation results of the evaluation index and the reference intervals, the operating status of various evaluable indicators is quantitatively analyzed.

[0056] Step 2-1) Based on the actual traffic demand of the current signal-controlled intersection, a single control period is used as the time granularity for problem diagnosis, and a lightweight and implementable signal control optimization model is used. With traffic volume data as input, a signal timing pre-optimization plan for a single-point intersection is output.

[0057] In this process, considering that traffic volume has short-term volatility, even at intersections with low traffic volume levels, there will be oversaturation. It is difficult to provide sufficient effective information for the diagnosis of problems in a single cycle to serve the control optimization. Therefore, the present invention sets the problem diagnosis time granularity to the control period. On this basis, in order to comprehensively judge the operating status of the object under study (single point), the degree of deviation between the current traffic state and the reasonable traffic state can be known, so as to better serve the diagnosis of traffic problems. Based on the above conception, the present invention introduces the concept of pre-optimization, that is, based on the actual traffic demand of the target intersection, by using a lightweight and implementable signal control optimization model, a reasonable signal control scheme under the current traffic demand is obtained, and on this basis, the reference intervals of various evaluation indicators are derived. The pre-optimization step takes traffic volume data as input and outputs a signal timing optimization scheme for a single-point intersection. This process can be achieved through classic signal timing optimization methods, such as the WEBSTER method based on minimum delay, the HCM method based on saturation, etc.

[0058] Step 2-2) is based on the single-point evaluation index calculation method described in step 1) and the signal control optimization model described in step 2-1) to calculate the reference range of various evaluable indicators: that is, the reasonable range of the indicator value when the traffic conditions are good.

[0059] When obtaining the theoretical reference interval of each indicator, the assumption is that the traffic demand remains unchanged, that is, the optimization measures will not affect the number of vehicles arriving. Therefore, indicators such as traffic volume that characterize traffic demand do not have corresponding reference intervals. For various indicators with reference intervals, the determination methods include three types: reference interval derivation method based on saturation practical limit, reference interval determination method based on specifications / papers, and reference interval derivation method based on traffic wave reconstruction. For each type of method, this embodiment selects a representative indicator for illustration, and the reference intervals of other indicators of the same category can be obtained in the same way.

[0060] A. Reference interval derivation method based on practical limit of saturation: The practical limit of saturation can be determined according to each phase (called "phase saturation practical limit"), or it can be determined according to the entire intersection (called "intersection saturation practical limit"). The research results show that some characteristic parameters reflecting the operating conditions of vehicles when passing through intersections: average vehicle delay, average number of stops, and queue length are all related to the actual allowable saturation (i.e. practical limit). Domestic corresponding research shows that when the practical limit of saturation is set between 0.8 and 0.9, the intersection can obtain better operating conditions. Webster and Cauchy used 0.9 as the practical limit of saturation. Based on the above considerations, the practical range of other indicators (such as phase demand imbalance coefficient and phase supply and demand imbalance index) can be calculated. Here, the green light vacant time ratio indicator is used as an example for explanation, and the calculation method is as follows.

[0061]

[0062] Among them, Q E1 is the reference value of green light utilization rate (in %); q G,i is the number of green light time passes in the i-th cycle of the lane in the pre-optimization scheme (in vehicles); t G,i is the pre-optimization of the green light time of the i-th cycle of the down lane (in seconds); h is the saturated headway (in seconds), which needs to be calibrated based on the actual situation;

[0063] On this basis, the reference value of the green light idle time ratio can be further calculated:

[0064] t E =1-Q E1

[0065] Among them, t E Q is the reference value of the idle time ratio of the lane in the cycle (in %); E1 It is the reference value of green light utilization rate (in %).

[0066] Taking into account the short-term volatility of traffic volume, the traffic volume is set to fluctuate within ±20% as the upper limit and lower limit of the reference interval, as shown in Table 2.

[0067] Table 2

[0068]

[0069]

[0070] On this basis, the reference ranges of indicators such as green light utilization rate, phase supply and demand imbalance index, inter-phase demand imbalance index, and intra-phase demand imbalance index can be further obtained through secondary calculation, which will not be repeated here.

[0071] B. Reference interval determination method based on standards / papers: For indicators that need to be calculated based on real traffic detection data, such as headway variation coefficient, short lane queue overflow cycle ratio, and section queue overflow cycle ratio, their reference intervals can be determined based on existing standards or literature. Here, the travel time indicator is used as an example for calculation, and its reference interval can be reversed based on the travel time ratio indicator TTI. The calculation method of the travel time ratio indicator TTI is as follows:

[0072]

[0073] Wherein, TTI represents the travel time ratio parameter (dimensionless); Represents the average driving speed of the studied road section during the current control period (km / h); t f Represents the free-flow speed of the studied road section (km / h).

[0074] On this basis, based on the relationship between travel time ratio and operating status level in the national standard GB / T 33171-2016, the reference interval of travel speed can be reversed to obtain the scores corresponding to the levels of excellent, good, medium and poor. Assuming that L represents the length of the intersection section, the results are shown in Table 3, and the intermediate state score can be obtained by linear interpolation.

[0075] Table 3

[0076]

[0077] C. Reference interval derivation method based on traffic wave reconstruction: Figure 2 As shown in the figure, by assuming the arrival rate of vehicles in the cycle, the running status of vehicles in the road section in the cycle can be restored based on the traffic wave reconstruction method, so as to obtain the reference interval of indicators such as the maximum queue length and the average vehicle delay. Here, the maximum queue length indicator is used as an example to illustrate. Assuming that the traffic demand of the intersection under study remains unchanged, combined with the signal timing parameters obtained by pre-optimization, the cycle traffic volume is first calculated. Assuming that the cycle length of the original timing plan is C o , the cycle length of the pre-optimization solution is C n, the hourly traffic volume in the control period is Q, then the number of vehicles passing through the cycle in the pre-optimization scheme can be obtained:

[0078]

[0079] Among them, Q a is the number of vehicles passing through the pre-optimization period (in vehicles); Q is the hourly traffic volume in the analyzed control period (in vehicles); C m is the cycle length in the pre-optimization scheme (in seconds).

