A traffic signal control method and device, electronic equipment and storage medium

Abstract

The application provides a traffic signal control method and device, electronic equipment and storage medium, which can be applied to the field of traffic. The method comprises the following steps: a central agent obtains an estimated vehicle flow under each central traffic phase corresponding to each traffic signal lamp in a target area, and sends the estimated vehicle flow under each central traffic phase to each first agent; continuously receiving first interaction messages sent by each first agent in a message passing manner, wherein the initial first interaction message sent by each first agent comprises predicted congestion information between the first agent and the central agent; when a preset receiving end condition is met, selecting a central traffic phase from the central traffic phases based on the latest received first interaction messages, and controlling the traffic signal lamps in the target area, which can effectively alleviate the traffic congestion problem.

Description

Technical Field

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

[0002] With rapid social development and improved living standards, the increasing number of cars is causing traffic congestion. Currently, intelligent traffic signal control schemes are commonly used to control traffic lights. Specifically, these schemes make real-time control decisions based on dynamically changing traffic flow to alleviate traffic congestion.

[0003] In related technologies, intelligent traffic signal control schemes can employ reinforcement learning methods. By interacting with the traffic environment, they learn the mapping from traffic states to signal action decisions, enabling them to learn decision-making actions in a complex and unknown environment. Specifically, an intelligent agent is deployed at each intersection equipped with traffic lights. This agent makes signal decisions based on the current surrounding traffic conditions, interacts with the environment to receive rewards, and continuously updates and learns reasonable strategies.

[0004] Reinforcement learning methods are effective at learning regular traffic flow changes, but they have the following problems:

[0005] When multiple agents learn simultaneously using multi-agent reinforcement learning, the traffic environment is often non-stationary, resulting in unstable learning outcomes and a tendency to get trapped in local optima. Consequently, the control decisions made by each agent can only alleviate congestion within its own target area, but may not effectively alleviate global traffic congestion.

[0006] For example, to alleviate congestion at the previous intersection, the agent controls a traffic light at that intersection to be green for 40 seconds. At the same time, to alleviate congestion at the next intersection, the agent controls a traffic light at that intersection to be red for 60 seconds. However, the traffic flow that passed through the previous intersection may cause congestion at the next intersection.

[0007] Therefore, it is necessary to redesign a traffic light control scheme to overcome the above-mentioned defects. Summary of the Invention

[0008] This disclosure provides a traffic signal control method, apparatus, electronic device, and storage medium to enable intelligent agents at each intersection to make accurate control decisions, effectively alleviating traffic congestion.

[0009] On one hand, embodiments of this application provide a traffic signal control method applied to a central agent among multiple agents, each agent controlling at least one traffic light within a corresponding area, including:

[0010] For each central traffic phase corresponding to each traffic light within the target area, the estimated traffic flow passing through the target area within the execution cycle of the central traffic phase is obtained; wherein, each central traffic phase represents: a signal combination mode corresponding to each traffic light;

[0011] The estimated traffic flow corresponding to each central traffic phase is sent to at least one adjacent first agent.

[0012] Receive the first interactive messages that are iteratively sent by each first intelligent agent using a message passing method. The initial first interactive message sent by each first intelligent agent includes: the predicted congestion information generated by the first intelligent agent when selecting a matching first traffic phase based on the estimated traffic flow under the central traffic phase for each central traffic phase.

[0013] When the preset reception termination condition is met, a target central traffic phase is selected from the central traffic phases based on the latest received first interaction messages, and the traffic lights in the target area are controlled.

[0014] On one hand, embodiments of this application provide a traffic signal control method applied to a non-central agent among multiple intelligent agents, wherein each intelligent agent is used to control at least one traffic light within a corresponding area, including:

[0015] For each third traffic phase corresponding to each traffic light in the target area, the estimated traffic flow passing through the target area within the execution cycle of the third traffic phase is obtained; wherein, each third traffic phase represents: a signal combination mode corresponding to each traffic light;

[0016] The estimated traffic flow corresponding to each of the third traffic phases is sent to at least one adjacent fourth agent.

[0017] The fourth interaction message is iteratively sent to each of the adjacent fourth agents using a message passing method until a preset reception end condition is met; wherein, the initial fourth interaction message sent to each of the fourth agents includes: for each fourth traffic phase of the fourth agent, based on the estimated traffic flow under the fourth traffic phase, the predicted congestion information generated when selecting the matching third traffic phase.

[0018] The system continuously receives fifth interactive messages iteratively sent by each fourth agent using a message passing method; wherein, the initial fifth interactive message sent by each fourth agent includes: for each third traffic phase of the non-central agent, based on the estimated traffic flow under the third traffic phase, the system selects the matching fourth traffic phase and generates predicted congestion information.

[0019] When the receiving end condition is met, based on the latest received fifth interaction messages, a target third traffic phase is selected from the third traffic phases, and the traffic lights in the target area are controlled.

[0020] On one hand, embodiments of this application provide a traffic signal control device applied to a central intelligent agent among multiple intelligent agents, each of which controls at least one traffic light within a corresponding area, including:

[0021] The first acquisition module obtains the estimated traffic flow passing through the target area within the execution cycle of each central traffic phase corresponding to each traffic light in the target area; wherein, each central traffic phase represents: a signal combination mode corresponding to each traffic light;

[0022] The first sending module is used to send the estimated traffic flow corresponding to each central traffic phase to at least one adjacent first intelligent agent;

[0023] The first receiving module is used to continuously receive the first interactive messages sent by each first intelligent agent using a message passing method. The initial first interactive message sent by each first intelligent agent includes: the predicted congestion information generated by the first intelligent agent when selecting a matching first traffic phase based on the estimated traffic flow under the central traffic phase for each central traffic phase.

[0024] The first selection module is used to select a target central traffic phase from the central traffic phases based on the latest received first interaction messages when the preset reception end condition is met, and to control the traffic lights in the target area.

[0025] Optionally, the iteratively updated first interaction message sent by each of the first agents includes:

[0026] The first agent, in conjunction with at least one adjacent second agent, iteratively sends second interactive messages using a message passing method, generating new predicted congestion information when reselecting a matching first traffic phase for each central traffic phase; and

[0027] Based on the predicted congestion information between the first agent and each second agent obtained from each second interaction message;

[0028] The initial second interaction message sent by each second agent includes: the predicted congestion information generated by the second agent when selecting a matching second traffic phase based on the estimated traffic flow under the first traffic phase for each first traffic phase.

