Signal sensing control method, electronic device and system

Abstract

The application discloses a signal induction control method, comprising the following steps: based on the number of stored vehicles, the number of incoming vehicles and the number of outgoing vehicles of a current intersection, predicting the optimal decision point of each phase of the current intersection through a global induction model; adjusting the optimal decision point based on the flow direction comparison result between the current phase and the overlapping phase of the current intersection; and performing phase switching according to the adjusted optimal decision point, so as to solve the problem of long-time emptying of a signal lamp.

Description

Technical Field

[0001] This application relates to the field of signal control technology, and in particular to a signal sensing control method, electronic device and system. Background Technology

[0002] Traffic light intersections typically use radar-based detection devices to monitor vehicles and road conditions. The method of using these devices to sense vehicles and control the switching of traffic light phases is called traffic sensing control. Because this method extends the green light time (e.g., by 12 seconds) as soon as it detects a vehicle crossing the stop line, if vehicle A, which has just crossed the stop line, has no following vehicles, the corresponding traffic flow may experience a prolonged period of idling after vehicle A leaves the intersection (i.e., 11 seconds of green light with no vehicles passing through). Simultaneously, other traffic flows may become congested due to this extended waiting time. Summary of the Invention

[0003] The purpose of this application is to provide a traffic sensing control method based on radar vision equipment, thereby solving the problem of long idle time for some traffic lights while congestion occurs in other traffic lights.

[0004] In a first aspect, embodiments of this application provide a signal sensing control method, including: Based on the current number of vehicles parked, approaching, and departing at the intersection, the optimal decision point for each phase of the current intersection is predicted using a global sensing model. The optimal decision point is adjusted based on the flow direction comparison between the current phase and the overlapping phase at the current intersection. Phase switching is performed based on the adjusted optimal decision point.

[0005] Optionally, predicting the optimal decision point for each phase of the current intersection using the global sensing model includes: For the current intersection, a single-loop mode is used to preset multiple phases, and a model predictive control (MPC) framework is used to predict the number of vehicles stored at multiple decision points for each phase of the current intersection. The decision point corresponding to the minimum predicted number of vehicles is taken as the optimal decision point.

[0006] Optionally, based on the flow direction comparison result between the current phase and the overlapping phase of the current intersection, the optimal decision point is adjusted, including: Compare the current phase of the current intersection with the flow direction of its overlapping phase to obtain the flow direction comparison result; When the flow direction comparison result indicates that there is an independent flow direction between the current phase and its overlapping phase, the duplicate flow direction in the current phase is deleted, and the predicted number of vehicles in the current phase at multiple decision points is predicted based on the independent flow direction and the duplicate flow direction in the overlapping phase. When the flow direction comparison result indicates that there is no independent flow direction between the current phase and its overlapping phase, the predicted number of vehicles in the current phase at multiple decision points is predicted based on the difference in the number of vehicles in the repeated flow directions between the current phase and its overlapping phase. The decision point corresponding to the minimum predicted number of vehicles is taken as the optimal decision point.

[0007] Optionally, predicting the optimal decision point for each phase of the current intersection using the global sensing model includes: For any phase of the current intersection: After the corresponding key flow direction switches phases at each decision point, the predicted number of vehicles in the key flow direction is calculated. The key flow direction is the flow direction corresponding to the green light of the current phase and / or the overlapping phase. By comparing the predicted number of vehicles at each decision point for the key flow direction, the decision point with the fewest predicted vehicles is taken as the optimal decision point for that phase.

[0008] Optionally, predicting the predicted number of vehicles in each phase of the current intersection at multiple decision points includes: For each phase of the current intersection, calculate the total number of vehicles in all directions after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase of the current intersection, calculate the total number of vehicles remaining in the original flow direction of the phase after switching phases at each decision point, and use this as the predicted number of vehicles remaining in the phase at each decision point; For each phase at the current intersection, calculate the total remaining number of vehicles in the corresponding flow direction of that phase and the overlapping phase after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles remaining in the lane corresponding to that phase after switching phases at each decision point, and use this as the predicted number of vehicles remaining in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles remaining in the lane with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles remaining in that phase at each decision point; or For each phase of the current intersection, calculate the number of vehicles in the lane with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point.

