Intelligent traffic signal control system

Abstract

The application provides a traffic signal lamp intelligent regulation system, and relates to the technical field of traffic regulation.The system comprises a traffic information collection unit and an upper computer software system;by fusing modern frontier technologies such as the Internet of Things, big data and geomagnetic field measurement, and combining scientific theories such as automatic control theory, operations research and artificial intelligence, a traffic signal lamp self-adaptive control system based on period time expansion is constructed, and the green light release time is automatically adjusted according to the real-time arrival condition of the traffic flow at the intersection.

Description

Technical Field

[0001] This invention relates to the field of traffic control technology, and in particular to a smart traffic signal control system. Background Technology

[0002] With the continuous improvement of people's living standards in my country, the number of cars has been growing rapidly year by year. The most prominent social problem brought about by the surge in car ownership is traffic congestion, and intersections have become the key locations in every city most prone to traffic congestion. Currently, traffic lights at intersections in China all adopt a fixed-duration working mode. At most, some more humane intersections simply switch between different fixed durations at different times.

[0003] Since 2019, several cities have attempted to implement "smart traffic light" systems to address traffic congestion. These systems, using high-definition cameras and millimeter-wave radar at intersections, acquire real-time traffic flow data and dynamically adjust traffic light durations accordingly, initially showing significant results. However, these projects have failed to gain wider adoption, primarily due to the following issues:

[0004] 1. High construction costs. Millimeter-wave radar and high-definition cameras are both precision data acquisition devices, which are relatively expensive. Furthermore, to ensure data validity, radar and cameras must be installed in pairs in each direction at intersections, resulting in the cost of data acquisition equipment alone reaching tens of thousands of yuan per intersection. In addition, the processing of radar and vision data requires enormous computing power, necessitating high-performance servers, which, due to their high cost, further increase the system's construction costs.

[0005] 2. Significant Construction Challenges. Millimeter-wave radar and high-definition cameras generate large amounts of data, which wireless transmission bandwidth cannot meet. Therefore, separate wired power and data transmission lines need to be laid. Furthermore, to ensure data complementarity between paired radars and cameras, the data from both devices must be in the same coordinate system, making accurate calibration extremely difficult.

[0006] 3. Low system stability. Radar waves are reflected or refracted, and are also subject to interference from electromagnetic signals of the same frequency in the environment, leading to false alarms or missed detections of targets; visual detection is greatly affected by visibility, and will basically lose its imaging ability in rainy, foggy, light-in-saturation, or when sunlight shines directly on the lens. Therefore, it is greatly affected by the environment and has poor system stability.

[0007] 4. High maintenance costs. Over time, due to the influence of external environmental factors such as temperature, humidity, and thermal expansion and contraction, the data consistency of paired radars and cameras will deteriorate. Therefore, continuous matching and calibration of the paired radars and cameras are necessary during use. In addition, the server software used to process radar and camera data also requires regular maintenance to ensure the accuracy of data complementarity between the paired devices, resulting in high maintenance costs.

[0008] 5. The control methods are too simplistic. The number of intersections in the cities that have tried the "smart traffic light" system is relatively small, making it impossible to coordinate multiple intersections within the area. Furthermore, in order to save costs, the construction units have used simple core algorithms, resulting in an overly simplistic control method.

[0009] 6. Promotion costs increase exponentially. The computing power requirements of recognition systems based on millimeter-wave radar and high-definition cameras increase arithmetically with the increase in quantity. However, the price of servers increases geometrically with the increase in computing power. Therefore, the promotion cost is not just an arithmetic progression, but increases geometrically with the increase in the number of intersections where the system is applied. This is one of the important reasons why existing "smart traffic light" systems often stop at trial construction and have not been widely promoted. Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides a smart traffic light control system. By integrating cutting-edge technologies such as the Internet of Things, big data, and geomagnetic field measurement, and combining them with scientific theories such as automatic control theory, operations research, and artificial intelligence, a traffic light adaptive control system based on periodic time expansion is constructed. This system enables traffic lights to automatically adjust green light times based on real-time traffic flow at intersections.