[0029] Optionally, the first selection module is further configured to:

[0030] Based on the latest received first interaction messages, the predicted congestion information in each first interaction message corresponding to each central traffic phase is determined;

[0031] From the various central traffic phases, select a target central traffic phase whose predicted congestion information meets the preset conditions.

[0032] Optionally, the device further includes:

[0033] The third receiving module is used to receive the estimated traffic flow under each first traffic phase sent by each of the first intelligent agents;

[0034] The fourth sending module is used to perform the following operations for each of the first intelligent agents:

[0035] Based on the estimated traffic flow under each first traffic phase of the first intelligent agent and the estimated traffic flow under the target center traffic phase, the predicted congestion information under each first traffic phase is obtained.

[0036] A third interaction message is sent to the first intelligent agent, so that the first intelligent agent selects a first traffic phase from the first traffic phases based on the third interaction message and controls the traffic lights under its jurisdiction; wherein, the third interaction message includes predicted congestion information for each first traffic phase.

[0037] Optionally, the first acquisition module is further configured to:

[0038] Based on the current traffic flow, preset vehicle speed, and traffic turning information of the target area, the estimated traffic flow passing through the jurisdiction area within the execution cycle of each central traffic phase is obtained.

[0039] On one hand, embodiments of this application provide a traffic signal control device applied to a non-central intelligent agent among multiple intelligent agents, wherein the first intelligent agent is used to control at least one traffic light within a target area, including:

[0040] The second acquisition module is used to obtain the estimated traffic flow passing through the target area within the execution cycle of each third traffic phase corresponding to each traffic light in the target area; wherein, each third traffic phase represents a signal combination mode corresponding to each traffic light.

[0041] The second sending module is used to send the estimated traffic flow corresponding to each of the third traffic phases to at least one adjacent fourth agent.

[0042] The third sending module is used to iteratively send fourth interaction messages to each adjacent fourth agent using a message passing method until a preset reception end condition is met; wherein, the initial fourth interaction message sent to each of the fourth agents includes: for each fourth traffic phase of the fourth agent, based on the estimated traffic flow under the fourth traffic phase, the predicted congestion information generated when selecting a matching third traffic phase.

[0043] The second receiving module is used to continuously receive the fifth interaction messages sent iteratively by each of the fourth intelligent agents using a message passing method; wherein, the initial fifth interaction message sent by each of the fourth intelligent agents includes: for each third traffic phase of the non-central intelligent agent, based on the estimated traffic flow under the third traffic phase, the predicted congestion information generated when selecting the matching fourth traffic phase;

[0044] The second selection module is used to select a target third traffic phase from the third traffic phases based on the latest received fifth interaction messages when the receiving end condition is met, and to control the traffic lights in the target area.

[0045] Optionally, the iteratively updated fourth interaction message sent to each of the fourth agents includes:

[0046] The non-central agent, in conjunction with at least one adjacent fifth agent, iteratively sends a sixth interactive message using a message passing method. This generates new predicted congestion information when reselecting a matching third traffic phase for each of the fourth traffic phases of the fourth agent; and

[0047] Based on the predicted congestion information between the non-central agent and each fifth agent obtained from each sixth interaction message;

[0048] The initial sixth interaction message sent by each of the fifth intelligent agents includes: the predicted congestion information generated by the fifth intelligent agent when selecting a matching fifth traffic phase based on the estimated traffic flow under the third traffic phase for each of the third traffic phases.

[0049] Optionally, the second selection module is also used for:

[0050] Based on the latest received fifth interaction messages, determine the predicted congestion information in the fifth interaction messages corresponding to each third traffic phase;

[0051] From the aforementioned third traffic phases, select a target third traffic phase whose predicted congestion information meets preset conditions.

[0052] On one hand, embodiments of this application provide an electronic device including a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of any of the traffic signal control methods described above.

[0053] On one hand, embodiments of this application provide a computer storage medium including a computer program, which, when run on an electronic device, causes the electronic device to perform the steps of any of the traffic signal control methods described above.

[0054] On one hand, embodiments of this application provide a computer program product, which includes a computer program stored in a computer-readable storage medium; when a processor of an electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, causing the electronic device to perform the steps of any of the traffic signal control methods described above.

[0055] Since the embodiments of this application adopt the above-described technical solution, they have at least the following technical effects:

[0056] In the scheme of this application embodiment, the central intelligent agent first obtains the estimated traffic flow under each central traffic phase and sends the estimated traffic flow to each adjacent first intelligent agent. Then, the central intelligent agent continuously receives first interaction messages sent by each adjacent first intelligent agent using a message passing method. The initial first interaction message includes: the predicted congestion information generated by the first intelligent agent for each central traffic phase based on the estimated traffic flow under that central traffic phase when selecting a matching first traffic phase. When a preset reception termination condition is met, the central intelligent agent selects a central traffic phase from each central traffic phase based on the latest received first interaction messages and controls each traffic light in the target area. Therefore, the central intelligent agent can combine the first interaction messages sent by each first intelligent agent to obtain the predicted congestion information between itself and each adjacent first intelligent agent, as well as the predicted congestion information between other intelligent agents transmitted by the first intelligent agent, and then select a suitable central traffic phase based on the predicted congestion information to effectively alleviate traffic congestion problems.

[0057] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a schematic diagram illustrating an application scenario of a traffic signal control method provided in an embodiment of this application.

[0060] Figure 2 A schematic diagram of a connection network for various intelligent agents provided in an embodiment of this application;

[0061] Figure 3 A flowchart of a traffic signal control method provided in this application embodiment;

[0062] Figure 4A A schematic diagram of a crossroads provided for an embodiment of this application;

[0063] Figure 4B A schematic diagram of traffic phases at an intersection provided for an embodiment of this application;

[0064] Figure 5A A schematic diagram of a traffic balance scenario provided in an embodiment of this application;

[0065] Figure 5B A schematic diagram of a traffic imbalance scenario provided in an embodiment of this application;

[0066] Figure 6 A collaborative graph of various intelligent agents is provided in an embodiment of this application;

[0067] Figure 7 A message passing factor diagram for each agent is provided in an embodiment of this application;

[0068] Figure 8 A flowchart of another traffic signal control method provided in this application embodiment;

[0069] Figure 9 A flowchart illustrating another traffic signal control method provided in this application embodiment;

[0070] Figure 10 A structural block diagram of a traffic signal control device provided in an embodiment of this application;

[0071] Figure 11 A structural block diagram of a traffic signal control device provided in an embodiment of this application;

[0072] Figure 12This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;

[0073] Figure 13 This is a schematic diagram of the structure of another electronic device in an embodiment of this application. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0075] To facilitate a better understanding of the technical solutions of this application by those skilled in the art, the terms involved in this application are introduced below.