[0009] Optionally, predicting the optimal decision point for each phase of the current intersection using a global sensing model includes: The current number of vehicles stored at the current intersection, collected by the radar-sensing device at the current moment, is input into the global sensing model; The global sensing model predicts the first arrival rate and the first departure rate of the current phase based on the current number of vehicles at the current intersection at the current time and the historical data of the current intersection. Based on the flashing duration, the first arrival rate and the first departure rate, it predicts the predicted number of vehicles after the current phase switches phases at multiple decision points. The decision point with the fewest predicted vehicles is taken as the optimal decision point. Specifically, when the radar device detects a large number of discrete green-end traffic flows at the current intersection, the first departure rate is greater than 0.5 pcu / s. When the radar device detects the presence of vehicles of a special type at the current intersection, the first departure rate is less than 0.3 pcu / s.

[0010] Optionally, predicting the predicted number of vehicles remaining after the current phase switches phases at multiple decision points based on the flash duration, the first arrival rate, and the first departure rate includes: For each decision point of the current phase, Calculate the current phase based on the first arrival rate and the first departure rate, and the first number of arriving vehicles and the first number of departing vehicles from the current moment to the second before the decision point; Based on the saturation flow rate of the current phase detected by the radar vision device, calculate the second number of vehicles passing through the current phase within the flashing duration, wherein the flashing duration includes the countdown and / or the yellow light time; Based on the first number of incoming vehicles, the first number of outgoing vehicles, and the second number of outgoing vehicles, the predicted number of vehicles to be stored at the decision point for the current phase is determined.

[0011] Optionally, before performing a phase switch based on the adjusted optimal decision point, the method further includes: One second before the execution time of the current phase reaches the optimal decision point, the number of vehicles at the current intersection is obtained through the radar camera, input into the global sensing model, and the optimal decision point of each phase of the current intersection is predicted again. If the optimal decision point is predicted again to be at second 0, a phase switch is executed; otherwise, the current phase is extended by one unit of time.

[0012] Optionally, the optimal decision points for each phase of the current intersection are predicted using a global sensing model, including: Multiple decision points are set for the current phase of the current intersection, and the time window between each decision point is fixed; The maximum number of vehicles that can be released at the multiple decision points is calculated based on the vehicle departure rate. The number of vehicles that can be released within the flashing duration is calculated based on the saturated headway. If the number of vehicles that can be released within the flashing duration is greater than the maximum number of vehicles that can be released at the multiple decision points, the number of decision points is increased until the number of vehicles that can be released within the flashing duration is not greater than the predicted maximum number of vehicles that can be released at the multiple decision points. When the number of vehicles that can be released during the flashing duration is less than the maximum number of vehicles that can be released at the plurality of decision points, a phase switch is performed at second 0.

[0013] Secondly, embodiments of this application provide an electronic device, including: at least one memory and at least one processor, wherein the at least one memory stores executable code, and the at least one processor is used to execute the executable code in the at least one memory to implement the above-described signal sensing control method.

[0014] Thirdly, embodiments of this application provide a traffic control system, including: traffic lights and the aforementioned electronic equipment.

[0015] This application provides a signal sensing control method that can reduce the number of vehicles at the entire intersection through a global sensing control model, thereby optimizing the overall traffic efficiency of the current intersection. In addition, by combining the flow direction comparison results between the current phase and the overlapping phase, the output of the global sensing model is adjusted, taking into account the vehicle traffic situation after the current phase switches to the overlapping phase. This can improve and reduce the idle time that occurs after the phase switch, thereby avoiding the problem of long idle time in some flows while congestion occurs in other flows. Attached Figure Description

[0016] Figure 1 A flowchart of a signal sensing control method provided in an embodiment of this application; Figure 2 A schematic diagram illustrating the normal queue length and detection range of the radar-guided equipment as provided in the embodiments of this application; Figure 3 A flowchart of a signal sensing control method provided in an embodiment of this application; Figure 4 This is a schematic diagram of the time window sliding in an embodiment of this application; Figure 5 This is a schematic diagram of the traffic light colors according to an embodiment of this application; Figure 6 This is one of the methods for calculating the second number of vehicles to depart during the countdown phase in this application embodiment; Figure 7 This is a schematic diagram of the flow direction of two phases in Example 1 of this application embodiment; Figure 8 This is a schematic diagram of the flow direction of the two phases in Example 2 of the embodiments of this application; Figure 9 This is a schematic diagram of the flow direction of the two phases in Example 3 of this application. Detailed Implementation

[0017] The present application will be described in detail below with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present application. Any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the protection scope of the present application.

[0018] Currently, most traffic sensing control methods used in my country employ a single-loop mode based on a phase-phase structure. If the control frequency of a traffic sensing control method is fixed, for example, making a decision every 3 seconds, its control principle can be: for a certain phase, every 3 seconds, check if a vehicle has crossed the stop line; if so, extend the green light for that phase by 3 seconds; otherwise, switch to the next phase. The inventors discovered that because this fixed-interval traffic sensing control method does not consider future information, such as the status of the adjacent phases after the current phase, or vehicles accelerating during the countdown period, it may result in prolonged periods of inactivity.