[0011] A smart traffic signal control system, specifically comprising: a traffic information acquisition unit and a host computer software system;

[0012] The traffic information collection unit includes several wireless geomagnetic sensors, which are installed at set intervals on the center line of the lanes at the intersection to detect the length of the convoy in each approach lane and to label the wireless geomagnetic sensors according to their deployment locations.

[0013] The host computer software system includes a network information transmission unit, a decision management unit, and a front-end signal control unit;

[0014] The network information transmission unit receives the detection data from the wireless geomagnetic sensor and sends the detection data to the decision management unit;

[0015] The decision management unit calculates the detection data, the corresponding wireless geomagnetic sensor number, and the deployment location information using a fuzzy time series prediction model to obtain real-time vehicle congestion data at the intersection. Based on this, it uses a traffic light intelligent control algorithm to calculate the optimal red and green light duration under the current congestion situation, converts it into control commands, and sends them to the front-end signal control unit.

[0016] The fuzzy time series prediction model combines fuzzy theory and time series analysis. It applies the characteristics of fuzzy binary functions to the estimation of empirical model parameters. By introducing the concept of ignoring, it simulates actual test data as fuzzy data knowledge, and then automatically identifies model parameters through fuzzy inference for fuzzy time series prediction.

[0017] The traffic light control algorithm divides the intersection into four directions: east, south, west, and north. Since right turns are not restricted by traffic lights, and the green light times for north-to-east and south-to-west turns are the same, as are the green light times for west-to-north and east-to-south turns, the intersection is designed as a simplified model including four directions: north-south straight, north-south left turn, east-west straight, and east-west left turn. This simplified model uses these directions as the optimization benchmark unit, constructing an extended network based on General Graph Optimization (GEM). For each node, a copy is added to the extended network. All node copies in the extended network have the same time step, which serves as the green light time benchmark period, determined by the traffic flow acquisition period and the traffic light control period. The start time of the green light for each direction is considered the start time of a green light time benchmark period. To obtain real-time traffic flow information for that direction, several equally spaced detection time points are set within the green light time. While continuously predicting real-time traffic flow for each direction, the green light time for each direction is allocated based on the currently acquired traffic flow information. During this period, the following constraints are set:

[0018] (1) The green light signal for each direction can only switch to green light once per cycle.

[0019] (2) Green lights cannot be displayed simultaneously in intersecting directions to ensure traffic safety; green lights must be displayed simultaneously in the same direction.

[0020] (3) There must be sufficient time to clear the intersection before any other direction turns green; each green and red phase must have a minimum duration, which must take into account the time required for vehicles to react to changes in green light timing and the safety of pedestrians in the intersection.

[0021] The intelligent traffic signal control system adopts a star topology, with the decision management unit at the center. It obtains traffic flow information by connecting to a cluster of traffic information collection units through a network information transmission unit, and finally sends the actual control signal to the front-end signal control unit to realize traffic light control.

[0022] The input to the software system is traffic flow information acquired by the traffic information collection unit. The traffic information collection unit uses a wireless geomagnetic sensor to detect traffic flow information and detects the length of the convoy on each entrance lane. This information is used as prior knowledge for the decision-making and management unit, and a fuzzy time series prediction model is used to predict the traffic flow.

[0023] The output of the software system is the front-end signal control unit, which controls the traffic lights at the intersection. The control signals come from the decision-making unit, which adaptively adjusts the remaining time of the traffic lights in each direction based on the traffic flow information at the intersection.

[0024] The decision management unit, as the core of the entire system, receives data from the traffic information collection unit, sends control signals to the front-end signal control unit, and calculates and automatically updates the green light time for each direction based on traffic flow information in each direction and a fixed green light time reference cycle.

[0025] The front-end signal control unit receives control commands to achieve real-time control of each traffic light at the intersection.