[0080] Based on the number of vehicles arriving in a known cycle, the traffic wave can be reconstructed based on different green light arrival ratios. This embodiment refers to the classification of green light arrival ratios in "HCM 2010" to obtain the arrival ratio during the red light period and the arrival ratio during the green light period in the cycle:

[0081] Excellent (Arrival Type V, Arrival Type VI): In this case, the red light arrival ratio is 20%, and the green light arrival ratio is 80%. In the specific red light period and green light period, it can be assumed that vehicles arrive evenly, and the maximum queue length reference value q can be calculated. max,1 (minimum queue length);

[0082] Good (Arrival Type III, Arrival Type IV): In this case, the proportion of vehicles arriving during the red light is 40%, and the proportion of vehicles arriving during the green light is 60%. Vehicles arrive evenly in different time periods. The calculated maximum queue length reference value is q max,2 ;

[0083] Medium (arrival type II): In this case, the proportion of vehicles arriving during the red light is 60%, and the corresponding proportion of vehicles arriving during the green light is 40%. Vehicles arrive evenly in different time periods, and the calculated maximum queue length reference value is q max,3 ;

[0084] Grade difference (arrival type I): In this case, the red light arrival ratio is 80%, and 80% of the vehicles arrive at the initial red light moment (i.e., the remaining queue length of the previous cycle). The remaining 20% ​​of vehicles are assumed to arrive evenly within the cycle (including the red light period and the green light period). The maximum queue length reference value q can be calculated max,4 ; When the maximum queue length is equal to the periodic traffic volume Q a When , it means that all vehicles in the cycle are queuing.

[0085] Based on the four scenarios shown above, traffic waves can be reconstructed to further obtain the maximum queue length reference value. First, the density of queued vehicles can be calculated as follows:

[0086]

[0087]

[0088]

[0089] Among them, q join Q is the arrival flow rate during the entire cycle (not when the red light is on) (unit: veh / s); a is the traffic volume passing through the whole cycle (in veh); α c is the proportion of arrivals during the entire cycle (here 20%); t c is the cycle length (unit: s); k join is the density of the convoy before joining the queue (in veh / h); v join The speed of the vehicle joining the queue (in m / s), which needs to be calibrated according to historical data; k leave The density of the platoon formed by the release of queued vehicles (unit: veh / h); q leave is the saturation flow rate, which can be obtained from historical observation data; v leave In order to saturate the vehicle's travel speed, it needs to be calibrated based on historical data.

[0090] On this basis, the velocity of the gathering wave can be further calculated as follows:

[0091]

[0092]

[0093] Among them, w f is the velocity of the gathering wave (in m / s); w d is the velocity of the evanescent wave (in m / s); k queue is the density of queued vehicles, which can be obtained from observed data.

[0094] Finally, the maximum queue length can be calculated as follows:

[0095] Q a,rstart =Q a ×α rstart

[0096]

[0097] Among them, Q a,rstart is the number of vehicles arriving at the initial red light (unit: veh); α rstart is the initial arrival ratio of the red light (here 80%); q max is the maximum queue length (in meters); t r is the duration of the red light (unit: s).

[0098] Based on the arrival type classification shown above, the maximum queue length values ​​for excellent, good, medium and poor levels can be obtained respectively. The scores are shown below. The states between different scores can be processed by linear interpolation.

[0099]

[0100] Step 2-3) Based on the evaluation index calculation results and reference intervals, quantitative analysis is performed on the operating status of various evaluable indicators.

[0101] Step 3) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed. Based on the quantitative analysis results of the operating status of evaluable indicators, the correspondence between problem diagnosis and operation evaluation is constructed, potential optimization goals are determined, and problem-oriented control signal optimization is established to achieve traffic problem diagnosis.

[0102] Step 3-1) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed.

[0103] In this embodiment, the traffic problem diagnosis model is a primary-secondary-third-level architecture. The primary diagnostic problems are divided into two categories: unreasonable supply and unreasonable demand. Such conclusions are often relatively macroscopic and can provide theoretical basis and data support for the formulation and regulation of corresponding macroeconomic policies; the secondary diagnostic problems are further subdivided from aspects such as signal control, traffic organization, and channelization design, which can directly correspond to the optimization measures required to be taken; although the secondary diagnostic problems can directly correspond to the optimization measures, it is difficult to directly establish a corresponding relationship with the signal control parameters. Therefore, the present invention sets a three-level diagnostic problem as the smallest unit for characterizing traffic system problems, which is convenient for establishing a corresponding relationship with potential optimization targets and optimization parameters; in addition, the present invention also divides the severity of the diagnostic problems (no problem, slight problem, significant problem, serious problem), accurately determines the optimization object, and better serves traffic managers.

[0104] The three-level diagnostic questions in this embodiment include:

[0105] (1) Large fluctuations in traffic volume during the TOD period: refers to the phenomenon that the analyzed traffic parameters have large fluctuations during the control period under study, showing an obvious segmented distribution. This traffic problem can directly reflect the rationality of the division of intersection control periods. Compared with indicators such as signal control delay and queue length, traffic volume indicators have the advantages of fewer calculation input parameters and high efficiency. The present invention selects traffic volume indicators to diagnose this problem. The time granularity of the problem diagnosis is the TOD period level / day level, and the spatial granularity is the intersection level.

[0106] This problem can be directly analyzed based on real data without the need to provide a corresponding reference interval in the pre-optimization phase. The present invention uses 5 minutes as the minimum analysis time granularity to aggregate the intersection flow, and a traffic volume time series containing 288 elements can be obtained throughout the day. This sequence will be divided into n subsequences of unequal lengths by the TOD time division scheme, and the fluctuation of traffic parameters within each subsequence needs to be analyzed separately, as shown below.

[0107]

[0108] Where: c i,v represents the coefficient of variation of traffic volume in the ith TOD period (in %). i,v When the range is 0-15%, it is a small variation phenomenon. i,v When the range is between 16% and 35%, it is considered as medium variation. i,v When it is greater than 36%, it is a highly variable phenomenon, indicating that the traffic volume fluctuates greatly during the TOD period, that is, the TOD period division is unreasonable; σ(q i ) represents the standard deviation of traffic volume in the ith TOD period (i=1,2,…,n); u(q i ) represents the mean traffic volume in the i-th TOD period (i=1,2,…,n).