[0019] Please refer to Figure 1 This application provides a traffic sensing control method, including: S101: Based on the current number of vehicles at the intersection, the number of vehicles approaching, and the number of vehicles leaving, the optimal decision point for each phase of the intersection is predicted using a global sensing model. S102: Adjust the optimal decision point based on the flow direction comparison between the current phase and the overlapping phase at the current intersection; S103: Perform a phase switch based on the adjusted optimal decision point.

[0020] In an optional embodiment of this application, multiple phases are preset for the current intersection using a single-loop mode. An MPC (Model Predictive Control) framework is employed to predict the predicted number of vehicles remaining at each phase of the current intersection at multiple decision points. The decision point corresponding to the lowest predicted number of vehicles remaining is selected as the optimal decision point. That is, the aforementioned global sensing model can be implemented using an MPC framework.

[0021] The MPC framework can be comprised of four parts: a dynamic prediction model, a feedback correction module, a rolling optimization module, and a reference trajectory module. The dynamic prediction model can include state transition equations. The feedback correction module calculates the difference between the actual number of vehicles remaining and the predicted number after each phase switch based on the number of incoming and outgoing vehicles detected by the radar equipment. This difference is used to correct the parameters of the prediction model, improving its accuracy. The rolling optimization module employs a rolling finite-time optimization strategy, predicting the number of vehicles remaining for N decision points in a preset window, resulting in higher accuracy compared to solutions that only consider the next step. The reference trajectory module adjusts the data output by the prediction model based on expectations and smooth curves.

[0022] If MPC decision-making only considers the information at the current moment, it is easy to have long periods of empty parking without any vehicles passing through. The embodiments of this application comprehensively consider the predicted number of vehicles in each phase at multiple decision points and compare them to find the optimal decision point. This can reduce the green light parking time and improve traffic efficiency while adapting to scenarios with frequent traffic flow fluctuations.

[0023] The radar-based vision system includes detection methods such as radar, video, and images. It can detect the number of vehicles, vehicle trajectories, and vehicle speeds at various road sections of the current intersection.

[0024] For example, the radar-sensing device can be installed on a traffic light pole, and its detection area can include intersections and the normal queuing area behind the stop line. (See reference...) Figure 2 In this embodiment, the number of vehicles currently parked on the road segment can be detected by the radar-guided vehicle (RBV) device, and the detection range of the RBV device is greater than the normal queue length. The normal queue length can be considered as the normal queue area from the stop line to the vehicles. For example, the normal queue area may include, but is not limited to, the portion from the stop line to the solid lane line. For example, a four-way intersection can be equipped with four RBV devices, one for each exit direction.

[0025] In this embodiment of the application, the global sensing model can predict the number of vehicles in the current road segment in the next few seconds, and determine the optimal decision point for switching phases based on the predicted number of vehicles in the next few seconds.

[0026] For example, when the number of vehicles in the current intersection and current road segment is 4, the original preset unit phase duration of the traffic light is T=60 seconds, and the preset unit extension time of the induction extension is t=3 seconds. There are 5 decision points for the current phase, namely the current time (second 0), and the 3rd, 6th, 9th, and 12th seconds after the current time. The total number of vehicles in all phases after the traffic light switches phases at the 5 decision points is calculated by the global induction model. The decision point with the smallest number of vehicles is taken as the optimal decision point, and the second before the decision point is taken as the starting time of the next decision.

[0027] Assuming the calculated optimal decision point is at second 6, the decision results of the global sensing model at seconds 0, 3, 6, 9, and 12 are: Hold, Hold, Switch, Hold, Hold. After this decision result is issued, the green light for that phase will be extended by one unit, and the time window will slide once. At second 3, the global sensing model restarts the next decision process. If the next decision result is: Hold, Switch, Hold, Hold, Hold, Hold, the green light for that phase will be extended for another 3 seconds, and the time window will slide once more. This continues until the result is: Switch, Hold, Hold, Hold, Hold, Hold, where the decision result at second 0 is Switch, at which point the current phase is directly switched to the next phase.

[0028] The configurable parameters of the global sensing model include: 1. Unit extension time t (i.e., the duration of a single slide of the time window, also known as the green light extension time); 2. Number of prediction units K, which is the number of decision points predicted by the global sensing model in one decision, where t can be in seconds, microseconds, etc., and K can be a non-zero integer. Since the interval between two adjacent decision points can be the sliding duration of a single time window, i.e., the unit extension time t, the total time length predicted by the global sensing model can be calculated by multiplying the number of prediction units K by the unit extension time t.

