Trunk signal cooperative control method for emergencies in networked traffic environment
By dynamically optimizing the phase structure and signal timing scheme in the networked traffic environment, the response lag problem of traditional signal control in emergencies is solved, and rapid response to emergencies and efficient traffic flow is achieved.
Patent Information
- Application Number
- CN202510830828.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-20
- Publication Date
- 2025-08-08
AI Technical Summary
Traditional signal control strategies are difficult to respond quickly and adapt to abnormal traffic flows from multiple sources in the face of emergencies, resulting in accumulation of traffic congestion and reduced system dynamic adaptability, especially in the connected traffic environment, lacking a refined representation mechanism for spatial and temporal correlation characteristics.
The trunk signal collaborative control method in the networked traffic environment is adopted, and the vehicle information and signal light status are collected, the response index is evaluated, and the feasible phase activation scheme in the rolling time domain is dynamically generated. The optimal phase activation scheme is determined based on the objective function evaluation, and the signal timing scheme is optimized to deal with emergencies.
It has achieved rapid identification and response to emergencies, reduced the number of vehicles queuing on the trunk line, shortened the duration of congestion, and improved traffic efficiency by more than 11%.
Smart Images
Figure CN120452228A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of transportation engineering, and in particular relates to a trunk line signal collaborative control method for emergencies in a networked traffic environment. Background Art
[0002] With the acceleration of urbanization, arterial roads, as the main arteries of urban transportation networks, play a crucial role in managing traffic flow. However, during emergencies such as holidays, large-scale events, or temporary traffic control, traffic disturbances caused by the short-term convergence of multiple sources of traffic can easily lead to localized oversaturation of arterial roads, overflow queues, and even cascading congestion across the road network. These scenarios exhibit significant spatiotemporal heterogeneity: dispersed traffic sources, sharp demand peaks, and dynamically expanding impact areas. This poses significant challenges to traditional signal control strategies. Timed signal control based on historical data struggles to adapt to dynamic road conditions, while conventional adaptive signal control exhibits delayed response and slow adjustment to sudden and abnormal disturbances (such as short-term traffic surges and the impact of emergencies), which can easily lead to cumulative congestion. In particular, under the influence of multiple sources of abnormal traffic, traditional control methods, lacking a refined representation of spatiotemporal correlations, significantly reduce the system's dynamic adaptability and real-time response efficiency, making it difficult to promptly curb the spatiotemporal propagation of traffic congestion and achieve rapid recovery of the road network.
[0003] Based on this, a trunk signal coordinated control method for emergencies in a connected traffic environment is proposed. While ensuring the accurate identification and rapid dissipation of local congestion caused by emergencies, it reduces the impact on the overall traffic efficiency of the intersection. Summary of the Invention
[0004] In order to solve the above problems, the present invention proposes a trunk signal collaborative control method for emergencies in a networked traffic environment.
[0005] The present invention relates to a trunk line signal coordinated control method for emergencies in a networked traffic environment, comprising the following steps:
[0006] Step 1: Collect vehicle information and traffic light status at the entrances of each intersection on the main road and evaluate the response index;
[0007] Step 2: Determine whether the traffic flow meets the control model switching conditions;
[0008] Step 3: Dynamically generate a set of feasible phase activation schemes within the rolling time domain T;
[0009] Step 4: Evaluate the objective function of the feasible phase activation schemes and determine the optimal phase activation scheme.
[0010] Furthermore, in step 1, a conventional signal control scheme is set up at each intersection of the trunk line, and the response index of each lane is evaluated with the lanes controlled by each phase at each intersection as the basic unit.
[0011] Collect vehicle operation status information at each intersection entrance. See the schematic diagram of the trunk intersection scene for details. Figure 2 , let (s,l) represent the entrance lane group l of intersection s, let (s,l,b) represent lane b in the entrance lane group l of intersection s, and the lane numbers are sorted from outside to inside as b1, b2, ..., b n Collect the number of vehicles queuing in each lane and the number of vehicles arriving at the stop line after the phase is opened, and record represents the number of vehicles in the queue at time t (s, l, b), Δq s,l,b (t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b).
[0012] (1) When the phase k of the intersection s begins to release traffic, the real-time overflow risk of the phase k controlled lane (s, l, b) is calculated and can be expressed as:
[0013]
[0014] Where, Represents the maximum number of vehicles allowed to queue on (s,l). Indicates the length of (s,l). d v Indicates the vehicle length, d safe Indicates the safe distance between adjacent vehicles in the queue state. r represents the set queue overflow risk threshold. s,l,b (t) represents the total number of dynamic vehicles at time t (s, l, b), which is the sum of the number of queued vehicles and the number of moving vehicles. tanh represents the hyperbolic tangent function, which is expressed as q s,l,b (t) increases; γ, ε, and χ are positive constants, and γ·ε=1.