[0076] Intelligent agents are entities possessing intelligence. Based on the cloud and centered on artificial intelligence, they construct a comprehensive, collaborative, precise, continuously evolving, and open intelligent system. Any independent entity capable of thought and interaction with its environment can be abstracted as an intelligent agent. In information technology, especially in the fields of artificial intelligence and computer science, an intelligent agent can be viewed as anything capable of perceiving its environment through sensors and acting upon that environment through actuators.

[0077] Traffic phase: Each control state of a traffic signal, that is, the combination of different light colors displayed for different directions of various approach lanes, is called a traffic phase.

[0078] MPC: Model Predictive Control, is a control method based on predicting the behavior of the controlled object. It is a real-time control method capable of responding instantly to traffic changes.

[0079] The word “exemplary” as used below means “serving as an example, embodiment, or illustration.” Any embodiment illustrated as an “exemplary” need not be construed as superior to or better than other embodiments.

[0080] The terms "first" and "second" used in this document are for descriptive purposes only and should not be construed as indicating relative importance or implying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0081] This application relates to Intelligent Traffic Systems (ITS), also known as Intelligent Transportation Systems. ITS effectively integrates advanced technologies (information technology, computer technology, data communication technology, sensor technology, electronic control technology, automatic control theory, operations research, artificial intelligence, etc.) into transportation, service control, and vehicle manufacturing. It strengthens the connection between vehicles, roads, and users, thereby forming a comprehensive transportation system that ensures safety, improves efficiency, enhances the environment, and conserves energy. Alternatively;

[0082] Intelligent Vehicle Infrastructure Cooperative Systems (IVICS) are a development direction of Intelligent Transportation Systems (ITS). IVICS utilizes advanced wireless communication and next-generation Internet technologies to implement comprehensive, real-time dynamic information exchange between vehicles and infrastructure. Based on the collection and fusion of dynamic traffic information across all times and spaces, it conducts active vehicle safety control and cooperative road management, fully realizing effective collaboration between people, vehicles, and roads. This ensures traffic safety, improves traffic efficiency, and ultimately forms a safe, efficient, and environmentally friendly road traffic system.

[0083] In related technologies, intelligent traffic signal control schemes typically employ reinforcement learning methods. By interacting with the traffic environment, they learn the mapping from traffic states to signal action decisions, enabling them to learn decision-making actions in a complex and unknown environment. Specifically, an intelligent agent is deployed at each intersection equipped with traffic lights. This agent makes signal decisions based on the current surrounding traffic conditions, interacts with the environment to receive rewards, and continuously updates and learns reasonable strategies.

[0084] Reinforcement learning methods are effective at learning regular traffic flow changes, but they have the following problems:

[0085] When multiple agents learn simultaneously using multi-agent reinforcement learning, the traffic environment is often non-stationary, resulting in unstable learning outcomes and a tendency to get trapped in local optima. Consequently, the control decisions made by each agent can only alleviate congestion within its own target area, but may not effectively alleviate global traffic congestion.

[0086] Furthermore, reinforcement learning methods require deploying simulators to perform millions of traffic flow simulations in order to learn a good strategy for an intersection, and the number of simulations increases exponentially with the number of agents. It is also difficult to be fully applicable to online traffic control scenarios, such as when sudden traffic events cause unexpected changes in traffic flow (e.g., traffic accidents or sudden road closures).

[0087] To address the aforementioned issues, this application implements an intelligent traffic signal control scheme based on model predictive control (MMC) methods from the control field. For example, traffic conditions at one or more intersections can be predicted by constructing linear programming or Markov decision process models, and efficient dynamic programming methods can be used to generate real-time strategies. Furthermore, considering that MMC is a real-time control method capable of responding instantly to traffic changes, but centralized computational methods place excessive demands on computation and communication, affecting the scalability of the scheme, online distributed MMC schemes are necessary for network-scale traffic control scenarios. This means that multiple agents control the traffic lights at each intersection in parallel.

[0088] In view of this, embodiments of this application provide a traffic signal control method, apparatus, electronic device, and storage medium. The central intelligent agent first estimates the traffic flow in each central traffic phase during the next execution cycle and sends the estimated traffic flow to each adjacent first intelligent agent. Then, it continuously receives first interaction messages sent by each adjacent first intelligent agent using a message passing method. When a preset reception termination condition is met, the central intelligent agent selects a central traffic phase from the central traffic phases based on the latest received first interaction messages and controls the traffic lights in the target area. In this way, the central intelligent agent can combine the first interaction messages from each first intelligent agent to obtain predicted congestion information between itself and each adjacent first intelligent agent, and then select a suitable central traffic phase based on the predicted congestion information, effectively alleviating traffic congestion problems.

[0089] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application. Furthermore, the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0090] like Figure 1 The diagram illustrates an application scenario in this application. This scenario includes multiple intelligent agents 110, each positioned at an intersection with traffic lights and connected to intelligent agents 110 at adjacent intersections, forming a traffic control network. One of the multiple intelligent agents 110 serves as the central intelligent agent. Adjacent intelligent agents 110 can communicate with each other via a communication network. Optionally, the communication network can be a wired network or a wireless network. Adjacent intelligent agents 110 can be directly or indirectly connected via wired or wireless communication; this application does not impose any limitations on this.

[0091] Some methods in the embodiments of this application can be executed by the intelligent agent 110 as the central intelligent agent, while other methods can be executed by the intelligent agent 110 as a non-central intelligent agent. Each intelligent agent 110 can estimate the traffic flow in each central traffic phase in the next execution cycle and communicate with neighboring intelligent agents 110 to determine the optimal control decision, that is, to select the target traffic phase from multiple traffic phases to control each traffic light.

[0092] The embodiments of this application do not limit the number of intelligent agents 110, nor do they limit the topology structure formed by multiple intelligent agents 110. Those skilled in the art will understand that flexible selection can be made as needed in different application scenarios.

[0093] It should be noted that, Figure 1 This is an example of the application scenario of the traffic signal control method of this application. However, the application scenarios to which the method in the embodiments of this application can be applied are not limited to this.

[0094] The specific implementation methods of the traffic signal control method according to the embodiments of this application will be described below.