[0029] For example, the optimal decision point can be adjusted based on the flow direction comparison between the current phase and the overlapping phase at the current intersection, but is not limited to the following methods: Compare the current phase of the intersection with the flow direction of its adjacent phase to obtain the flow direction comparison result; When the flow direction comparison result indicates that there is an independent flow direction between the current phase and its overlapping phase, the duplicate flow direction in the current phase is deleted, and the predicted number of vehicles in the current phase at multiple decision points is predicted based on the independent flow direction and the duplicate flow direction in the overlapping phase. When the flow direction comparison results indicate that there is no independent flow direction between the current phase and its overlapping phase, the predicted number of vehicles in the current phase at multiple decision points is predicted based on the difference in the number of vehicles in the repeated flow directions between the current phase and its overlapping phase. The decision point corresponding to the minimum predicted number of vehicles is taken as the optimal decision point.

[0030] When calculating the predicted number of vehicles remaining for each phase, all lanes in that phase can be calculated, and the sum of the number of vehicles remaining in all lanes can be used as the predicted number of vehicles remaining for that phase. Alternatively, the number of vehicles remaining in the critical lanes of that phase can be used as the predicted number of vehicles remaining for that phase, or the number of vehicles remaining in the flow direction corresponding to that phase can be used as the predicted number of vehicles remaining for that phase. The critical lane can be the lane with the longest queue length and / or the lane with the highest number of vehicles remaining.

[0031] Optionally, for any phase of the current intersection: After the corresponding key flow direction switches phases at each decision point, the predicted number of vehicles in the key flow direction is calculated. The key flow direction is the flow direction corresponding to the green light of the current phase and / or the overlapping phase. The predicted number of vehicles in the key flow direction at each decision point is compared, and the decision point with the fewest predicted vehicles in the key flow direction is taken as the optimal decision point for that phase.

[0032] In this way, calculating the predicted number of vehicles in key flow directions (i.e., the flow directions corresponding to the green light of the current phase and / or the overlapping phase) can more easily and quickly determine the optimal decision point for switching phases, thereby ensuring the traffic efficiency of the intersection.

[0033] For example, multiple phases can be preset for the current intersection using a single-loop mode. The predicted number of vehicles remaining after switching phases at each decision point can be determined, but is not limited to, in the following ways: Calculate the total number of vehicles in all directions at the current intersection after the phase switch at each decision point, and use this as the predicted number of vehicles in each direction at each decision point for that phase; or After the current intersection switches phases at each decision point, calculate the total number of vehicles remaining in the direction originally corresponding to that phase, and use this as the predicted number of vehicles in that phase at each decision point; Calculate the total remaining number of vehicles in the current intersection and the corresponding flow direction after the phase switch at each decision point, and use this as the predicted number of vehicles in the current phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase of the current intersection, calculate the total number of vehicles in the lane corresponding to that phase after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles remaining in the lane with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles remaining in that phase at each decision point; or For each phase of the current intersection, calculate the number of vehicles in the lane with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point.

[0034] For example, the global sensing model can calculate the predicted number of vehicles in each phase of the current intersection using the following formula:

[0035]

[0036] In the above formula, This represents a list of all phases at the current intersection calculated by the global sensing model, where i represents the phase number and k represents the decision point number. This represents the predicted number of vehicles at the intersection at the k-th decision point in the i-th phase. This represents the number of vehicles arriving at the k-th decision point in the i-th phase. This represents the number of vehicles leaving at the k-th decision point in the i-th phase (when the phase is green, the number of vehicles leaving is calculated according to the above formula; when the phase is red, the number of vehicles leaving is 0). This represents the arrival rate of the i-th phase (in pcu / s). This represents the departure rate of the i-th phase (unit: pcu / s).

[0037] The number of arriving vehicles mentioned above can be the number of vehicles that enter the intersection or the detection range of the radar equipment between the (k-1)th decision point and the kth decision point in the traffic flow corresponding to the i-th phase.

[0038] The number of vehicles leaving can be the number of vehicles in the traffic flow corresponding to the i-th phase that leave the stop line or leave the detection range of the radar equipment between the (k-1)-th decision point and the k-th decision point.