[0015] (2) When the phase k of the intersection s ends and the lane is released, the real-time saturation of the phase k controlled lane (s, l, b) is calculated and can be expressed as:
[0016]
[0017] Where, t k-1 represents the initial time of phase k, t k Indicates the end time of phase k. Δq s,l,b (t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b). s,l,b (t) indicates the time from t k-1 The duration of the green light at time t (s, l, b). δ s,l,b(t) represents the start-up loss time of the vehicle in the queue at phase k at time t (s, l, b). If phase k is started for the first time at time t, then δ s,l,b (t) = δ0; otherwise, δ s,l,b (t)=0.
[0018] (3) At the end of the nth cycle at intersection s, the response index is evaluated and can be expressed as:
[0019]
[0020] Where, represents the emergency control mode response index, represents the control mode response index of the normal state. α1, α3, α2, and α4 are the thresholds of overflow risk and saturation, respectively.
[0021] Furthermore, in step 2, it is determined whether to switch the control mode according to the current signal control mode:
[0022] (1) If the current signal control mode is the normal state control mode
[0023] Evaluate the response index with the cycle length as the step When the response index meets the following conditions,
[0024]
[0025] Where Z0 represents the response index threshold. N Represents a set of N intersections (s, l, b) controlled by coordinated linkage on the arterial line.
[0026] This indicates that the conventional state control mode cannot meet the normal operation of traffic flow, the phase is continuously oversaturated or (s, l, b) has a continuous overflow risk, then switch to the emergency control mode, execute step 3, reset Otherwise, there is no need to switch the control mode and return to step 1.
[0027] (2) If the current signal control mode is emergency control mode
[0028] Evaluate the response index with the cycle length as the step When the response index meets the following conditions,
[0029]
[0030] This means that the emergency control mode causes the green light time to be wasted and the traffic efficiency to be reduced. Then switch to the normal state control mode, return to step 1, and reset Otherwise, go to step 3.
[0031] Furthermore, in step 3, the phase structure adopts a flexible double-ring structure (see Figure 3 ), when the phase activation scheme at time t (the scheme determined by optimization at time t-1) is completed, the phase schemes allowed to be activated within the rolling time domain T are determined:
[0032] (1) At each time step (τ), only one phase can be activated in each loop β, which can be expressed as:
[0033]
[0034] Where, Indicates whether phase k is activated in the signal cycle n loop β of the intersection s at time t+τ, t+τ∈[t,t+T-1]. K s,β represents the phase set in the ring β of intersection s, |K s,β | represents the phase set K s,β The number of elements in .
[0035] (2) The phase activated in loop β at time t+τ must be the phase activated at time t+τ-1, or the next phase in loop β, which can be expressed as:
[0036]
[0037] Where phase k-1 represents the previous phase of phase k.
[0038] (3) In the double-ring structure, phase activation must follow the “barrier” restriction between the east-west and north-south phases, which can be expressed as:
[0039]
[0040] Where, Indicates whether the activation state of phase k at time t+τ changes compared to time t+τ-1 in the loop β of signal cycle n at intersection s. When the variable takes the value of -1, it means that phase k stops being activated at time t+τ. K ew and K sn These represent the east-west and north-south phase sets, respectively. Due to barrier restrictions, the east-west and north-south phases in both rings must be activated or deactivated simultaneously. Specifically, phases 1 and 5, and 3 and 7 must be activated simultaneously, while phases 2 and 6, and 4 and 8 must be deactivated simultaneously.
[0041] Furthermore, in step 4, based on the feasible phase activation schemes within the rolling horizon T obtained in step 3, a vehicle dynamics simulation is performed using a "store-and-forward" model. Based on the vehicle dynamics simulation results, an objective function evaluation is performed to determine the optimal phase activation scheme. The specific method is as follows:
[0042] (1) Calculate the pressure weight at time t+τ (s, l, b), which can be expressed in basic form as follows:
[0043]
[0044] Where, Represents the lane set in the downstream intersection entrance lane group that leaves (s, l, b). s+1,l+,b+ (t+τ) represents the proportion of vehicles leaving (s, l, b) at time t+τ that enter the downstream intersection lane (s+1, l+, b+). s,l,b (t+τ) represents the larger value between the number of vehicles in the queue at time t+τ (s, l, b) and the maximum number of vehicles that can pass through the intersection stop line based on the distance constraint, which can be expressed as:
[0045]
[0046] Where, v f Indicates the free flow speed of vehicles on (s,l,b). Count(0,v f ·τ) represents the number of vehicles between the stop line and the farthest position a vehicle can reach at free-flow speed within one time step (τ).