[0026] The beneficial effects of adopting the above technical solution are as follows:

[0027] This invention provides a smart traffic signal control system, which has the following advantages:

[0028] 1. Low construction cost. Geomagnetic sensors are mature technologies for measuring the Earth's magnetic field. They are inexpensive and perform stably. Data from each sensor can be used directly after simple conversion, requiring almost no computing power. There's no need to purchase separate servers for computational support, making the hardware cost far lower than radar / camera solutions. Furthermore, the number of geomagnetic sensors deployed is related to the number of lanes in each direction at the intersection. At intersections with multiple lanes, the number of sensors is lower than with radar / camera solutions; intersections with fewer lanes further reduce construction costs.

[0029] 2. Low construction difficulty. The amount of data information from geomagnetic sensors is extremely low, and the bandwidth requirements for data transmission can be met through wireless IoT. Moreover, geomagnetic sensors consume very little power; a 38AH lithium-ion battery can meet the needs for several years. Therefore, geomagnetic sensors are easy to install and maintain, eliminating the need for laying wired lines for power supply or transmission, and eliminating the need to close lanes. They cause minimal damage to the road surface (ground-mounted types do not damage the road surface, and buried types only require drilling a hole 11.5 cm in diameter and 11 cm deep in the road surface). During maintenance, only the geomagnetic sensor needs to be inspected, and the sensor is not affected by road movement and is not easily damaged.

[0030] 3. High system stability. The geomagnetic sensor detects ferromagnetic objects by utilizing changes in the Earth's magnetic field as they pass through, thus remaining unaffected by weather conditions and exhibiting high system stability. By adjusting the sensitivity settings, the size of ferromagnetic objects can be identified, allowing for a rough determination of the vehicle type. Furthermore, the geomagnetic sensor does not react to non-ferromagnetic objects, effectively reducing false detections.

[0031] 4. Low maintenance cost. The geomagnetic sensors operate independently, requiring no systematic comprehensive maintenance and are not connected by wires. Therefore, only periodic battery replacement (once every few years) or occasional equipment failure repair is needed, making maintenance simple and extremely low-cost.

[0032] 5. Diverse control methods. Based on a system design aimed at large-scale promotion, it combines individual intersection control with coordinated control of multiple intersections within a region, offering diverse control methods that are more practical and can meet various control needs under different road conditions, time periods, and scenarios.

[0033] 6. Facilitates widespread application. Low construction costs, extremely low construction and maintenance difficulty, and extremely high system stability form the basis for the system's widespread application. Attached Figure Description

[0034] Figure 1 This is a structural diagram of the intelligent traffic signal control system in an embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram of the hardware layout of the intelligent traffic signal control system in an embodiment of the present invention;

[0036] Figure 3 This is a traffic model diagram of a crossroads in an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram illustrating that green lights cannot be displayed simultaneously in intersecting directions according to an embodiment of the present invention;

[0038] Figure 5 This is a schematic diagram of a green wave road segment in an embodiment of the present invention. Detailed Implementation

[0039] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0040] A smart traffic signal control system, such as Figure 1 As shown, it specifically includes: a traffic information collection unit and a host computer software system;

[0041] The traffic information collection unit includes several wireless geomagnetic sensors, such as... Figure 2 As shown, the wireless magnetic sensors are installed at set intervals on the center line of the lanes at the intersection to detect the length of the convoy in each entrance lane. The wireless magnetic sensors are then labeled according to their deployment locations.

[0042] The Earth's magnetic field remains relatively constant within a few kilometers, but large ferromagnetic objects can cause significant disturbances. Geomagnetic sensors can detect changes in the Earth's magnetic field as small as one six-thousandth of a second. The impact of a vehicle passing through the magnetic field is a fraction of its intensity, making geomagnetic sensors highly sensitive for vehicle detection. Geomagnetic sensors detect vehicles by detecting the disturbances they cause to the Earth's magnetic field. At intersections, a number of wireless geomagnetic sensors are deployed at specific intervals along the center lines of the lanes and numbered. These sensors transmit real-time data via a network information transmission unit (NB-IoT, Narrow Band Internet of Things) at a specified frequency to a decision management unit. The decision management unit uses a dedicated mathematical model to calculate the received data, along with the corresponding geomagnetic sensor numbers and deployment locations, to obtain real-time traffic congestion data for the intersection. Based on this data, a core algorithm is used to calculate the optimal traffic light duration for the current congestion situation and sends this information to the front-end signal control unit. The front-end signal control unit then controls the traffic lights at the intersection in real-time according to the instructions.