[0039] Reference Figure 3 This application provides a signal sensing control method, wherein the process of determining the optimal decision points for each phase of the current intersection through a global sensing model may include: Step 301: For the i-th phase of the current intersection, predict the number of vehicles to be parked at the k-th decision point of the i-th phase using the global sensing model. Step 302: Determine whether the decision point with the minimum predicted number of vehicles in the current phase is the 0th second. If yes, proceed to step 303; otherwise, continue to maintain the current phase. Step 303: Switch phase.

[0040] For example, the global sensing model predicting the optimal decision point for each phase of the current intersection may include: The current number of vehicles stored at the current intersection, collected by the radar-sensing device at the current moment, is input into the global sensing model; The global sensing model predicts the first arrival rate and first departure rate of the current phase based on the current number of vehicles at the intersection at the current moment and the historical data of the intersection. Based on the flashing duration, the first arrival rate, and the first departure rate, it predicts the predicted number of vehicles remaining after switching phases at multiple decision points. The decision point with the fewest predicted vehicles remaining is selected as the optimal decision point. The flashing duration can include the green light countdown and the yellow light duration. Since the decision range takes the flashing duration into account, it is assumed that the optimal decision point is at the 3rd second. However, since the flashing duration is 6 seconds, it can be assumed that all remaining vehicles can be cleared within 6 seconds. Therefore, the phase can be switched at the current moment without waiting until the 3rd second for another decision.

[0041] Reference Figure 4 For example, the total duration of the decision point predicted by the global sensing model is usually greater than the flashing duration. After the global sensing model determines the phase switching at second 0, the green light in the current phase of the traffic light starts to flash, enters the green light countdown, and then enters the yellow light. At the beginning of the flashing duration of the current phase, that is, after entering the green light countdown, the global sensing model no longer makes a switching decision for the current phase, but instead executes the overlapping phase and makes a prediction of the switching decision for the overlapping phase.

[0042] Reference Figure 5 As shown, since the time taken to switch phases is much less than 1 second, the time to switch phases can be considered as the 0th second after the phase switch is executed following the decision.

[0043] This application embodiment predicts the first arrival rate and the first departure rate of the current phase based on the current number of vehicles at the current intersection at the current time and the historical data of the current intersection. Based on the flashing duration, the first arrival rate, and the first departure rate, it predicts the predicted number of vehicles remaining after switching phases at multiple decision points, and selects the decision point with the fewest predicted vehicles remaining as the optimal decision point. For each decision point in the current phase, Calculate the current phase based on the first arrival rate and the first departure rate, and the first number of arriving vehicles and the first number of departing vehicles from the current moment to the second before the decision point; Based on the saturation flow rate of the current phase detected by the radar vision device, calculate the second number of vehicles departing within the flashing duration for the current phase; Based on the first number of incoming vehicles, the first number of outgoing vehicles, and the second number of outgoing vehicles, the predicted number of vehicles to be stored at the decision point for the current phase is determined.

[0044] See Figure 6 As shown in the optional embodiment of this application, since vehicles may accelerate through the stop line during the flashing duration (green light countdown and yellow light time), the predicted number of vehicles to be stored can be fine-tuned based on the predicted number of vehicles leaving during the flashing duration.

[0045] In this embodiment of the application, the above-mentioned saturation flow rate can be calculated based on the saturation headway and the vehicle departure rate. It is assumed that the saturation headway of the vehicle can be set to a fixed value of 3s.

[0046] If the number of vehicles that can be released at each decision point and the number of vehicles that can be released within the flashing duration are calculated based on the vehicle departure rate and the saturation headway, considering that vehicles will accelerate through during the countdown, in an optional embodiment, the number of vehicles that can be released within the flashing duration can be calculated using the saturation headway. Then, based on the number of vehicles that can be released during the flashing duration and the predicted maximum number of vehicles that can be released at multiple decision points, the number of predicted decision points is determined.

[0047] When the number of vehicles that can be released during the flashing duration exceeds the maximum number of vehicles that can be released at multiple decision points, the number of decision points is increased, the number of prediction units K is increased, and the total prediction duration is extended.

[0048] For example, assuming the saturation headway is 3 seconds per vehicle and the total flashing duration (green light countdown and yellow light) is 15 seconds, then the number of vehicles that can be released during the flashing duration is 5. If there are 3 prediction units, each with a duration of 3 seconds, and a departure rate of 0.5 pcu / s, then the maximum number of vehicles that can be released at the 3 decision points corresponding to the 3 prediction units is 4.5. Since the number of vehicles that can be released during the flashing duration is greater than the maximum number of vehicles that can be released at the 3 decision points, the number of prediction units is increased by one. This increases the maximum number of vehicles in the corresponding 4 decision points to 6, which is greater than the number of vehicles that can be released during the flashing duration, thus completing the adjustment of the number of prediction units K.