[0047] Calculate Q for each lane of the entrance road of intersection s+1 downstream of intersection s s+1,l+,b+ When (t+τ), queue overflow needs to be considered, which can be expressed as:
[0048]
[0049] Where S s,l,b (t+τ) is a binary variable indicating whether a green light was granted to the vehicle at (s,l,b) at time t+τ. If so, the value is 1, which corresponds to the traffic signal update step τ; otherwise, the value is 0. Here, we assume the most extreme case: all vehicles leaving (s,l,b) at time t+τ enter (s+1,l+,b+) and join the queue. If the number of vehicles queuing at (s+1,l+,b+) is less than the overflow risk threshold, the value is 0, indicating that these vehicles will not cause the downstream queue to overflow. Therefore, the pressure on the queue at (s,l,b) is not reduced, maximizing the pressure on the queue at (s,l,b). Otherwise, the value remains unchanged, preventing overflow in the queue at (s+1,l+,b+).
[0050] (2) Based on the above pressure weight calculation, the pressure at time t+τ (s, l, b) can be expressed as:
[0051]
[0052] (3) Note μ s,l,b(t)∈{0,1} represents the variable on (s,l,b) used for coordinated control with the upstream green light on, which can be expressed as:
[0053]
[0054] Where, is the set of lanes in the arterial internal intersection lane group. This variable affects the phase coordination control of the arterial internal intersection lane group at the start of the rolling optimization window (τ = 0) by mapping the green light on state of the upstream intersection (s-1) at the previous time (t + τ - 1). Its core mechanism is: when the green light of (s-1, l-, b-) is on at time t + τ - 1, the green light signal of the corresponding phase of the current intersection (s) is preferentially turned on at time t + τ. For phase opening decisions with τ ≥ 1 within the rolling time domain, each intersection adopts an independent optimization strategy and no longer imposes coordination constraints.
[0055] (4) Based on the above pressure calculation and collaborative control variable definition, the objective function of phase activation in the rolling time domain T can be expressed as:
[0056]
[0057] Where, Represents all lane groups and associated lane sets in intersection s. s,l,b (t+τ) represents the weight for collaborative control on (s, l, b), satisfying: when S s,l,b When (t+τ)=1, η s,l,b The upper limit of (t+τ) is determined by the lane saturation flow rate C s,l,b OK, when S s,l,b When (t+τ)=0, η s,l,b (t+τ)=0.
[0058] (5) Determine the phase activation scheme for multiple time steps (τ) within the rolling horizon T using the above objective function. At time t, implement the decision plan for the current time (τ = 0) in the phase activation scheme. At time t + 1, return to step 1 and re-determine a new phase activation scheme.
[0059] Beneficial effects
[0060] The present invention provides a trunk line signal collaborative control method for emergencies in a networked traffic environment. It can give full play to the communication perception and data sharing capabilities of networked vehicles, quickly identify the disturbances caused by emergencies to the operation status of trunk line traffic, and comprehensively evaluate the queue pressure of each phase at each intersection of the trunk line based on queue overflow risk and lane saturation. Under the premise of ensuring the basic needs of branch road traffic, the phase structure and signal timing plan are dynamically optimized to achieve rapid response and collaborative relief of sudden congestion on the trunk line. By minimizing the number of vehicles queuing on the trunk line, the duration of congestion is effectively shortened and the scale of congestion is reduced, thereby improving the overall traffic efficiency of the trunk line by more than 11%. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Figure 1 It is an implementation flow chart of the present invention.
[0062] Figure 2 Schematic diagram of the trunk collaborative control scenario of the present invention.
[0063] Figure 3 Schematic diagram of the basic phase structure optimized by the present invention.
[0064] Figure 4 Schematic diagram of the intersection entrance road of the present invention. DETAILED DESCRIPTION
[0065] The following describes the embodiments of the present invention in detail.
[0066] The present invention provides a trunk signal coordinated control method for emergencies in a networked traffic environment. The implementation process is summarized as follows:
[0067] First, conventional signal control schemes are set up at each intersection of the trunk line. During traffic operation, the saturation of each phase and the overflow risk of its controlled lane are calculated, and the response index is calculated with the signal cycle as the step size.