[0109] Based on the calculated coefficient of variation of each TOD period, the TOD period division of the intersection throughout the day is further evaluated, and the diagnostic classification is as follows:

[0110] No problem: The proportion of TOD periods with high variability is less than 20%;

[0111] This problem exists (minor): the length of the TOD period with high variability accounts for 20% to 40%;

[0112] · This problem exists (significant): the length of the TOD period with high variability accounts for 40% to 60%;

[0113] This problem exists (serious): the TOD period length with high variability accounts for more than 60%.

[0114] (2) Severe traffic intersection conflicts: refers to a phenomenon in which motor vehicles and motor vehicles, or motor vehicles and non-motor vehicles in a certain flow direction / phase / intersection conflict due to unreasonable setting of phase and sequence parameters. The present invention mainly considers motor vehicle conflicts, and the time granularity of the problem diagnosis is the phase level, and the spatial granularity is the lane group level.

[0115] First, it is necessary to analyze the signal timing scheme in the existing TOD period, analyze each phase, and find the phase containing the cross-conflict point. Assuming that there are n phases containing cross-conflict points in this cycle, it is necessary to analyze the phases containing the conflict points one by one, and use this as the judgment condition:

[0116] A i,Q =q h,a ×q h,b

[0117] Where: q h,a represents the average hourly flow rate of a single lane in the conflicting flow direction a during the control period (in vehicles); q h,b A represents the average hourly flow rate of a single lane in conflicting flow direction b during the control period (in vehicles); i,Q is the conflict flow product of phase i of the conflict phase. Based on the intersection channelization design requirements, when A i,Q If it is greater than 50,000, it means that the traffic flow in this phase is seriously conflicting, which can be expanded to the cycle level as follows:

[0118] No such problem: no conflict point in the cycle;

[0119] This problem exists (minor): There are conflict points, but there is no phase where traffic crosses and conflicts are serious;

[0120] · This problem exists (significant): There is a conflict point, and there is one phase with serious traffic intersection conflict;

[0121] This problem exists (serious): There are conflict points and multiple phases where traffic flows cross and conflict seriously.

[0122] (3) Through traffic is disturbed by turning traffic: when through vehicles on the entrance lane enter the intersection, if the right-turning vehicles on the right entrance lane are not controlled, the through vehicles on the entrance lane will be disturbed by the right-turning vehicles on the right entrance lane, forming a merging conflict point, thus leading to a reduction in traffic capacity. The time granularity of the problem diagnosis is the time period level, and the spatial granularity is the entrance lane level. Taking the conventional four-limb intersection (right turns are all uncontrolled) as an example, there are four spatial objects analyzed: north entrance through and west entrance right turn, east entrance through and north entrance right turn, south entrance through and west entrance right turn, and west entrance through and south entrance right turn.

[0123] First, it is necessary to analyze the signal timing in the existing TOD period, analyze each phase, and find the phase containing the confluence conflict point. Assuming that there are n phases containing the confluence conflict point in this cycle (under the standard four-phase signal timing, n=4), it is necessary to analyze each confluence point one by one, and use this as the judgment condition:

[0124] A i,c =q h,th ×qh,r

[0125] Where: q h,th represents the average hourly flow rate of a single straight lane in the merging conflict direction (in vehicles); q h,r It represents the average hourly flow rate of a single lane of right turn traffic flowing to the adjacent intersection due to merging conflicts (in vehicles); A i,C is the product of the merging conflict flow of phase i of the merging conflict phase (i=1,2,…,n). If it is greater than 50000, it means that the straight traffic flow of this phase is disturbed by the turning traffic flow, and there is room for signal optimization. It is expanded to the periodic level as follows:

[0126] No such problem: There is no confluence conflict point, or there is less than one phase with this problem;

[0127] · This problem exists (minor): There is a confluence conflict point, and there are 1 to 2 phases with this problem;

[0128] · This problem exists (significant): There is a confluence conflict point, and there are 3 to 4 phases with this problem;

[0129] · This problem exists (serious): There is a confluence conflict point, and more than 4 phases have this problem.

[0130] (4) Large flow imbalance coefficient: refers to the imbalance of traffic flow in the same direction or opposite direction, which causes the saturation imbalance coefficient of one or several pairs of flows at the intersection to be greater than the set threshold, that is, the imbalance node. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the intersection level. It can be calculated using the inter-phase demand imbalance coefficient, intra-phase demand imbalance coefficient, and phase supply and demand imbalance index indicators constructed by the evaluation index system.

[0131] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, the score of the diagnostic problem can be further obtained, as shown in the following formula.

[0132]

[0133] Where: M1 is the diagnostic problem score of "large flow imbalance coefficient" (dimensionless); are the evaluation index scores obtained by the inter-phase demand imbalance coefficient, the intra-phase demand imbalance coefficient, and the phase supply and demand imbalance index indicators respectively; δ1, δ2, and δ3 are the weights of the inter-phase demand imbalance coefficient, the intra-phase demand imbalance coefficient, and the phase supply and demand imbalance index indicators.

[0134] On this basis, based on the score result of the "large flow imbalance coefficient" diagnosis problem, the problem can be diagnosed. When the diagnosis score is in the interval [100,80), the corresponding level is "excellent", indicating that the intersection has not diagnosed this problem during the period; if the diagnosis score is in the interval [80,70), the corresponding level is "good", indicating that the intersection has diagnosed this problem during the period, but the degree is mild; if the diagnosis score is in the interval [70,60), the corresponding level is "medium", indicating that the intersection has diagnosed this problem during the period, and the degree is significant; if the diagnosis score is in the interval [60,0], the corresponding level is "poor", indicating that the intersection has diagnosed this problem during the period, and the degree is serious (the corresponding relationship between the following diagnosis problem scores and levels is the same as above).