[0049] Therefore, when the number of vehicles that can be released during the flashing duration is greater than the maximum number of vehicles that can be released at multiple decision points as predicted, the number of prediction units can be increased until the number of vehicles that can be released during the flashing duration is no greater than the maximum number of vehicles that can be released at multiple decision points as predicted. If the number of vehicles that can be released during the flashing duration is less than the maximum number of vehicles that can be released at the multiple decision points, a phase switch is performed at second 0.

[0050] The first arrival rate of the current phase can be calculated based on the number of vehicles parked and passing through in historical seconds prior to the current time (e.g., 10 seconds, 15 seconds, etc.). For example, the historical time of the i-th phase can be calculated using the following formula:

[0051] in, This indicates the time elapsed from the current moment to a preset historical time point (e.g., 10 seconds, 15 seconds, etc.). This represents the number of vehicles stored at the current moment in the i-th phase. This represents the number of vehicles stored at a preset historical time point in the i-th phase. This represents the number of vehicles that leave during the i-th phase between the current time and a preset historical time point.

[0052] For example, the departure rate of a phase can also be set according to empirical values ​​(the empirical departure rate can be set between [0.1~1], such as 0.3 pcu / s, 0.5 pcu / s, etc.), or the departure rate of a phase can also be calculated based on historical information during the green light of that phase. For example, the departure rate of a phase can be determined by the number of vehicles leaving the i-th phase between the current time and the preset historical time point and the duration between the current time and the preset historical time point. When the departure rate is set according to empirical values, it can be set according to the following correspondence: If the user needs to cut off the discrete traffic flow at the end of the green light, the departure rate should be greater than 0.5 pcu / s. Setting the departure rate higher will result in more vehicles remaining at the start of the countdown. After the discrete traffic flow at the end of the green light gathers, it can be released in the next cycle, which can improve traffic efficiency. If the radar camera detects that there are many large vehicles on site with slow speeds, the departure rate can be set lower. This will result in fewer vehicles remaining at the start of the countdown, ensuring that large vehicles have enough time to pass through the intersection.

[0053] Optionally, the optimal decision points for each phase of the current intersection are predicted using a global sensing model, including: The current number of vehicles stored at the current intersection, collected by the radar-sensing device at the current moment, is input into the global sensing model; The global sensing model predicts the first arrival rate and the first departure rate of the current phase based on the current number of vehicles at the current intersection at the current time and the historical data of the current intersection. Based on the flashing duration, the first arrival rate, and the first departure rate, it predicts the predicted number of vehicles remaining after phase switching at multiple decision points for the current phase. The decision point with the fewest predicted vehicles remaining is selected as the optimal decision point. Specifically, when the radar-guided device detects a large number of discrete green-end traffic flows at the current intersection, the first departure rate is greater than 0.5 pcu / s; when the radar-guided device detects the presence of special vehicle types at the current intersection, the first departure rate is less than 0.3 pcu / s. These special vehicle types can include water trucks, construction vehicles, trucks, and freight trucks, etc., and are not limited here.

[0054] For example, the departure rate of the phase can also be calculated using the following formula. :

[0055] in, This represents the number of vehicles that departed during the i-th phase between the current time and a preset historical time point. This indicates the time elapsed from the current moment to the preset historical time point.

[0056] The global sensing model in this application uses the number of parked vehicles, the number of departing vehicles, and the number of arriving vehicles to predict the number of parked vehicles at multiple future time points, and obtains a decision result: the time point with the fewest predicted parked vehicles is determined as the decision point for switching phases, and the other time points are set as the time points for maintaining phases. After issuing the decision point for switching phases and the corresponding decision result to the traffic lights, the time window is slid according to the decision point.

[0057] In the embodiments of this application, since a flow direction may appear continuously in multiple adjacent phases, if the existence of repeated flow directions is not considered in the single-loop mode, after the repeated flow direction in the current phase clears the queued vehicles, the repeated flow direction may also appear in the overlapping phase when the phase is switched to the overlapping phase. This will result in the repeated flow direction in the overlapping phase being empty.