[0068] Secondly, if the response index does not reach the threshold, the conventional signal control scheme continues to be executed; if it reaches the threshold, the emergency control mode is switched.
[0069] Then, according to the phase activation sequence of the dual-loop structure and the "barrier" restriction, a set of feasible phase activation schemes in the rolling time domain is dynamically generated.
[0070] Finally, the phase scheme and signal timing are optimized in the rolling time domain T with a time step of τ. That is, at time t, the objective function is evaluated in the time domain t+τ∈[t,t+T-1] to establish the optimal phase activation scheme, execute the decision plan for the current time t in the solution, and re-optimize at the next time (t+1).
[0071] The optimization process of the trunk signal coordinated control method for emergencies in a networked traffic environment is shown in Figure 1 , the optimization steps are as follows:
[0072] Step 1: Collect vehicle information and traffic light status at each intersection entrance, and evaluate the response index
[0073] Conventional signal control schemes are set up at each intersection of the trunk line, and the response index of each lane is evaluated with the lane controlled by each phase at each intersection as the basic unit.
[0074] Collect vehicle operation status information of each intersection entrance lane. Let (s, l) represent the entrance lane group l of intersection s, and let (s, l, b) represent lane b in the entrance lane group l of intersection s. The lane numbers are sorted from outside to inside as b1, b2, ..., b n Collect the number of vehicles queuing in each lane and the number of vehicles arriving at the stop line after the phase is opened, and record represents the number of vehicles in the queue at time t (s, l, b), Δq s,l,b (t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b).
[0075] (1) When the phase k of the intersection s begins to release traffic, the real-time overflow risk of the phase k controlled lane (s, l, b) is calculated and can be expressed as:
[0076]
[0077]
[0078] Where, Represents the maximum number of vehicles allowed to queue on (s,l). Indicates the length of (s,l). d v Indicates the vehicle length, d safe Indicates the safe distance between adjacent vehicles in the queue state. r represents the set queue overflow risk threshold. s,l,b (t) represents the total number of dynamic vehicles at time t (s, l, b), which is the sum of the number of queued vehicles and the number of moving vehicles. tanh represents the hyperbolic tangent function, which is expressed as q s,l,b (t) increases; γ, ε, and χ are positive constants, and γ·ε=1.
[0079] (2) When the phase k of the intersection s ends and the lane is released, the real-time saturation of the phase k controlled lane (s, l, b) is calculated and can be expressed as:
[0080]
[0081] Where, t k-1 represents the initial time of phase k, t k Indicates the end time of phase k. Δq s,l,b(t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b). s,l,b (t) indicates the time from t k-1 The duration of the green light at time t (s, l, b). δ s,l,b (t) represents the start-up loss time of the vehicle in the queue at phase k at time t (s, l, b). If phase k is started for the first time at time t, then δ s,l,b (t) = δ0; otherwise, δ s,l,b (t)=0.
[0082] (3) At the end of the nth cycle at intersection s, the response index is evaluated and can be expressed as:
[0083]
[0084] Where, represents the emergency control mode response index, represents the control mode response index of the normal state. α1, α3 and α2, α4 are the thresholds of real-time overflow risk and real-time saturation respectively.
[0085] like Figure 2 The main road scene intersection 2 is shown as follows: Figure 4 As shown in Table 1, the sudden incident caused a sudden increase in the straight traffic volume from west to east on the main road. The number of queued vehicles at the initial moment of each lane phase, the total number of dynamic vehicles, and the number of vehicles arriving at the stop line during the green light period in three cycles are shown in Table 1. Overflow risk threshold r = 0.8, vehicle length d v =5m, safe distance d between adjacent vehicles in queue safe =1m, constants γ, ε, and χ are 0.02, 50, and 50 respectively, and the green light time g s,l,b =50s, saturation flow rate C s,l,b = 2100 puc / h = 0.58 pcu / s, phase green light loss time δ0 = 2s, the thresholds α1, α3, α2, and α4 of real-time overflow risk and real-time saturation are 1.0, 1.0, and 1.0, 0.6 respectively. The overflow risk R of each lane in the three cycles s,l,b , saturation Φ s,l,b and response index As shown in Table 2.
[0086] Table 1. The number of queued vehicles, the total number of dynamic vehicles, and the number of vehicles arriving at the stop line during the green light period in each lane during the three cycles.