[0095] The traffic signal control method of this application embodiment can be applied to traffic signal control scenarios in a designated area. For example, the designated area can be a city or a region of a city, without limitation. It obtains information on all intersections equipped with traffic lights within the designated area, as well as the traffic network formed by these intersections. An intelligent agent is set up at each intersection to control the traffic lights. Based on the traffic network of each intersection, a connection network of the intelligent agents can be constructed. A central intelligent agent is selected from among the intelligent agents based on this connection network.

[0096] For example, assuming each intersection is a crossroads, the connection network of each intelligent agent is as follows: Figure 2 As shown, this includes agents a1-a9. Specifically, for any agent a... i and a j Let its shortest path be d. ij Define intelligent agents As the central intelligent agent, the central intelligent agent satisfies the distance... The furthest agent d ij Minimum; Selected Next, the diameter of the connected network is defined as the distance. The farthest node d ij ,exist Figure 2 In this context, a5 is a medium-sized intelligent agent, and the diameter of the network it connects to is 2.

[0097] It should be noted that the above Figure 2The structure of the connection network in the text is only an example, and the specific structure can be determined according to the actual traffic network. It is not limited here.

[0098] Figure 3 The diagram illustrates a traffic signal control method provided in an embodiment of this application. This method can be executed by a central intelligent agent, which controls at least one traffic light within a target area.

[0099] like Figure 3 As shown, the traffic signal control method of this application embodiment may include the following steps S301-S304:

[0100] Step S301: The central agent obtains the estimated traffic flow passing through the target area within the execution cycle of each central traffic phase corresponding to each traffic light in the target area. Each central traffic phase represents a signal combination mode corresponding to each traffic light.

[0101] The target area can be an intersection, and the traffic lights include multiple central traffic phases. Figure 4A Taking the intersection shown as an example, the east-west lanes include l1, l2, l3, l4, l5, and l6; the north-south lanes include l7, l8, l9, l10, l11, and l12; as shown... Figure 4B As shown, various central traffic phases can include: traffic phase 1 (l2, l5), traffic phase 2 (l1, l6), traffic phase 3 (l8, l11), and traffic phase 4 (l9, l10). Traffic phase 1 refers to having a green light in the l2 and l5 directions, traffic phase 2 refers to having a green light in the l1 and l6 directions, traffic phase 3 refers to having a green light in the l8 and l11 directions, and traffic phase 4 refers to having a green light in the l9 and l10 directions.

[0102] Figure 4B The traffic phases of the intersection shown in the image are just examples. The specific phases can be determined based on the actual situation of the intersection. For example, in addition to the intersection, it can also include a three-way intersection. There is no limitation on this.

[0103] Optionally, the estimated traffic flow corresponding to each central traffic phase can be obtained in the following way: based on the current traffic flow, preset speed and traffic turning information of the target area, the estimated traffic flow passing through the jurisdiction within the execution cycle of each central traffic phase can be obtained.

[0104] The current traffic flow can include the traffic flow in each lane of the target area. Specifically, the location of each vehicle can be determined based on its location information. The preset speed can be the saturation flow speed of traffic movement. For example, if the maximum number of vehicles passing through within a set time period is known, the saturation flow speed can be obtained based on the number of vehicles and the set time period. The traffic flow turning information can be obtained based on the turning of each vehicle, which can be obtained based on the vehicle's navigation information.

[0105] The execution cycle of each central traffic phase can be set as needed, for example, 20 seconds. Based on the current traffic flow in each lane of the target area, the turning direction of each vehicle in each lane, and the preset speed, the estimated traffic flow passing through the target area within the execution cycle of each central traffic phase can be obtained.

[0106] Step S302: The estimated traffic flow corresponding to each central traffic phase is sent to at least one adjacent first agent.

[0107] The central agent can pre-determine the adjacent first agents based on the aforementioned connection network of multiple agents. This connection network includes the connection relationships between multiple agents, such as... Figure 2 In the network shown, the central agent is a5, and the first agents adjacent to a5 include a2, a4, a6, and a8.

[0108] Step S303: Receive the first interactive messages that each first agent iteratively sends using a message passing method. The initial first interactive message sent by each first agent includes: the predicted congestion information generated by the first agent when selecting a matching first traffic phase based on the estimated traffic flow under the central traffic phase for each central traffic phase.

[0109] The message passing method can be understood as follows: In the first round of message passing, each first agent sends a first interaction message to the adjacent central agent. In the first round of message passing, each first agent sends the second interaction message received from other agents, as well as the first interaction message, to the central agent, and so on.

[0110] In this embodiment, for the updated number of vehicles on each road at the intersection managed by each agent, a load balancing function is calculated. This load balancing function is used to evaluate the quality of the traffic conditions reflected by the estimated traffic flow under the corresponding traffic phase. For example, the load balancing function of agent i is defined as follows (1):

[0111] B i (t)=∑ (l,h):l∈I(i),h∈∈O(i) |q(l,h)| 2 (1)

[0112] Where q(l,h) represents the updated traffic flow of lane (l,h), l represents the lanes entering the intersection, and h represents the lanes exiting the intersection. The smaller the value of this function, the more balanced the traffic conditions of each lane at the intersection, and the smaller the average waiting time of the traffic flow. Figure 5A As shown, in the traffic balance scenario, the traffic flow in each lane of the intersection is relatively balanced, and the average waiting time for vehicles is short; for example... Figure 5B As shown, in traffic imbalance scenarios, the traffic flow in each lane of the intersection is uneven, and the average waiting time for vehicles is relatively long.

[0113] In step S303 above, in the initial first interaction message sent by each first agent, the predicted congestion information generated by the first agent when selecting the matching first traffic phase based on the estimated traffic flow under the central traffic phase for each central traffic phase can be calculated based on the above load balancing function. Specifically, the above predicted congestion information is calculated by the following formula (2):

[0114] c ij (x i ,x j )=∑ h∈O(i) [q(l ij ,h)] 2 +∑ h∈O(j) [q(l ji ,h)] 2 (2)

[0115] Where, x i x represents the first traffic phase of the first intelligent agent. j c represents the second traffic phase of the central agent. ij (x i ,x j The first agent assumes that the central agent executes x. j Choose to execute x i The predicted congestion information generated in time, q(l ij q(l) represents the estimated traffic flow of the corresponding lane from the first agent to the central agent. ji ,h) represents the estimated traffic flow of the corresponding lane from the central agent to the first agent.

[0116] Based on the above Figure 2 Taking the connected network shown as an example, congestion information can be predicted between every two adjacent agents based on the above equation (2), and the following can be obtained: Figure 6 The collaboration diagram shown.