[0058] To address the above situation, this embodiment compares the vehicle flow direction of the current phase at the intersection with that of the subsequent overlapping phases to determine whether there is an independent flow direction. The flow direction comparison result between the current phase and the overlapping phase adjusts the optimal decision point of the global sensing model, including: Compare the current phase of the intersection with the flow direction of its adjacent phase to obtain the flow direction comparison result; When the flow direction comparison result indicates that there is an independent flow direction between the current phase and its overlapping phase, the duplicate flow direction in the current phase is deleted, and the predicted number of vehicles stored in the current phase at the multiple decision points is predicted based on the independent flow direction and the duplicate flow direction in the overlapping phase. When the flow direction comparison result indicates that there is no independent flow direction between the current phase and its overlapping phase, the predicted number of vehicles stored at the multiple decision points is predicted based on the difference in the number of vehicles stored between the repeated flow directions between the current phase and its overlapping phase.

[0059] Reference Figure 7 As shown, downward-pointing arrows represent northbound traffic flow, and upward-pointing arrows represent southbound traffic flow. In single-loop mode, if the global sensing model's decision-making process does not consider the existence of overlapping traffic flows, the system will only switch to phase 2 after the northbound queue in phase 1 is cleared. This will result in the northbound queue in phase 2 being left empty, increasing the overall cycle of the traffic lights and affecting the traffic flow control effect at the intersection.

[0060] See Figure 8As shown, assuming the current phase is phase 1 and the subsequent overlapping phase is phase 2, when there is an independent flow direction between the current phase and the overlapping phase (i.e., the current phase has a different flow direction than the subsequent overlapping phase), the north-to-north direction in phase 1 can be deleted during the decision-making process because phase 1 and phase 2 have a duplicate north-to-north direction. In this way, the duration of phase 1 is mainly determined by the independent flow direction (north left turn) during the decision-making process.

[0061] See Figure 9 As shown, assuming the current phase is phase 1 and the subsequent overlapping phase is phase 2, when the current phase has no independent flow direction compared to the overlapping phase (i.e., the current phase has no different flow direction compared to the subsequent phase), the duration of phase 1 is determined by the independent flow direction in phase 2 - south-straight. Therefore, the number of vehicles in phase 1 is processed as follows: number of vehicles in phase 1 = number of vehicles in phase 1 north-straight - number of vehicles in phase 2 south-straight.

[0062] Based on the same inventive concept, this application also provides an electronic device, including: at least one memory and at least one processor, wherein the at least one memory stores executable code, and the at least one processor is used to execute the executable code in the at least one memory to implement the above-described signal sensing control method.

[0063] This application also provides a traffic control system, including: traffic lights and the aforementioned electronic equipment.

[0064] At least one of the aforementioned memories can be used to store a computer program, which may include instructions and data to implement the steps of any of the methods described above. The memory may be random access memory, read-only memory, non-volatile, programmable ROM, erasable PROM, electrically erasable, flash memory, optical memory, and registers, etc. The processor 801 may be a general-purpose processor, which is a processor that performs specific steps and / or operations by reading and executing the computer program stored in the memory. The general-purpose processor may use the memory stored in the memory during the execution of the steps and / or operations. The general-purpose processor may be a central processing unit, ASIC, and FPGA, etc. The electronic device may also include a communication interface, which may include input / output interfaces, physical interfaces, and logical interfaces for interconnecting devices within the network device. In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The methods disclosed in the embodiments of this application can be directly implemented by a hardware processor or by a combination of hardware and software modules in the processor.

Claims

1. A signal sensing control method, characterized in that, The method includes: Based on the current number of vehicles parked, approaching, and departing at the intersection, the optimal decision point for each phase of the current intersection is predicted using a global sensing model. Compare the current phase of the current intersection with the flow direction of its overlapping phase to obtain the flow direction comparison result; When the flow direction comparison result indicates that there is an independent flow direction between the current phase and its overlapping phase, the duplicate flow direction in the current phase is deleted, and the predicted number of vehicles in the current phase at multiple decision points is predicted based on the independent flow direction and the duplicate flow direction in the overlapping phase. When the flow direction comparison result indicates that there is no independent flow direction between the current phase and its overlapping phase, the predicted number of vehicles in the current phase at multiple decision points is predicted based on the difference in the number of vehicles in the repeated flow directions between the current phase and its overlapping phase. The decision point corresponding to the minimum predicted number of vehicles to be stored is taken as the optimal decision point; Phase switching is performed based on the adjusted optimal decision point.

2. The method as described in claim 1, characterized in that, The step of predicting the optimal decision points for each phase of the current intersection using the global sensing model includes: For the current intersection, a single-loop mode is used to preset multiple phases, and a model predictive control (MPC) framework is used to predict the number of vehicles stored at multiple decision points for each phase of the current intersection. The decision point corresponding to the minimum predicted number of vehicles is taken as the optimal decision point.