[0087]
[0088] Table 2 Overflow risk, saturation and response index for three periods
[0089]
[0090] Step 2: Determine whether the traffic flow meets the control model switching conditions
[0091] (1) If the current signal control mode is the normal state control mode
[0092] Evaluate the response index with the cycle length as the step When the response index meets the following conditions,
[0093]
[0094] Where Z0 represents the response index threshold. N Represents a set of N intersections (s, l, b) controlled by coordinated linkage on the arterial line.
[0095] This indicates that the conventional state control mode cannot meet the normal operation of traffic flow, the phase is continuously oversaturated or (s, l, b) has a continuous overflow risk, then switch to the emergency control mode, execute step 3, reset Otherwise, there is no need to switch the control mode and return to step 1.
[0096] (2) If the current signal control mode is emergency control mode
[0097] Evaluate the response index with the cycle length as the step When the response index meets the following conditions,
[0098]
[0099] This means that the emergency control mode causes the green light time to be wasted and the traffic efficiency to be reduced. Then switch to the normal state control mode, return to step 1, and reset Otherwise, go to step 3.
[0100] like Figure 2 The three-cycle response index of the west-to-east entrance of intersection 2 in the trunk scene shown As shown in Table 2, the response index threshold is Z0 = 3, then the response index of lanes b2 and b3 after three cycles is Threshold Z0 has been reached and the control mode is switched to emergency control mode.
[0101] Step 3: Dynamically generate a set of feasible phase activation solutions within the rolling time domain T
[0102] The phase structure adopts a flexible double-ring structure (see Figure 3 ), when the phase activation scheme at time t (the scheme determined by optimization at time t-1) is completed, the phase schemes allowed to be activated within the rolling time domain T are determined:
[0103] (1) At each time step (τ), only one phase can be activated in each loop β, which can be expressed as:
[0104]
[0105] Where, Indicates whether phase k is activated in the signal cycle n loop β of the intersection s at time t+τ, t+τ∈[t,t+T-1]. K s,β represents the phase set in the ring β of intersection s, |K s,β | represents the phase set K s,β The number of elements in .
[0106] (2) The phase activated in loop β at time t+τ must be the phase activated at time t+τ-1, or the next phase in loop β, which can be expressed as:
[0107]
[0108] Where phase k-1 represents the previous phase of phase k.
[0109] (3) Phase activation in the double-ring structure must follow the "barrier" restriction between the east-west and north-south phases, which can be expressed as:
[0110]
[0111] Where, Indicates whether the activation state of phase k at time t+τ changes compared to time t+τ-1 in the loop β of signal cycle n at intersection s. When the variable takes the value of -1, it means that phase k stops being activated at time t+τ. K ew and K sn These represent the east-west and north-south phase sets, respectively. Due to barrier restrictions, the east-west and north-south phases in both rings must be activated or deactivated simultaneously. Specifically, phases 1 and 5, and 3 and 7 must be activated simultaneously, while phases 2 and 6, and 4 and 8 must be deactivated simultaneously.
[0112] Table 3 Set of feasible phase activation schemes
[0113]
[0114] like Figure 2 In the arterial scenario shown, the activated phase at intersection 2 at time t-1 is the north-south straight phase (1,5). When the rolling time domain T=2(τ∈[0,T-1]), the set of feasible phase activation schemes is shown in Table 3.
[0115] Step 4: Evaluate the objective function of the feasible phase activation schemes and determine the optimal phase activation scheme
[0116] (1) Based on the feasible phase activation scheme within the rolling time domain T obtained in step 3, the vehicle dynamics simulation is performed using the “store-and-forward” model. According to the vehicle dynamics simulation results, the pressure weight (s, l, b) at time t+τ is calculated. Its basic form can be expressed as:
[0117]
[0118] Where, Represents the lane set in the downstream intersection entrance lane group that leaves (s, l, b). s+1,l+,b+ (t+τ) represents the proportion of vehicles leaving (s, l, b) at time t+τ that enter the downstream intersection lane (s+1, l+, b+). s,l,b (t+τ) represents the larger value between the number of vehicles in the queue at time t+τ (s, l, b) and the maximum number of vehicles that can pass through the intersection stop line based on the distance constraint, which can be expressed as:
[0119]
[0120] Where, v f Indicates the free flow speed of vehicles on (s,l,b). Count(0,v f ·τ) represents the number of vehicles between the stop line and the farthest position a vehicle can reach at free-flow speed within one time step (τ).