[0117] Furthermore, the message passing order among multiple agents can be determined based on the distance between each agent in the network and the central agent, for example: in the above... Figure 6In the diagram, the central agent is a5. For two adjacent agents a1 and a4, if the distance between a1 and a5 is greater than the distance between a4 and a5, then a1 will send a message to a4. Based on this, Figure 6 The collaboration graph shown is converted into a message passing factor graph, where each node represents an agent, and the predicted congestion information between nodes is represented as function nodes, such as... Figure 7 As shown, each agent can determine neighboring agents with whom it needs to exchange messages based on the message passing factor graph.

[0118] Assuming, Represents agent a i Subsequent neighboring agents, Represents agent a i The predecessor's neighboring agents. In each iteration of message passing, each agent a... i Repeatedly collect Neg prev (i) Send the message and send the message to Neg foll (i).

[0119] In step S303 above, the iteratively updated first interaction message sent by each first agent may include the following A1 and A2:

[0120] A1. The first intelligent agent, together with at least one adjacent second intelligent agent, iteratively sends second interactive messages using a message passing method, generating new predicted congestion information when reselecting a matching first traffic phase for each central traffic phase.

[0121] Among them, the new predicted congestion information in A1 can also be calculated based on the above equation (2), except that the first agent can select a new first traffic phase for each central traffic phase.

[0122] A2. The predicted congestion information between the first agent and each second agent obtained based on each second interaction message; wherein, the initial second interaction message sent by each second agent includes: the predicted congestion information generated by the second agent when selecting a matching second traffic phase based on the estimated traffic flow under the first traffic phase for each first traffic phase.

[0123] For example: the first intelligent agent uses Figure 7 Taking a4 as an example, the first interaction message sent by the first agent a4 to the central agent a5 includes c 45That is, the first agent assumes that the central agent executes each central traffic phase x5, and selects the predicted congestion information generated when executing the first traffic phase x6. Simultaneously with the first agent a4 sending the first interaction message to the central agent a5, the first agent a4 receives second interaction messages sent by adjacent second agents a1 and a7, respectively. The second interaction message sent by the second agent a1 contains c. 14 The second interactive message sent by the second intelligent agent a7 contains c 47 Furthermore, based on the received second interaction messages, the first intelligent agent a4 redetermines the new c. 45 That is, assuming the central agent executes each central traffic phase x5, and selects to execute the new first traffic phase x6, the predicted congestion information generated respectively.

[0124] The first interactive message sent by the first agent a4 to the central agent a5 after the second iteration update includes: the updated c 45 c 14 and c 47 ; where c 14 and c 47 With the above c 45 The meaning is similar and will not be elaborated here.

[0125] exist Figure 7 In this scenario, since the second agents a1 and a7 have no preceding neighboring agents, they do not receive any interaction messages. The second interaction message sent iteratively to the first agent a4 remains unchanged. It can be seen that the number of iterations for message passing is equal to the diameter of the connection network of multiple agents, i.e., the distance between the agent farthest from the central agent and the central agent. For example... Figure 2 If the diameter of the connection network in the network is 2, then the number of message passing iterations is 2.

[0126] For example, in each iteration of message passing, the first agent a i Towards the central intelligent agent a j Send the first interactive message Q ij Q ij It can be calculated using the following formula (3):

[0127]

[0128] Among them, Q ij The meaning is that, assuming the central agent a... j Make a decision x j The first intelligent agent a i Collaborate with this value to select decision x i To minimize the accumulated cost during message delivery, which includes: predicting congestion information cij (x i ,x j ) and the first intelligent agent a i Received from Neg prev (i) cumulative cost.

[0129] Step S304: When the preset reception termination condition is met, select a target central traffic phase from the central traffic phases based on the latest received first interaction messages, and control the traffic lights in the target area.

[0130] In this step, the receiving end condition can be that the number of iterations reaches a set number, which can be determined based on the connection network of multiple agents, as described in the above embodiment.

[0131] In some embodiments, selecting a target central traffic phase from the central traffic phases based on the latest received first interaction messages in step S304 above may include the following steps B1-B2:

[0132] B1. Based on the latest received first interaction messages, determine the predicted congestion information in each first interaction message corresponding to each central traffic phase;

[0133] B2. From each central traffic phase, select a target central traffic phase whose predicted congestion information meets the preset conditions.

[0134] Specifically, multiple agents interact with each other until the propagated messages converge, i.e., the number of iterations reaches a set number. Finally, the central agent selects the central traffic phase that minimizes the sum of costs received in the last iteration as its optimal control decision, i.e., the target central traffic phase.

[0135] In this embodiment, the central agent can combine with each first agent to send a first interactive message using a message passing method to obtain the predicted congestion information between itself and each of the adjacent first agents, as well as the predicted congestion information between other agents transmitted by the first agents. Then, based on the predicted congestion information, a suitable central traffic phase can be selected to effectively alleviate traffic congestion problems.

[0136] In some embodiments, after determining the target central traffic phase, the central agent can interact with each of the adjacent first agents, and the first agents can interact with each of the adjacent second agents, and so on. After the message transmission is completed, each first agent can select the target first traffic phase and control the traffic lights under its jurisdiction. Similarly, the second agents and other agents can also make control decisions.

[0137] Optionally, the central agent method may also perform the following steps C1-C2:

[0138] C1. Receive the estimated traffic flow for each first traffic phase sent by each first intelligent agent.

[0139] C2. For each first agent, perform the following operations C21-C22 respectively:

[0140] C21. Based on the estimated traffic flow under each first traffic phase of the first intelligent agent and the estimated traffic flow under the target center traffic phase, obtain the predicted congestion information under each first traffic phase.

[0141] Among them, the central agent has determined the target central traffic phase to be executed. Specifically, the above formula (2) can be used to calculate the predicted congestion information under each first traffic phase when the target central traffic phase cooperates.

[0142] C22. Send a third interaction message to the first intelligent agent so that the first intelligent agent selects a first traffic phase from the first traffic phases based on the third interaction message and controls the traffic lights under its jurisdiction; wherein, the third interaction message includes predicted congestion information for each first traffic phase.

[0143] Specifically, in the above embodiments, multiple agents interact with each other until the propagated messages converge, that is, the number of iterations reaches a set number. Finally, the central agent selects the target central traffic phase based on the first interaction message received in the last iteration, which can be used as the first stage of message transmission.

[0144] Next, in the second phase of message passing, the message passing direction of the first phase is reversed, that is, the central agent interacts with the first agent, the first agent interacts with the second agent, and so on, and the message passing process of the first phase is repeated. The difference from the first phase is that the control decision of the central agent in the second phase (i.e. the target central traffic phase) has been determined.