[0135] (5) Low green light time utilization rate: refers to the low green light time utilization rate of a certain flow direction or intersection during the studied period, that is, the phenomenon of green light vacancy. The most intuitive manifestation is that the green light vacancy time of the intersection is too long during the cycle. The effective diagnosis of this indicator can further serve the subsequent signal timing optimization, especially the redesign of signal timing parameters, such as green-to-signal ratio, cycle duration, etc. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the intersection level. It can be calculated using the saturation, green light vacancy time ratio, phase supply and demand imbalance index, and green light arrival ratio indicators constructed by the evaluation indicator system.

[0136] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, and on the basis of setting the weights of each indicator, we can further obtain the score of the diagnostic problem "low green light time utilization rate". The calculation method and diagnostic problem level classification are the same as the diagnostic problem "large flow imbalance coefficient", so we will not go into details here.

[0137] (6) High saturation: refers to the phenomenon that the saturation of a certain flow direction or intersection is high and greater than a certain threshold during the studied period. It is different from the "congestion bottleneck" in unreasonable demand. The former refers to the phenomenon that the saturation of the main flow direction is high due to unreasonable signal timing design, which leads to a high overall saturation of the intersection, while the latter refers to the phenomenon that all entrances to the intersection are at high saturation due to excessive traffic demand. This phenomenon cannot be improved based on signal timing optimization and needs to be improved from the perspective of traffic demand management. Reasonable diagnosis of "high saturation" can serve the subsequent signal timing optimization, especially the redesign of signal timing parameters. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the intersection level. The saturation, average number of stops, average signal control delay per vehicle, green light empty time ratio, and phase supply and demand imbalance index constructed by the evaluation index system can be used for calculation. By comprehensively analyzing the indicators from different methods, the impact of the calculation error of a single "saturation" indicator can be reduced.

[0138] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, and on the basis of setting the weights of each indicator, we can further obtain the score of the diagnostic problem "high saturation". The calculation method and the classification of diagnostic problems are the same as the diagnostic problem "large flow imbalance coefficient", which will not be repeated here.

[0139] (7) Long queue length: refers to a phenomenon in which the queued vehicles in a certain flow direction or intersection reach or exceed the position of the detector in the road, resulting in the failure of the detection data. In severe cases, it will also cause queue overflow, affecting the driving process of vehicles in other directions. The effective diagnosis of the traffic problem "long queue length" can serve the subsequent signal timing optimization, thereby reducing the probability of road overflow. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the flow level / intersection level. The maximum queue length, remaining queue length, queue length at the time when the green light is on, queue length growth rate, and queue clearing time ratio indicators constructed by the evaluation index system can be used for calculation. By comprehensively analyzing the indicators from different methods, the impact of the calculation error of the single "maximum queue length" indicator can be reduced.

[0140] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, and on the basis of setting the weights of each indicator, we can further obtain the score of the diagnostic problem "long queue length". The calculation method and the diagnostic problem level classification are the same as the diagnostic problem "large flow imbalance coefficient", so we will not go into details here.

[0141] (8) Lane overflow: refers to the phenomenon that the queued vehicles exceed their parking space, that is, the queued vehicles in the short lane or adjacent lane of the intersection exceed their parking space, blocking the vehicles in the adjacent lane from entering the parking space, such as Figure 3 As shown. The effective diagnosis of the traffic problem "lane overflow" can serve the optimization design of phase sequence parameters, thereby reducing the probability of lane overflow. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the entrance level / intersection level. The "short lane queue overflow cycle ratio" indicator constructed by the evaluation index system can be directly used for calculation.

[0142] Based on the calculation results of the evaluation indicator "short lane queue overflow cycle ratio", when the short lane queue overflow cycle ratio is 0% ≤ N s When the short lane queue overflow cycle ratio is 20%≤N s When the short lane queue overflow cycle ratio is 30%≤N sWhen it is less than 40%, it means that this problem can be diagnosed and the degree is significant; when the short lane queue overflow cycle ratio is greater than 40%, it means that this problem can be diagnosed and the degree is serious.

[0143] (9) High number of stops: refers to the phenomenon that the average number of stops in a flow direction or intersection is greater than a certain threshold. It can reflect the overall operation status of the vehicle when passing through the current intersection. Compared with the "high saturation" diagnostic problem, this problem can better reflect the entire operation process of the vehicle; in addition, this problem can reflect the proportion of green light arrivals and quantitatively analyze the coordination effect of the trunk line. The effective diagnosis of the traffic problem "high number of stops" can serve the design of coordination solutions at the trunk line optimization level. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the flow level / intersection level. The "average number of stops" indicator constructed by the evaluation index system can be directly used for calculation.

[0144] Based on the scoring results of the evaluation indicator "average number of parking times", when the score is 80<M≤100, it means that there is no such diagnostic problem; when the score is 70<M≤80, it means that this problem can be diagnosed, but the degree is mild; when the score is 60<M≤70, it means that this problem can be diagnosed and the degree is significant; when the score is lower than 60, it means that this problem can be diagnosed and the degree is serious.

[0145] (10) High proportion of multiple queues: refers to the phenomenon that the proportion of multiple queue vehicles (i.e. vehicles that have stopped more than once) in a lane, flow direction or intersection is greater than a certain threshold. It can reflect the parking status of the intersection together with the traffic problem "high number of stops". The former can reflect the distribution status of specific parking times, while the latter reflects the average parking status of vehicles in the intersection. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the flow level / intersection level. The "proportion of multiple queue vehicles" indicator constructed by the evaluation index system can be directly used to calculate each flow direction, and the intersection-level result can be obtained by flow weighting.

[0146] Based on the scoring results of the evaluation indicator "Proportion of vehicles queued for multiple times", when the score is 80<M≤100, it means that there is no such diagnosis problem; when the score is 70<M≤80, it means that this problem can be diagnosed, but the degree is mild; when the score is 60<M≤70, it means that this problem can be diagnosed and the degree is significant; when the score is lower than 60, it means that this problem can be diagnosed and the degree is serious.