[0121] Calculate Q for each lane of the entrance road of intersection s+1 downstream of intersection s s+1,l+,b+ When (t+τ), queue overflow needs to be considered, which can be expressed as:
[0122]
[0123] Where S s,l,b (t+τ) is a binary variable indicating whether a green light was granted to the vehicle at (s,l,b) at time t+τ. If so, the value is 1, which corresponds to the traffic signal update step τ; otherwise, the value is 0. Here, we assume the most extreme case: all vehicles leaving (s,l,b) at time t+τ enter (s+1,l+,b+) and join the queue. If the number of vehicles queuing at (s+1,l+,b+) is less than the overflow risk threshold, the value is 0, indicating that these vehicles will not cause the downstream queue to overflow. Therefore, the pressure on the queue at (s,l,b) is not reduced, maximizing the pressure on the queue at (s,l,b). Otherwise, the value remains unchanged, preventing overflow in the queue at (s+1,l+,b+).
[0124] (2) Based on the above pressure weight calculation, the pressure at time t+τ (s, l, b) can be expressed as:
[0125]
[0126] (3) Note μ s,l,b (t)∈{0,1} represents the variable on (s,l,b) used for coordinated control with the upstream green light on, which can be expressed as:
[0127]
[0128] Where, is the set of lanes in the arterial internal intersection lane group. This variable affects the phase coordination control of the arterial internal intersection lane group at the start of the rolling optimization window (τ = 0) by mapping the green light on state of the upstream intersection (s-1) at the previous time (t + τ - 1). Its core mechanism is: when the green light of (s-1, l-, b-) is on at time t + τ - 1, the green light signal of the corresponding phase of the current intersection (s) is preferentially turned on at time t + τ. For phase opening decisions with τ ≥ 1 within the rolling time domain, each intersection adopts an independent optimization strategy and no longer imposes coordination constraints.
[0129] (4) Based on the above pressure calculation and collaborative control variable definition, the objective function of phase activation in the rolling time domain T can be expressed as:
[0130]
[0131] Where, Represents all lane groups and associated lane sets in intersection s. s,l,b (t+τ) represents the weight for collaborative control on (s, l, b), satisfying: when S s,l,b When (t+τ)=1, η s,l,b The upper limit of (t+τ) is determined by the lane saturation flow rate C s,l,b OK, when S s,l,b When (t+τ)=0, η s,l,b (t+τ)=0.
[0132] (5) Determine the phase activation scheme for multiple time steps (τ) within the rolling horizon T using the above objective function. At time t, implement the decision plan for the current time (τ = 0) in the phase activation scheme. At time t + 1, return to step 1 and re-determine a new phase activation scheme.
[0133] like Figure 2 In the trunk scenario, the phase activated at intersection 2 at time t-1 is the north-south straight phase (1, 5). The phase activation scheme set within the rolling time domain T = 2 at time t is shown in Table 3. At time t, intersection 1 activates the east-west straight phase (3, 7), intersection 3 activates the north-south left turn phase (2, 6), and the vehicle free flow speed v f =55km / h=15.3m / s, traffic signal update step τ=15s, saturation flow rate Cs,l,b =2100 puc / h=0.58 pcu / s, phase green light loss time δ s,l,b (t) = 2s, the yellow light time is 3s, the overflow risk threshold r = 0.8, and the collaborative control weight η s,l,b (t+τ)=5C s,l,b , then the objective function values corresponding to the phase activation scheme set shown in Table 3 in the rolling time domain T=2 are shown in Table 4.
[0134] Table 4 Objective function values
[0135]
[0136] The phase plan numbered A42 has the highest objective function value. Therefore, the optimal phase activation plan for the rolling time domain T = 2 at time t is [(2,6),(3,7)]. This plan activates the north-south left-turn phase first, followed by the east-west straight-through phase in the next time step. At time t, the north-south straight-through phase (1,5) is deactivated. During the time period (t, t+1), the north-south straight-through phase is activated with a yellow light for 3 seconds, followed by a green light for 12 seconds for the north-south left-turn phase. At time t+1, when the current optimal plan completes, the process returns to step 1 and re-establishes a new phase activation plan.
[0137] The above contents of the present invention are only preferred embodiments of the present invention and are not intended to limit the implementation scheme of the present invention. Ordinary technicians in this field can easily make corresponding changes or modifications based on the main concepts and spirit of the present invention. Therefore, the scope of protection of the present invention shall be based on the scope of protection required by the claims.