[0145] Specifically, multiple agents interact with each other by sending messages until the propagated messages converge, i.e., the number of iterations reaches a set number. Finally, each agent selects the decision that minimizes the sum of costs received in the last iteration as its optimal control decision.

[0146] Since the control decision of the central agent (i.e. the target central traffic phase) has been determined, the third interactive message received by the first agent from the central agent remains unchanged. Based on the third interactive message, a target first traffic phase is selected from each first traffic phase, and the traffic lights under its jurisdiction are controlled.

[0147] In this embodiment, intelligent agents are deployed at each intersection equipped with traffic lights within a designated area, and each agent makes real-time signal decisions in a distributed manner. Each agent senses the current traffic flow information at its intersection to predict traffic flow for the next time period, and communicates with agents at adjacent intersections to coordinate and plan the optimal decision, effectively alleviating traffic congestion.

[0148] In this embodiment, time can be divided into multiple time segments t = 1, 2, ..., T, each time segment having a set duration, such as 20 seconds, without limitation. Within each time segment, the central agent performs a signal decision using the traffic signal control method described in the above embodiment, i.e., selects the target central traffic phase.

[0149] Based on the same inventive concept, this application provides a traffic signal control method applied to a non-central intelligent agent among multiple intelligent agents. The non-central intelligent agent can be any intelligent agent other than the central intelligent agent, and each intelligent agent is used to control at least one traffic light in a corresponding area.

[0150] like Figure 8 As shown in the figure, a traffic signal control method provided in this application includes the following steps S801-S805:

[0151] Step S801: For each third traffic phase corresponding to each traffic light in the target area, obtain the estimated traffic flow passing through the target area within the execution cycle of the third traffic phase; wherein, each third traffic phase represents: a signal combination mode corresponding to each traffic light.

[0152] The third traffic phase controlled by the non-central agent is similar to the central traffic phase controlled by the central agent in the above embodiments, and will not be described again here.

[0153] Optionally, the estimated traffic flow through the target area within the execution cycle of each third traffic phase can be obtained in the following way: based on the current traffic flow, preset speed and traffic turning information of the target area, the estimated traffic flow through the jurisdiction within the execution cycle of each central traffic phase can be obtained respectively. See the above embodiment for details, which will not be repeated here.

[0154] Step S802: The estimated traffic flow corresponding to each third traffic phase is sent to at least one adjacent fourth agent.

[0155] Step S803: The fourth interaction message is iteratively sent to each of the adjacent fourth agents using a message passing method until the preset reception end condition is met; wherein, the initial fourth interaction message sent to each fourth agent includes: for each fourth traffic phase of the fourth agent, the predicted congestion information generated when selecting the matching third traffic phase based on the estimated traffic flow under the fourth traffic phase.

[0156] Step S803 can be understood as the message passing process in the first stage of the above embodiments. Each fourth agent can be understood as the successor of the executing entity (non-central agent) of step S803 and its adjacent agents.

[0157] Step S804: Continuously receive the fifth interaction messages sent iteratively by each fourth agent using message passing; wherein, the initial fifth interaction message sent by each fourth agent includes: for each third traffic phase of the non-central agent, the predicted congestion information generated when selecting the matching fourth traffic phase based on the estimated traffic flow under the third traffic phase.

[0158] Step S803 can be understood as the second-stage message passing process in the above embodiments.

[0159] Furthermore, the iteratively updated fourth interaction message sent to each fourth agent includes the following D1 and D2:

[0160] D1. The non-central agent, together with at least one adjacent fifth agent, iteratively sends a sixth interactive message using message passing, generating new predicted congestion information when reselecting a matching third traffic phase for each fourth traffic phase of the fourth agent.

[0161] Each fifth agent can be understood as a neighboring agent that is the predecessor of the non-central agent.

[0162] D2. Predicted congestion information between the non-central agent and each fifth agent obtained from each sixth interaction message; wherein, the initial sixth interaction message sent by each fifth agent includes: predicted congestion information generated by the fifth agent when selecting a matching fifth traffic phase based on the estimated traffic flow under the third traffic phase for each third traffic phase.

[0163] Wherein, D1 and D2 are similar to A1 and A2 in the above embodiments. Specifically, the non-central agent can be understood as the first agent in A1, the fifth agent can be understood as the second agent in A1, and the sixth interaction message can be understood as the second interaction message in A1.

[0164] Step S805: When the reception termination condition is met, based on the latest received fifth interaction messages, select a target third traffic phase from the third traffic phases and control the traffic lights in the target area.

[0165] In some embodiments, selecting a target third traffic phase from among the third traffic phases based on the latest received fifth interaction messages and controlling the traffic lights in the target area may include the following steps E1-E2:

[0166] E1. Based on the latest received fifth interaction messages, determine the predicted congestion information in each fifth interaction message corresponding to each third traffic phase;

[0167] E2. From each third traffic phase, select a target third traffic phase whose predicted congestion information meets the preset conditions.

[0168] The specific implementation process of steps E1-E2 can be found in steps B1-B2 of the above embodiment. In the message transmission process of the second stage in the above embodiment, multiple agents interact with each other until the propagated messages converge, that is, the number of iterations reaches the set number. Finally, the non-central agent selects the target third traffic phase based on the fifth interaction message received in the last iteration and controls each traffic light in the target area.

[0169] Specifically, the non-central agent selects the third traffic phase that minimizes the sum of costs from the fifth interaction messages received in the last iteration as its optimal decision, i.e., the target third traffic phase, based on all costs in the fifth interaction messages received in the last iteration. These costs include the predicted congestion information received by the non-central agent from the fourth agents.

[0170] The following is combined with Figure 9 The traffic signal control process of the non-centralized intelligent agent in the embodiments of this application will be described.

[0171] Figure 9 The traffic signal control process shown can be executed by a non-centralized intelligent agent, including the following steps S901-S912:

[0172] Step S901: Before the previous round of decision-making ends, begin a new round of decision-making.

[0173] In this embodiment, time can be divided into multiple time segments t = 1, 2, ..., T, each time segment having a set duration, such as 20 seconds, without limitation. Within each time segment, the non-central agent performs a signal decision using the traffic signal control method described in the above embodiment, that is, selects the target traffic phase (which can be understood as the aforementioned target third traffic phase).

[0174] For example, start a new round of decision-making a few seconds before the end of the previous round of decision-making (e.g., the first 3 seconds).