[0043] The host computer software system includes a network information transmission unit, a decision management unit, and a front-end signal control unit;

[0044] The network information transmission unit receives the detection data from the wireless geomagnetic sensor and sends the detection data to the decision management unit;

[0045] The decision management unit calculates the detection data, the corresponding wireless geomagnetic sensor number, and the deployment location information using a fuzzy time series prediction model to obtain real-time vehicle congestion data at the intersection. Based on this, it uses a traffic light intelligent control algorithm to calculate the optimal red and green light duration under the current congestion situation, converts it into control commands, and sends them to the front-end signal control unit.

[0046] The fuzzy time series prediction model combines fuzzy theory and time series analysis to solve the prediction problem of uncertain time series data. Unlike traditional time series analysis, which solves for hard constraints on definite parameters, the fuzzy time series prediction model applies the characteristics of fuzzy binary functions to the estimation of empirical model parameters. By introducing the concept of ignoring, it simulates actual test data as fuzzy data knowledge, and then uses fuzzy inference to automatically identify model parameters for fuzzy time series prediction.

[0047] The traffic light control algorithm divides the intersection into four directions: east, south, west, and north. Figure 3 As shown, since right turns are not restricted by traffic lights, and the green light times for north-to-east and south-to-west turns are the same, as are the green light times for west-to-north and east-to-south turns, the intersection is designed as a simplified intersection model including four directions: north-south straight, north-south left turn, east-west straight, and east-west left turn. The simplified intersection model breaks down the intersection into directions as the baseline unit for optimization, and uses these directions as nodes to build an extended network based on General Graph Optimization. When extending adaptive control from a single traffic light intersection to a multi-intersection traffic light problem based on dynamic flow control, a time-extended network is used to solve the problem. To create an extended network from a simple network, for each node, a replica of that node is added to the extended network. All node replicas in the extended network have the same time step, which serves as the green light time reference period, determined by the traffic flow acquisition period and the traffic light control period. The start time of the green light for each direction is considered the start time of a green light time reference period. To obtain real-time traffic flow information for that direction, several equally spaced detection time points are set within the green light time. While continuously predicting real-time traffic flow for each direction, the green light time for each direction is allocated based on the currently acquired traffic flow information. Figure 4 , Figure 5 As shown, the following constraints are set during this period:

[0048] (1) The green light signal for each direction can only switch to green light once per cycle.

[0049] (2) Green lights cannot be displayed simultaneously in intersecting directions to ensure traffic safety; green lights must be displayed simultaneously in the same direction to form a green wave and improve traffic efficiency.

[0050] (3) There must be sufficient time to clear the intersection before any other direction turns green; each green and red phase must have a minimum duration, which must take into account the time required for vehicles to react to changes in green light timing and the safety of pedestrians in the intersection.

[0051] The constructed cyclic time-extended network model can be used to simultaneously optimize traffic flow and traffic signals; this model allows for rigorous mathematical programming. Therefore, a solver can be used to guarantee the optimality of the solutions for the remaining green light times in different directions at each traffic light intersection. This is a significant advantage compared to other methods using genetic programming or similar heuristics.

[0052] The intelligent traffic signal control system adopts a star topology, with the decision management unit at the center. It obtains traffic flow information by connecting to a cluster of traffic information collection units through a network information transmission unit, and finally sends the actual control signal to the front-end signal control unit to realize traffic light control.

[0053] The input to the software system is traffic flow information acquired by the traffic information collection unit. The traffic information collection unit uses a wireless geomagnetic sensor to detect traffic flow information and detects the length of the convoy on each entrance lane. This information is used as prior knowledge for the decision-making and management unit, and a fuzzy time series prediction model is used to predict the traffic flow.

[0054] The output of the software system is the front-end signal control unit, which controls the traffic lights at the intersection. The control signals come from the decision-making unit, which adaptively adjusts the remaining time of the traffic lights in each direction based on the traffic flow information at the intersection.