Claims
1. The method for coordinated control of trunk line signals for emergencies in a connected traffic environment is characterized by: The steps include: Step 1: Collect vehicle information and traffic light status at the entrances of each intersection on the main road and evaluate the response index; Step 2: Determine whether the traffic flow meets the control model switching conditions; Step 3: Dynamically generate a set of feasible phase activation schemes within the rolling time domain T; Step 4: Evaluate the objective function of the feasible phase activation schemes and determine the optimal phase activation scheme.
2. The trunk signal coordinated control method for emergencies in a networked traffic environment according to claim 1 is characterized in that: In step 1, conventional signal control schemes are set up at each intersection of the trunk line, and the response index of each lane is evaluated with the lanes controlled by each phase at each intersection as the basic unit. Collect vehicle operation status information at the entrance lanes of each intersection. The schematic diagram of the trunk intersection scene is shown in Figure 2. (s, l) represents the entrance lane group l of intersection s, and (s, l, b) represents lane b in the entrance lane group l of intersection s. The lane numbers are sorted from the outside to the inside as b1, b2, ..., b n Collect the number of vehicles queuing in each lane and the number of vehicles arriving at the stop line after the phase is opened, and record represents the number of vehicles in the queue at time t (s, l, b), Δq s,l,b (t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b). (1) When the phase k of the intersection s begins to release traffic, the real-time overflow risk of the phase k controlled lane (s, l, b) is calculated and can be expressed as: Where, Represents the maximum number of vehicles allowed to queue on (s,l). Indicates the length of (s,l). d v Indicates the vehicle length, d safe Indicates the safe distance between adjacent vehicles in the queue state. r represents the set queue overflow risk threshold. s,l,b (t) represents the total number of dynamic vehicles at time t (s, l, b), which is the sum of the number of queued vehicles and the number of moving vehicles. tanh represents the hyperbolic tangent function, which is expressed as q s,l,b (t) increases; γ, ε, and χ are positive constants, and γ·ε=1. (2) When the phase k of the intersection s ends and the lane is released, the real-time saturation of the phase k controlled lane (s, l, b) is calculated and can be expressed as: Where, t k-1 represents the initial time of phase k, t k Indicates the end time of phase k. Δq s,l,b (t) represents the number of vehicles that have newly arrived at the stop line at time t (s, l, b). s,l,b (t) indicates the time from t k-1 The duration of the green light at time t (s, l, b). δ s,l,b (t) represents the start-up loss time of the vehicle in the queue at phase k at time t (s, l, b). If phase k is started for the first time at time t, then δ s,l,b (t) = δ0; otherwise, δ s,l,b (t)=0. (3) At the end of the nth cycle at intersection s, the response index is evaluated and can be expressed as: Where, represents the emergency control mode response index, represents the control mode response index of the normal state. α1, α3, α2, and α4 are the thresholds of overflow risk and saturation, respectively.
3. The trunk signal coordinated control method for emergencies in a networked traffic environment according to claim 1 is characterized in that: In step 2, determine whether to switch the control mode based on the current signal control mode: (1) If the current signal control mode is the normal state control mode Evaluate the response index with the cycle length as the step When the response index meets the following conditions, Where Z0 represents the response index threshold. N Represents a set of N intersections (s, l, b) controlled by coordinated linkage on the arterial line. This indicates that the conventional state control mode cannot meet the normal operation of traffic flow, the phase is continuously oversaturated or (s, l, b) has a continuous overflow risk, then switch to the emergency control mode, execute step 3, reset Otherwise, there is no need to switch the control mode and return to step 1. (2) If the current signal control mode is emergency control mode Evaluate the response index with the cycle length as the step When the response index meets the following conditions, This means that the emergency control mode causes the green light time to be wasted and the traffic efficiency to be reduced. Then switch to the normal state control mode, return to step 1, and reset Otherwise, go to step 3.