[0175] Step S902: Obtain the current traffic flow information at the intersection.

[0176] The current traffic flow information may include the current traffic volume, preset vehicle speed, and traffic flow turning information in the above embodiments.

[0177] Step S903: Enumerate each traffic phase of the adjacent intersection and estimate the traffic flow passing through the intersection within the execution cycle of each traffic phase based on the current traffic flow information.

[0178] The method for estimating traffic flow can be found in the above embodiments, and will not be repeated here.

[0179] Step S904: Initialize the predicted congestion information between the agent and its successor neighboring agents.

[0180] Specifically, the initial predicted congestion information between the agent and its successor neighboring agents is as follows: for each traffic phase of the successor neighboring agent, the predicted congestion information is generated when a matching traffic phase is selected; the initial successor interaction message is obtained based on the predicted congestion information under each traffic phase.

[0181] Step S905: Send a successor interaction message to the successor neighboring agent and receive a predecessor interaction message sent by the predecessor neighboring agent.

[0182] Step S906: Update the successor interaction message based on the received predecessor interaction message.

[0183] Steps S905-S906 can be understood as the first stage of the message transmission process in the above embodiments.

[0184] Step S907: Determine whether the number of convergence steps for message passing has been reached. If yes, proceed to step S908; otherwise, return to step S905.

[0185] The number of convergences is the set number in the above embodiments, which can be specifically determined based on the connection network of multiple agents.

[0186] Step S908: Send a predecessor interaction message to the predecessor neighboring agent and receive a successor interaction message sent by the successor neighboring agent.

[0187] Step S909: Update the predecessor interaction message based on the received successor interaction message.

[0188] Steps S908-S909 can be understood as the second-stage message passing process in the above embodiments.

[0189] Step S910: Determine whether the number of convergence attempts for message passing has been reached. If yes, proceed to step S911; otherwise, return to step S908.

[0190] Step S911: Based on the received subsequent interaction messages, select the target traffic phase as the final decision.

[0191] Step S912: Execute the target traffic phase continuous setting duration.

[0192] The duration can be set as needed, for example, 20 seconds, and there is no limit to it.

[0193] In this embodiment, intelligent agents are set up at each intersection. Relevant traffic flow information is input to the intelligent agents, and the intelligent agents communicate with each other via message passing to coordinate distributed constraint optimization. After the communication ends, each intelligent agent decides the final decision for the intersection (selecting the target traffic phase). The target traffic phase is sustained for Ts (e.g., T = 20s), and a new round of distributed calculation for phase decision is performed before the end of each duration.

[0194] In practical applications, the traffic signal control scheme of this application embodiment was deployed on a traffic simulation platform and compared with other schemes in related technologies. The other schemes compared included a first traditional method (referred to as Method 1) and a second traditional method (referred to as Method 2), a first single-agent reinforcement learning method (referred to as Method 3) and a second single-agent reinforcement learning method (referred to as Method 4), a first multi-agent reinforcement learning method (referred to as Method 5) and a second multi-agent reinforcement learning method (referred to as Method 6). The test dataset included three virtual road data grids of different sizes. 4x4 Grid 15x15 and Grid 20x20 The data also includes road network data from four real cities: City 1, City 2, City 3, and City 4, with road network sizes of 3x4, 4x4, 16x3, and 1x16, respectively.

[0195] As shown in Table 1, by comparing the average travel time of the solution in this application embodiment with other solutions under virtual road data and real urban road network data, it can be seen that the solution in this application embodiment is superior to other solutions.

[0196] Table 1

[0197]

[0198] Furthermore, comparing the global pressure value of the system at each moment under the road network data of the two cities with the solution of this application embodiment and the above-mentioned method two, the pressure value of the solution of this application embodiment is lower than that of the above-mentioned method two for most of the time.

[0199] Based on the same inventive concept, this application provides a traffic signal control device applied to a central intelligent agent among multiple intelligent agents, each intelligent agent being used to control at least one traffic light in a corresponding area.

[0200] like Figure 10 As shown in the figure, a traffic signal control device provided in this application embodiment includes a first acquisition module 101, a first transmission module 102, a first receiving module 103, and a first selection module 104.

[0201] The first acquisition module 101 obtains the estimated traffic flow passing through the target area within the execution cycle of each central traffic phase corresponding to each traffic light in the target area; wherein, each central traffic phase represents: a signal combination mode corresponding to each traffic light;

[0202] The first sending module 102 is used to send the estimated traffic flow corresponding to each central traffic phase to at least one adjacent first intelligent agent respectively;

[0203] The first receiving module 103 is used to continuously receive the first interactive messages sent by each first intelligent agent using a message passing method. The initial first interactive message sent by each first intelligent agent includes: the predicted congestion information generated by the first intelligent agent when selecting a matching first traffic phase based on the estimated traffic flow under the central traffic phase for each central traffic phase.

[0204] The first selection module 104 is used to select a target central traffic phase from each central traffic phase based on the latest received first interaction messages when the preset reception end condition is met, and control each traffic light in the target area.

[0205] Optionally, the iteratively updated first interaction message sent by each first agent includes:

[0206] The first agent, in conjunction with at least one adjacent second agent, iteratively sends second interactive messages using a message-passing method. This generates new predicted congestion information when reselecting a matching first traffic phase for each central traffic phase.

[0207] Based on the predicted congestion information between the first agent and each second agent obtained from each second interaction message;

[0208] The initial second interaction message sent by each second agent includes: the predicted congestion information generated by the second agent when selecting a matching second traffic phase based on the estimated traffic flow under the first traffic phase for each first traffic phase.

[0209] Optionally, the first selection module 104 is also used for:

[0210] Based on the latest received first interaction messages, determine the predicted congestion information in each first interaction message corresponding to each central traffic phase;

[0211] From all the central traffic phases, select a target central traffic phase whose predicted congestion information meets the preset conditions.

[0212] Optionally, the device also includes:

[0213] The third receiving module is used to receive the estimated traffic flow under each first traffic phase sent by each first intelligent agent;

[0214] The fourth sending module is used to perform the following operations for each first agent:

[0215] Based on the estimated traffic flow under each first traffic phase of the first intelligent agent and the estimated traffic flow under the target center traffic phase, the predicted congestion information under each first traffic phase is obtained.

[0216] A third interactive message is sent to the first intelligent agent so that the first intelligent agent can select one first traffic phase from all first traffic phases based on the third interactive message and control the traffic lights under its jurisdiction; wherein, the third interactive message includes the predicted congestion information under each first traffic phase.