[0147] (11) High signal control delay: refers to a phenomenon in which the signal control delay of a flow direction, phase or intersection is greater than a set threshold. The specific parking time can be quantified based on the "high saturation" and "high number of stops" of traffic problems. The effective diagnosis of this traffic problem can serve the subsequent optimization of signal timing, especially the redesign of signal timing parameters. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the intersection level. The "average vehicle signal control delay" indicator constructed by the evaluation index system can be directly used for calculation.

[0148] Based on the scoring results of the evaluation indicator "average signal control delay per vehicle", when the score is 80<M≤100, it means that there is no such diagnosis problem; when the score is 70<M≤80, it means that this problem can be diagnosed, but the degree is mild; when the score is 60<M≤70, it means that this problem can be diagnosed and the degree is significant; when the score is lower than 60, it means that this problem can be diagnosed and the degree is serious.

[0149] (12) Slow dissipation of traffic flow at the end of a phase: refers to a phenomenon in which the proportion of motor vehicles remaining in a flow direction or intersection is greater than the set threshold, that is, from the end of the current green light to the turn-on of the green light in the next phase, vehicles are unable to leave the intersection due to unreasonable setting of the all-red time parameter. Effective diagnosis of this traffic problem can serve the subsequent optimization of signal timing, especially the selection of yellow light duration and all-red time. The time granularity of the problem diagnosis is the phase level / time period level, and the spatial granularity is the flow direction level / intersection level. The "intersection clearing remaining cycle ratio" indicator constructed by the evaluation index system can be directly used for calculation.

[0150] Based on the calculation results of the evaluation indicator "intersection clearance remaining period ratio", the clearance phenomenon of each phase in the period can be analyzed. For a single period, "sufficient clearance remaining time for this period" means that there is no "insufficient clearance remaining time" phenomenon in any phase, and "insufficient clearance remaining time for this period" means that there is at least one phase with "insufficient clearance remaining time" phenomenon; on this basis, the period-level diagnosis results can be further analyzed.

[0151] No problem: The proportion of cycles with insufficient "clearance remaining time" during the TOD period is less than 20%;

[0152] ·This problem exists (minor): the proportion of insufficient "clearance remaining time" in the TOD period is 20% to 40%;

[0153] · This problem exists (significant): the proportion of insufficient “clearance remaining time” in the TOD period is 40% to 60%;

[0154] This problem exists (serious): the proportion of insufficient "clearance remaining time" cycles during the TOD period is greater than 60%.

[0155] (13) Low saturation intersection: refers to a phenomenon in which the traffic flow at an intersection is far lower than the traffic capacity. It is an extreme case of low green light time utilization, that is, the green light utilization rate of each phase of the intersection is low. This phenomenon occurs when the traffic demand at the intersection is too low, but the cycle is set too long. The "low saturation intersection" problem will reduce the control efficiency of the intersection. When the "low saturation intersection" problem occurs, if the situation is relatively minor, it can be optimized by redesigning the signal timing plan. If the problem is serious, the control method can be changed, such as setting it to induction control. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the flow direction level / intersection level. The saturation, green light utilization rate, intra-phase imbalance coefficient, inter-phase demand imbalance coefficient, and phase supply and demand imbalance index constructed by the evaluation index system can be used for calculation. By comprehensively analyzing the indicators from different methods, the impact of the calculation error of a single "saturation" indicator can be reduced.

[0156] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, and on the basis of setting the weights of each indicator, the score of the diagnostic problem "low-saturation intersection" can be further obtained. The calculation method and the diagnostic problem level classification are the same as the diagnostic problem "large flow imbalance coefficient", which will not be repeated here.

[0157] (14) Congestion bottleneck: refers to the phenomenon that during the analyzed time period, the clearing capacity of each phase of the intersection is insufficient and the remaining queue length of each entrance lane is too high. The effective identification of this problem can serve the subsequent signal timing optimization and traffic demand management at the macro level. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the intersection level. It can be calculated using the remaining queue length, short lane queue overflow cycle ratio, section queue overflow cycle ratio, saturation, average number of stops, multiple stops ratio, and vehicle-average signal control delay indicators constructed by the evaluation index system.

[0158] First, based on the scores corresponding to the various evaluation indicators corresponding to this problem, and on the basis of setting the weights of each indicator, we can further obtain the score of the diagnostic problem "congestion bottleneck". The calculation method and the classification of diagnostic problems are the same as the diagnostic problem "large flow imbalance coefficient", which will not be repeated here.

[0159] (15) Road overflow: refers to the phenomenon that the queue length of the intersection entrance exceeds its safe parking space, affecting the dissipation of upstream vehicles, such as Figure 4 As shown. Effective identification of this problem can serve the subsequent optimization of signal timing and reduce the impact of vehicles queuing at downstream intersections on vehicles traveling at upstream intersections. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the flow level / intersection level. The "road section queue overflow cycle ratio" indicator constructed by the evaluation index system can be directly used for calculation.

[0160] Based on the calculation result of the evaluation indicator "Road Section Queue Overflow Cycle Ratio", when the road section queue overflow cycle ratio is 0, it means that there is no such diagnosis problem; when the road section queue overflow cycle ratio is between 0%≤N s When the ratio of queue overflow cycle is 10%≤N s When the traffic queue overflow ratio is less than 20%, it means that the problem can be diagnosed and the degree is significant; when the section queue overflow period ratio is greater than 20%, it means that the problem can be diagnosed and the degree is serious. Compared with the traffic problem "lane overflow", the diagnosis level of the traffic problem "section overflow" is more stringent. This is because in general, the queue length of the flow direction can easily exceed the short lane storage space, but because the queue length is short, the overflow vehicles dissipate quickly, and the blocked vehicles can enter the corresponding storage space within a certain period of time. In the section overflow, the queued vehicles have filled the entire section, and the dissipation time is longer, and it is easy to cause overall congestion at the upstream intersection. Therefore, its diagnostic conditions are more stringent.