4. The trunk signal coordinated control method for emergencies in a networked traffic environment according to claim 1 is characterized in that: In step 3, the phase structure adopts a flexible dual-loop structure (see Figure 3). When the phase activation plan at time t (the plan optimized and determined at time t-1) is completed, the phase plans allowed to be activated within the rolling time domain T are determined: (1) At each time step (τ), only one phase can be activated in each loop β, which can be expressed as: Where, Indicates whether phase k is activated in the signal cycle n loop β of the intersection s at time t+τ, t+τ∈[t,t+T-1]. K s,β represents the phase set in the ring β of intersection s, |K s,β | represents the phase set K s,β The number of elements in . (2) The phase activated in loop β at time t+τ must be the phase activated at time t+τ-1, or the next phase in loop β, which can be expressed as: Where phase k-1 represents the previous phase of phase k. (3) In the double-ring structure, phase activation must follow the "barrier" restriction between the east-west and north-south phases, which can be expressed as: Where, Indicates whether the activation state of phase k at time t+τ changes compared to time t+τ-1 in the loop β of signal cycle n at intersection s. When the variable takes the value of -1, it means that phase k stops being activated at time t+τ. K ew and K sn These represent the east-west and north-south phase sets, respectively. Due to "barrier" restrictions, the east-west and north-south phases in both rings must be activated or deactivated simultaneously. Specifically, phases 1 and 5, and 3 and 7 must be activated simultaneously, while phases 2 and 6, and 4 and 8 must be deactivated simultaneously.
5. The trunk line signal coordinated control method for emergencies in a networked traffic environment according to claim 1, characterized in that: In step 4, based on the feasible phase activation schemes within the rolling horizon T obtained in step 3, a "store-and-forward" model is used to simulate vehicle dynamics. Based on the results of the vehicle dynamics simulation, an objective function is evaluated and the optimal phase activation scheme is determined. The specific method is as follows: (1) Calculate the pressure weight at time t+τ (s, l, b), which can be expressed in basic form as follows: Where, Represents the lane set in the downstream intersection entrance lane group that leaves (s, l, b). s+1,l+,b+ (t+τ) represents the proportion of vehicles leaving (s, l, b) at time t+τ that enter the downstream intersection lane (s+1, l+, b+). s,l,b (t+τ) represents the larger value between the number of vehicles in the queue at time t+τ (s, l, b) and the maximum number of vehicles that can pass through the intersection stop line based on the distance constraint, which can be expressed as: Where, v f Indicates the free flow speed of vehicles on (s,l,b). Count(0,v f ·τ) represents the number of vehicles between the stop line and the farthest position a vehicle can reach at free-flow speed within one time step (τ). Calculate Q for each lane of the entrance road of intersection s+1 downstream of intersection s s+1,l+,b+ When (t+τ), queue overflow needs to be considered, which can be expressed as: Where S s,l,b (t+τ) is a binary variable that indicates whether a green light was granted to the vehicle at (s,l,b) at time t+τ. If so, the value is 1, which corresponds to the traffic signal update step τ; otherwise, the value is 0. Here, we assume the most extreme case: all vehicles leaving (s,l,b) at time t+τ enter (s+1,l+,b+) and join the queue. If the number of vehicles queuing at (s+1,l+,b+) is less than the overflow risk threshold, the value is 0, indicating that these vehicles will not cause overflow in the downstream queue. Therefore, the queue pressure at (s,l,b) is not reduced, and the queue pressure at (s,l,b) is maximized. Otherwise, the value remains unchanged to avoid (s+1,l+,b+) queue overflow. (2) Based on the above pressure weight calculation, the pressure at time t+τ (s, l, b) can be expressed as: (3) Note μ s,l,b (t)∈{0,1} represents the variable on (s,l,b) used for coordinated control with the upstream green light on, which can be expressed as: Where, is the set of lanes in the arterial internal intersection lane group. This variable affects the phase coordination control of the arterial internal intersection lane group at the start of the rolling optimization window (τ = 0) by mapping the green light on state of the upstream intersection (s-1) at the previous time (t + τ - 1). Its core mechanism is: when the green light of (s-1, l-, b-) is on at time t + τ - 1, the green light signal of the corresponding phase of the current intersection (s) is preferentially turned on at time t + τ. For phase opening decisions with τ ≥ 1 within the rolling time domain, each intersection adopts an independent optimization strategy and no longer imposes coordination constraints. (4) Based on the above pressure calculation and collaborative control variable definition, the objective function of phase activation in the rolling time domain T can be expressed as: Where, Represents all lane groups and associated lane sets in intersection s. s,l,b (t+τ) represents the weight for collaborative control on (s, l, b), satisfying: when S s,l,b When (t+τ)=1, η s,l,b The upper limit of (t+τ) is determined by the lane saturation flow rate C s,l,b OK, when S s,l,b When (t+τ)=0, η s,l,b (t+τ)=0. (5) Determine the phase activation scheme for multiple time steps (τ) within the rolling horizon T using the above objective function. At time t, implement the decision plan for the current time (τ = 0) in the phase activation scheme. At time t + 1, return to step 1 and re-determine a new phase activation scheme.