3. The method as described in claim 1, characterized in that, The step of predicting the optimal decision points for each phase of the current intersection using the global sensing model includes: For any phase of the current intersection: After the corresponding key flow direction switches phases at each decision point, the predicted number of vehicles in the key flow direction is calculated. The key flow direction is the flow direction corresponding to the green light of the current phase and / or the overlapping phase. By comparing the predicted number of vehicles at each decision point for the key flow direction, the decision point with the fewest predicted vehicles is taken as the optimal decision point for that phase.

4. The method as described in claim 2, characterized in that, The prediction of the predicted number of vehicles stored at multiple decision points for each phase of the current intersection includes: For each phase of the current intersection, calculate the total number of vehicles in all directions after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the total remaining number of vehicles in the original flow direction after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the total remaining number of vehicles in the corresponding flow direction of that phase and the overlapping phase after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles in the flow direction with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles remaining in the lane corresponding to that phase after switching phases at each decision point, and use this as the predicted number of vehicles remaining in that phase at each decision point; or For each phase at the current intersection, calculate the number of vehicles remaining in the lane with the longest queue length after switching phases at each decision point, and use this as the predicted number of vehicles remaining in that phase at each decision point; or For each phase of the current intersection, calculate the number of vehicles in the lane with the highest number of vehicles after switching phases at each decision point, and use this as the predicted number of vehicles in that phase at each decision point.

5. The method as described in claim 1, characterized in that, The step of predicting the optimal decision points for each phase of the current intersection using a global sensing model includes: The current number of vehicles stored at the current intersection, collected by the radar-sensing device at the current moment, is input into the global sensing model; The global sensing model predicts the first arrival rate and the first departure rate of the current phase based on the current number of vehicles at the current intersection at the current time and the historical data of the current intersection. Based on the flashing duration, the first arrival rate, and the first departure rate, it predicts the predicted number of vehicles after the current phase switches phases at multiple decision points. The decision point with the fewest predicted vehicles is selected as the optimal decision point. Specifically, when the radar-guided device detects a large number of discrete green-end traffic flows at the current intersection, the first departure rate is greater than 0.5 pcu / s. When the radar-guided device detects the presence of special vehicle types at the current intersection, the first departure rate is less than 0.3 pcu / s. The special vehicle types include water trucks, engineering vehicles, trucks, and freight trucks.

6. The method as described in claim 5, characterized in that, The method of predicting the predicted number of vehicles remaining after the current phase switches phases at multiple decision points, based on the flash duration, the first arrival rate, and the first departure rate, includes: For each decision point of the current phase, Calculate the current phase based on the first arrival rate and the first departure rate, and the first number of arriving vehicles and the first number of departing vehicles from the current moment to the second before the decision point; Based on the saturation flow rate of the current phase detected by the radar vision device, calculate the second number of vehicles passing through the current phase within the flashing duration, wherein the flashing duration includes the countdown and / or the yellow light time; Based on the first number of incoming vehicles, the first number of outgoing vehicles, and the second number of outgoing vehicles, the predicted number of vehicles to be stored at the decision point for the current phase is determined.

7. The method as described in claim 1, characterized in that, Before performing a phase switch based on the adjusted optimal decision point, the method further includes: One second before the execution time of the current phase reaches the optimal decision point, the number of vehicles at the current intersection is obtained through the radar camera, input into the global sensing model, and the optimal decision point of each phase of the current intersection is predicted again. If the optimal decision point is predicted again to be at second 0, a phase switch is executed; otherwise, the current phase is extended by one unit of time.

8. The method according to any one of claims 1-7, characterized in that, Using a global sensing model, the optimal decision points for each phase of the current intersection are predicted, including: Multiple decision points are set for the current phase of the current intersection, and the time window between each decision point is fixed; The maximum number of vehicles that can be released at the multiple decision points is calculated based on the vehicle departure rate. The number of vehicles that can be released within the flashing duration is calculated based on the saturated headway. If the number of vehicles that can be released within the flashing duration is greater than the maximum number of vehicles that can be released at the multiple decision points, the number of decision points is increased until the number of vehicles that can be released within the flashing duration is not greater than the predicted maximum number of vehicles that can be released at the multiple decision points. When the number of vehicles that can be released during the flashing duration is less than the maximum number of vehicles that can be released at the plurality of decision points, a phase switch is performed at second 0.

9. An electronic device, characterized in that, include: At least one memory and at least one processor, The at least one memory stores executable code, and the at least one processor executes the executable code in the at least one memory to implement the method as described in claims 1-7.

10. A traffic control system, characterized in that, include: Traffic lights and the electronic device as described in claim 9.

Images