[0161] (16) Low-speed convoy: refers to a phenomenon in which vehicles stop and go, the distance between vehicles is small, the speed is low (generally not more than 30 km / h), and the engine runs at low speed for a long time, that is, the low-speed following phenomenon, which is a kind of induced congestion, such as Figure 5 As shown in the figure. The effective identification of this problem can serve the subsequent signal timing optimization, especially the generation of trunk line coordinated control scheme. The time granularity of the problem diagnosis is the cycle level / time period level, and the spatial granularity is the inlet channel level.

[0162] When the queue length of the downstream entrance road under analysis exceeds the safe queue length of its section, the headway variation coefficient of the three flows b1, b2, and b3 in the upstream intersection can be further analyzed to diagnose the problem. Based on the study of probability theory, when the coefficient of variation is in the range of 0-15%, it is a small variation phenomenon; when the coefficient of variation is in the range of 16%-35%, it is a medium variation phenomenon; when the coefficient of variation is greater than 36%, it is a high variation phenomenon. On this basis, the traffic problem "low-speed convoy" is further diagnosed:

[0163] ·No such problem: the number of upstream flow directions with small variation coefficient of headway sequence is 0;

[0164] ·This problem exists (minor): the number of upstream flow directions with small variation coefficient of headway sequence is 1;

[0165] ·This problem exists (significant): the number of upstream flow directions with small variation coefficient of headway sequence is 2;

[0166] ·This problem exists (serious): The number of upstream flow directions with small variation coefficient of headway sequence is 3.

[0167] (17) Linear congestion spread: refers to the phenomenon that the current intersection forms a congestion bottleneck due to excessive traffic volume, and this trend continues to spread to the upstream intersection of the intersection under study. The effective identification of this problem can serve the subsequent signal timing optimization, especially the generation of trunk or regional coordinated control plans. The time granularity of the problem diagnosis is the time period level, and the spatial granularity is the entrance road level.

[0168] The diagnosis of this problem needs to consider the operating status of the intersection being analyzed and its upstream intersections, and diagnose the "congestion bottleneck" problem of the intersection and its upstream intersection respectively, and further diagnose the traffic problem "congestion spreading linearly" based on the results:

[0169] No problem: the intersection is not a congestion bottleneck, or the number of intersections that are congestion bottlenecks but upstream are congestion bottlenecks is 0;

[0170] This problem exists (minor): the intersection is a congestion bottleneck, and the number of intersections upstream that are congestion bottlenecks is 1;

[0171] · This problem exists (significant): the intersection is a congestion bottleneck, and the number of intersections upstream that are congestion bottlenecks is 2;

[0172] This problem exists (serious): the intersection is a congestion bottleneck, and the number of intersections that are congestion bottlenecks upstream is more than 2.

[0173] In this embodiment, unreasonable supply refers to the traffic problems caused by unreasonable signal control, geometric design, and traffic organization during the control period under study, which can be reflected in multiple aspects such as signal cycle, phase sequence, channelization design, etc. (i.e., secondary diagnosis problems), and is specifically manifested as follows:

[0174] a1) Unreasonable TOD division: refers to traffic problems caused by unreasonable division of TOD scheme (i.e. control period scheme), which is manifested as large fluctuations in traffic parameters within the TOD period. The time granularity of problem diagnosis is TOD period level, the spatial granularity is intersection level, and the diagnostic index is the coefficient of variation of the specified traffic parameter within each TOD period (intersection flow is selected in the present invention).

[0175] a2) Unreasonable green-to-signal ratio: refers to traffic problems caused by unreasonable green-to-signal ratio parameter settings. The time granularity of problem diagnosis is the TOD time period level, and the spatial granularity is the intersection level. It can be reflected in multiple third-level diagnostic problems such as low green light time utilization, high saturation, low-saturation intersections, high signal control delays, more parking times, high proportion of multiple queues, and large flow saturation imbalance coefficient.

[0176] a3) Unreasonable signal cycle: refers to traffic problems caused by unreasonable signal cycle parameter settings. The time granularity of problem diagnosis is the TOD time period level, and the spatial granularity is the intersection level. It can be reflected in multiple third-level diagnostic problems such as high signal control delays, high number of parking times, high proportion of multiple queues, long queue lengths, and mutual interference between straight and left queues (short lane overflow phenomenon).

[0177] a4) Unreasonable phase sequence: refers to traffic problems caused by unreasonable phase sequence parameter settings. The time granularity of problem diagnosis is the TOD time period level, and the spatial granularity is the intersection level. It can be reflected in multiple third-level diagnostic problems such as excessive overflow cycle ratio and long queue length.

[0178] a5) Unreasonable green interval: refers to the traffic problem caused by unreasonable setting of green light interval time parameters (yellow light time, all-red time). The time granularity of problem diagnosis is TOD time period level, and the spatial granularity is intersection level. It can be reflected in multiple third-level diagnostic problems such as serious traffic conflicts, slow dissipation of traffic at the end of the phase (insufficient all-red time setting), etc.

[0179] a6) Unreasonable channelization design: refers to traffic problems caused by unreasonable channelization design of intersections. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It can be reflected in multiple three-level diagnostic problems such as the interference of straight-moving traffic with turning traffic, low green light time utilization, high saturation, and high proportion of multiple queues.

[0180] a7) Unreasonable traffic organization: refers to traffic problems caused by unreasonable traffic organization at intersections. The time granularity of problem diagnosis is the TOD time period level, and the spatial granularity is the intersection level. It can be reflected in multiple third-level diagnostic problems such as high signal control delays, high number of parking times, high proportion of multiple queues, large flow imbalance coefficient (unbalanced nodes), long queue lengths (long queue phenomenon), and serious traffic conflicts.

[0181] Irrational demand refers to the traffic problems caused by unreasonable traffic demand management during the control period under study, which can be reflected in many aspects of secondary diagnosis, such as local overflow at intersections and global overflow at intersections. Specifically, it is manifested as follows:

[0182] b1) Local overflow: refers to traffic problems caused by unreasonable demand, which leads to local overflow. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is entrance road level / intersection level. It can be reflected in multiple third-level diagnostic problems such as road overflow, congestion bottleneck, frequent parking, and low-speed convoys.