[0055] The decision management unit, as the core of the entire system, receives data from the traffic information collection unit, sends control signals to the front-end signal control unit, and calculates and automatically updates the green light time for each direction based on traffic flow information in each direction and a fixed green light time reference cycle.

[0056] The front-end signal control unit receives control commands to achieve real-time control of each traffic light at the intersection.

[0057] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A smart traffic signal control system, characterized in that, This includes traffic information collection units and host computer software systems; The traffic information collection unit includes several wireless geomagnetic sensors, which are installed at set intervals on the center line of the lanes at the intersection to detect the length of the convoy in each approach lane and to label the wireless geomagnetic sensors according to their deployment locations. The host computer software system includes a network information transmission unit, a decision management unit, and a front-end signal control unit. The network information transmission unit receives the detection data from the wireless geomagnetic sensor and sends the detection data to the decision management unit. The decision management unit uses a fuzzy time series prediction model to calculate the real-time traffic congestion data of the intersection based on the detection data, the corresponding wireless geomagnetic sensor number, and the deployment location information. Then, based on this, it uses a traffic light intelligent control algorithm to calculate the optimal traffic light duration under the current congestion situation, converts it into control commands, and sends them to the front-end signal control unit. The intelligent traffic light control system adopts a star architecture, with the decision management unit as the center. It connects to the deployed traffic information collection unit cluster through the network information transmission unit to obtain traffic flow information, and finally sends the actual control signals to the front-end signal control unit to realize traffic light control. The fuzzy time series prediction model combines fuzzy theory and time series analysis. It applies the characteristics of fuzzy binary functions to the estimation of empirical model parameters. By introducing the concept of ignoring, it simulates actual test data as fuzzy data knowledge, and then automatically identifies model parameters through fuzzy inference for fuzzy time series prediction. The traffic light control algorithm divides the intersection into four directions: east, south, west, and north. Since right turns are not restricted by traffic lights, and the green light times for north-to-east and south-to-west turns are the same, as are the green light times for west-to-north and east-to-south turns, the intersection is designed as a simplified model including four directions: north-south straight, north-south left turn, east-west straight, and east-west left turn. This simplified model uses these directions as the optimization benchmark unit, constructing an extended network based on General Graph Optimization (GEM). For each node, a copy is added to the extended network. All node copies in the extended network have the same time step, which serves as the green light time benchmark period, determined by the traffic flow acquisition period and the traffic light control period. The start time of the green light for each direction is considered the start time of a green light time benchmark period. To obtain real-time traffic flow information for that direction, several equally spaced detection time points are set within the green light time. While continuously predicting real-time traffic flow for each direction, the green light time for each direction is allocated based on the currently acquired traffic flow information. During this period, the following constraints are set: (1) The green light signal for each direction can only switch to green once per cycle; (2) Green lights cannot be displayed simultaneously in intersecting directions to ensure traffic safety; green lights must be displayed simultaneously in the same direction. (3) There must be sufficient time to clear the intersection before any other direction turns green; each green and red phase must have a minimum duration, which must take into account the time required for vehicles to react to the change in green light time and the safety of pedestrians in the intersection. The input to the host computer software system is the traffic flow information obtained by the traffic information collection unit. The traffic information collection uses a wireless geomagnetic sensor to sense and detect the traffic flow information, detect the length of the platoon on each entrance lane, use it as the prior knowledge of the decision management unit, and use a fuzzy time series prediction model to predict the traffic flow. The output of the host computer software system is the front-end signal control unit, which is the control signal for each traffic light at the intersection. The control signal comes from the decision unit, which adaptively adjusts the remaining time of the traffic lights in each direction based on the traffic flow information at the intersection. The decision management unit receives data from the traffic information collection unit, sends control signals to the front-end signal control unit, and calculates and automatically updates the green light time for each direction based on traffic flow information in each direction and a fixed green light time reference cycle. The front-end signal control unit receives control commands to achieve real-time control of each traffic light at the intersection.

Images