[0183] b2) Global overflow: refers to the traffic problem of global overflow caused by unreasonable demand. It is an extreme case of local overflow problem, that is, the overflow phenomenon further spreads to adjacent intersections. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is intersection level. It can be reflected in multiple third-level diagnostic problems such as road overflow, congestion bottleneck, multiple parking, low-speed convoy, linear spread of congestion, etc.

[0184] Step 3-2) Based on the quantitative analysis results of the operating status of the evaluable indicators, a corresponding relationship between problem diagnosis and operation evaluation is established.

[0185] Step 3-3) Identify various types of third-level diagnostic problems, determine potential optimization targets based on the third-level diagnostic problems and their severity levels, and obtain corresponding optimization measures.

[0186] In this embodiment, an optimization measure solution library is constructed for urban road signal-controlled intersections. Specifically, 23 types of optimization measure solutions can be constructed in signal control, channelization design, traffic organization, etc., as shown in Table 4.

[0187] Table 4

[0188] symbol Single point optimization measures symbol Single point optimization measures A Optimize green letter ratio M Set pre-signal B Optimize signal period N One-way street C Optimizing phase structure O Design variable lanes / tidal lanes D Optimize phase structure and lane function P Double crossing / diagonal crossing E Optimization of TOD time division scheme Q Add pedestrian crossing islands / overpasses F Optimize signal control method R Traffic rules / parking line design G Left turn lane on the right S Travel generation optimization H Design left turn / straight ahead waiting area T Travel distribution optimization I Add guide line U Optimizing travel modes J Increase the number of lanes V Spatial route optimization K No left turn (detour left turn) W Time selection optimization L Right turn control

[0189] Based on the conclusions of the diagnosis problem (no problem, slight problem, significant problem, severe problem), the potential optimization objectives and optimization parameters of the traffic system are also clearer. Therefore, the relationship between the diagnosis problem and its severity level and the optimization solution can be further constructed as shown in Table 5.

[0190] Table 5

[0191]

[0192]

[0193] Step 3-4) Based on the identification of various third-level diagnostic problems, the diagnostic results of various third-level problems are comprehensively analyzed, and the first / secondary diagnostic problems are diagnosed to achieve macro control of the signal-controlled intersection.

[0194] The first and second level diagnosis of signal control can provide the most intuitive decision conclusions to traffic managers and decision makers. However, in actual application, it is difficult to directly construct the corresponding relationship between the second level diagnosis questions and the evaluation index parameters. Therefore, the present invention sets a third level diagnosis question, and obtains the conclusion of the first and second level diagnosis question through comprehensive analysis of the overall diagnosis effect of multiple third level diagnosis questions.

[0195] The specific steps include:

[0196] Step 3-4-1) Determine the correspondence between the first / second level diagnostic questions and the third level diagnostic questions, as shown in Table 6.

[0197] Table 6

[0198]

[0199]

[0200] Step 3-4-2) Based on the multi-level correspondence of diagnostic problems, on the basis of the scoring results of the third-level diagnostic problems, a comprehensive indicator evaluation method is used to calculate the scoring results of the first / secondary diagnostic problems.

[0201] The existing comprehensive index evaluation methods are mainly divided into two categories: subjective weighting method and objective weighting method. The former adopts a qualitative method, and experts make subjective judgments based on their experience to obtain the weights of each index, such as the hierarchical analysis method and the fuzzy comprehensive evaluation method, while the latter determines the weights based on the correlation coefficients between the indicators or the coefficients of variation of each indicator, such as the entropy method and the grey correlation method. This step can be calculated based on many existing methods and will not be repeated here.

[0202] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning, or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A traffic problem diagnosis method for urban road signal-controlled intersections, characterized in that: The following steps are involved: Step 1) Starting from the traffic status and control benefits, an evaluation index system for urban road signal-controlled intersections for traffic problem diagnosis is constructed, and the single-point evaluation index calculation method is used to perform cycle-level and lane-level calculations based on the data collected by traffic detection equipment to obtain the calculation results of various evaluation indicators; Step 2) Combined with the traffic demand of the signalized intersection being analyzed, a single control period is used as the time granularity for problem diagnosis, and the evaluation index calculation results are used as input. Based on the signal control optimization model, a pre-optimization scheme for the control signal is output, and the reference intervals of various evaluable indicators are determined; Based on the evaluation index calculation results and reference intervals, a quantitative analysis is performed on the operating status of various evaluable indicators, wherein the reference interval is a reasonable range in which the indicator value should be when the traffic conditions are good, and the evaluable indicator is an evaluation indicator with a reference interval; Step 3) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed. Based on the quantitative analysis results of the operating status of the evaluable indicators, the corresponding relationship between problem diagnosis and operation evaluation is constructed, the potential optimization goals are determined, and problem-oriented control signal optimization is established to realize traffic problem diagnosis; The step 3) specifically includes the following steps: Step 3-1) Starting from the two aspects of unreasonable supply and unreasonable demand, a traffic problem diagnosis model for urban road signal-controlled intersections is constructed, wherein the traffic problem diagnosis model is a primary-secondary-third-level architecture, the primary diagnosis problem is divided into two categories: unreasonable supply and unreasonable demand, and the classification conclusion of the primary diagnosis problem belongs to a macro conclusion; the secondary diagnosis problem is subdivided from multiple aspects, which directly corresponds to the optimization measures to be taken, wherein the multiple aspects include signal cycle, phase sequence, and channelization design; the tertiary diagnosis problem is used as the smallest unit for characterizing traffic system problems, and is used to establish a corresponding relationship between potential optimization targets and optimization parameters; at the same time, the severity of the diagnosis problem is divided; Step 3-2) Based on the quantitative analysis results of the operating status of the evaluable indicators, the corresponding relationship between problem diagnosis and operation evaluation is established; Step 3-3) Identify various third-level diagnostic problems, determine potential optimization targets based on the third-level diagnostic problems and their severity levels, and obtain corresponding optimization measures; Step 3-4) Based on the identification of various third-level diagnostic problems, comprehensively analyze the diagnostic results of various third-level problems, diagnose the first / secondary diagnostic problems, and achieve macro control of signalized intersections; The three-level diagnostic questions include: During the TOD period, the traffic volume fluctuates greatly, the cross-traffic conflicts are serious, the straight traffic is disturbed by the turning traffic, the flow imbalance coefficient is large, the green light time utilization rate is low, the saturation is high, the queue length is long, the lane overflows, the number of parking times is high, the proportion of multiple queues is high, the signal control delay is high, the traffic at the end of the phase dissipates slowly, the low-saturation intersection, the congestion bottleneck, the road section overflow, the low-speed convoy, and the congestion spreads linearly. The diagnostic conclusions of each three-level diagnostic problem are divided into four levels according to the severity level, namely: no problem, slight problem, significant problem and serious problem; The unreasonable supply refers to the traffic problems caused by unreasonable signal control, geometric design, and traffic organization during the control period under study, which are reflected in multiple aspects of the secondary diagnosis problem, including: a1) Unreasonable TOD division: refers to traffic problems caused by unreasonable TOD division, which is reflected in the large fluctuation of traffic parameters within the TOD period. The time granularity of the corresponding problem diagnosis is the TOD period level, the spatial granularity is the intersection level, and the diagnostic indicator is the coefficient of variation of the specified traffic parameters within each TOD period; a2) Unreasonable green-to-signal ratio: refers to traffic problems caused by unreasonable green-to-signal ratio parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: low green light time utilization, high saturation, low saturation intersections, high signal control delays, more parking times, high proportion of multiple queues, and large flow saturation imbalance coefficient; a3) Unreasonable signal cycle: refers to traffic problems caused by unreasonable signal cycle parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: high signal control delay, high number of parking times, high proportion of multiple queues, long queue length, and short lane overflow; a4) Unreasonable phase sequence: refers to traffic problems caused by unreasonable phase sequence parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: the overflow cycle ratio is too large and the queue length is long; a5) Unreasonable green interval: refers to traffic problems caused by unreasonable green light interval time parameter settings. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: severe traffic conflict, slow dissipation of traffic at the end of the phase; a6) Unreasonable channelization design: refers to traffic problems caused by unreasonable intersection channelization design. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is intersection level. It is reflected in the following three-level diagnosis problems: through traffic is disturbed by turning traffic, green light time utilization is low, saturation is high, and the proportion of multiple queues is high; a7) Unreasonable traffic organization: refers to traffic problems caused by unreasonable traffic organization at intersections. The time granularity of problem diagnosis is the TOD period level, and the spatial granularity is the intersection level. It is reflected in the following three-level diagnosis problems: high signal control delay, high number of parking times, high proportion of multiple queues, large flow imbalance coefficient, long queue length, and serious traffic conflict; The unreasonable demand refers to the traffic problems caused by unreasonable traffic demand management during the control period under study, which is reflected in multiple aspects of the secondary diagnosis problem, including: b1) Local overflow: refers to traffic problems caused by unreasonable demand, which leads to local overflow. The time granularity of problem diagnosis is TOD period level, and the spatial granularity is entrance road level / intersection level. It is reflected in the following three-level diagnosis problems: road overflow, congestion bottleneck, frequent parking, and low-speed convoy; b2) Global overflow: refers to the traffic problem of global overflow caused by unreasonable demand. It is an extreme case of local overflow problem, that is, the overflow phenomenon further spreads to adjacent intersections. The time granularity of problem diagnosis is TOD time period level, and the spatial granularity is intersection level. It is reflected in the following three-level diagnostic problems: road section overflow, congestion bottleneck, multiple parking, low-speed convoy, and linear spread of congestion.

2. A traffic problem diagnosis method for urban road signal controlled intersections according to claim 1, characterized in that: The evaluation indicators related to traffic conditions include traffic volume, headway variation coefficient, maximum queue length, queue length when the green light is on, queue length when the red light is on, queue length growth rate, short lane queue overflow cycle ratio, and road section queue overflow cycle ratio.

3. The method for diagnosing traffic problems at urban road signal-controlled intersections according to claim 1, characterized in that: The evaluation indicators related to control effectiveness include saturation, inter-phase demand imbalance index, intra-phase demand imbalance index, green light arrival ratio, average number of stops, proportion of vehicles queuing multiple times, average vehicle control delay, green light empty time ratio, green light utilization rate, queue clearing time ratio, phase supply and demand imbalance index, and intersection clearing remaining cycle ratio.

4. The method for diagnosing traffic problems at urban road signal-controlled intersections according to claim 2, characterized in that: The step 2) specifically includes the following steps: Step 2-1) Based on the actual traffic demand of the current signal-controlled intersection, a single control period is used as the time granularity for problem diagnosis, and a lightweight and implementable signal control optimization model is used. With traffic volume data as input, a signal timing pre-optimization plan for a single-point intersection is output; Step 2-2) Based on the single-point evaluation index calculation method described in step 1) and the signal control optimization model described in step 2-1), the reference intervals of various evaluable indicators are calculated, wherein the reference interval determination methods include three types: a reference interval derivation method based on saturation practical limit, a reference interval determination method based on specifications / papers, and a reference interval derivation method based on traffic wave reconstruction; Step 2-3) Based on the evaluation index calculation results and reference intervals, quantitatively analyze the operating status of various evaluable indicators.

5. A traffic problem diagnosis method for urban road signal-controlled intersections according to claim 1 or 4, characterized in that: The implementation method of the signal control optimization model includes a WEBSTER method based on minimum delay and a HCM method based on saturation.

6. The method for diagnosing traffic problems at urban road signal-controlled intersections according to claim 1, characterized in that: The steps 3-4) specifically include the following steps: Step 3-4-1) Determine the correspondence between the first / second level diagnostic questions and the third level diagnostic questions; Step 3-4-2) Based on the multi-level correspondence of diagnostic problems, on the basis of the scoring results of the third-level diagnostic problems, a comprehensive indicator evaluation method is used to calculate the scoring results of the first / secondary diagnostic problems, wherein the comprehensive indicator evaluation method is divided into two categories: subjective weighting method and objective weighting method. The subjective weighting method includes hierarchical analysis method and fuzzy comprehensive evaluation method, and the objective weighting method includes entropy method and grey correlation method.

Images