[0217] Optionally, the first acquisition module is also used for:

[0218] Based on the current traffic flow, preset speed, and traffic turning information of the target area, the estimated traffic flow passing through the jurisdiction area within the execution cycle of each central traffic phase is obtained.

[0219] Based on the same inventive concept, this application provides a traffic signal control device applied to a non-central intelligent agent among multiple intelligent agents, wherein the first intelligent agent is used to control at least one traffic light within a target area.

[0220] like Figure 11 As shown in the figure, this application provides a traffic signal control device, including a second acquisition module 111, a second transmission module 112, a third transmission module 113, a second receiving module 114, and a second selection module 115.

[0221] The second acquisition module 111 is used to obtain the estimated traffic flow passing through the target area within the execution cycle of each third traffic phase corresponding to each traffic light in the target area; wherein, each third traffic phase represents: a signal combination mode corresponding to each traffic light.

[0222] The second sending module 112 is used to send the estimated traffic flow corresponding to each third traffic phase to at least one adjacent fourth intelligent agent respectively;

[0223] The third sending module 113 is used to iteratively send fourth interaction messages to each adjacent fourth agent using a message passing method until a preset reception end condition is met; wherein, the initial fourth interaction message sent to each fourth agent includes: for each fourth traffic phase of the fourth agent, based on the estimated traffic flow under the fourth traffic phase, the predicted congestion information generated when selecting the matching third traffic phase.

[0224] The second receiving module 114 is used to continuously receive the fifth interactive messages sent iteratively by each fourth intelligent agent using a message passing method; wherein, the initial fifth interactive message sent by each fourth intelligent agent includes: for each third traffic phase of the non-central intelligent agent, the predicted congestion information generated when selecting the matching fourth traffic phase based on the estimated traffic flow under the third traffic phase.

[0225] The second selection module 115 is used to select a target third traffic phase from each third traffic phase based on the latest received fifth interaction messages when the receiving end condition is met, and to control each traffic light in the target area.

[0226] Optionally, the iteratively updated fourth interaction message sent to each fourth agent includes:

[0227] The non-central agent, in conjunction with at least one neighboring fifth agent, iteratively sends a sixth interactive message using message passing. This generates new predicted congestion information when reselecting a matching third traffic phase for each fourth traffic phase of the fourth agent; and...

[0228] Based on the predicted congestion information between the non-central agent and each fifth agent obtained from the sixth interaction messages;

[0229] The initial sixth interaction message sent by each fifth agent includes: the predicted congestion information generated by the fifth agent when selecting a matching fifth traffic phase based on the estimated traffic flow under the third traffic phase for each third traffic phase.

[0230] Optionally, the second selection module 115 is also used for:

[0231] Based on the latest received fifth interaction messages, determine the predicted congestion information in each fifth interaction message corresponding to each third traffic phase;

[0232] From each of the third traffic phases, select a target third traffic phase whose predicted congestion information meets the preset conditions.

[0233] For ease of description, the above sections are divided into modules (or units) according to their functions and described separately. Of course, in implementing this application, the functions of each module (or unit) can be implemented in one or more software or hardware components.

[0234] Regarding the apparatus in the above embodiments, the specific execution methods of each module have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0235] Those skilled in the art will understand that various aspects of this application can be implemented as devices, methods, or computer program products, with each part described separately as a module according to its function. Of course, in implementing this application, the functions of each module can be implemented in one or more software or hardware components. Therefore, various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, collectively referred to herein as a "circuit," "module," or "system."

[0236] Based on the same inventive concept as the above-described method embodiments, this application also provides an electronic device. The principle of this electronic device in solving the problem is similar to that of the above-described method. Therefore, the implementation of this electronic device can refer to the implementation of the above-described method, and repeated details will not be described again.

[0237] See Figure 12 As shown, the electronic device 120 may include at least a processor 121 and a memory 122. The memory 122 stores a computer program, which, when executed by the processor 121, causes the processor 121 to perform the steps of any of the traffic signal control methods described above.

[0238] In an exemplary embodiment, this application also provides a storage medium including a computer program, such as a memory 122 including a computer program, which can be executed by a processor 121 of an electronic device 120 to perform the traffic signal control method described above. Optionally, the storage medium can be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0239] The following reference Figure 13 To describe an electronic device 130 according to this embodiment of the present application. Figure 13 The electronic device 130 is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0240] like Figure 13The electronic device 130 is manifested in the form of a general electronic device. The components of the electronic device 130 may include, but are not limited to: at least one processing unit 131, at least one storage unit 132, and a bus 133 connecting different system components (including storage unit 132 and processing unit 131).

[0241] Bus 133 represents one or more of several bus structures, including a memory bus or memory controller, peripheral bus, processor, or local bus using any of the various bus structures.

[0242] Storage unit 132 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 1321 and / or cache memory 1322, and may further include read-only memory (ROM) 1323.

[0243] Storage unit 132 may also include a program / utility 1325 having a set (at least one) of program modules 1324, such program modules 1324 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0244] Electronic device 130 can also communicate with one or more external devices 134 (e.g., keyboard, pointing device, etc.), and with one or more devices that enable a user to interact with electronic device 130, and / or with any device that enables electronic device 130 to communicate with one or more other electronic devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 135. Furthermore, electronic device 130 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 136. As shown, network adapter 136 communicates with other modules used in electronic device 130 via bus 133. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0245] Based on the same inventive concept as the above-described method embodiments, this application provides a computer program product comprising a computer program stored in a computer-readable storage medium. A processor of an electronic device reads the computer program from the computer-readable storage medium and executes the computer program, causing the electronic device to perform the steps of any of the above-described traffic signal control methods.

[0246] Computer program products may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0247] The computer program product of the embodiments of this application may employ a portable compact disc read-only memory (CD-ROM) and include a computer program, and may run on an electronic device. However, the computer program product of this application is not limited thereto. In this document, the readable storage medium may be any tangible medium that contains or stores a program that may be used by or in conjunction with a command execution system, apparatus, or device.

[0248] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a readable computer program. This propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting a program for use by or in conjunction with a command execution system, apparatus, or device.

[0249] Computer programs contained on readable media may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0250] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0251] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A traffic signal control method, characterized in that, A central agent is applied to multiple agents within a defined area, the central agent being located at the center of a network connecting the multiple agents, each agent controlling at least one traffic light at an intersection within the defined area, including: For each central traffic phase corresponding to each traffic light at the target intersection within the defined area, the estimated traffic flow of each lane passing through the target intersection within the execution cycle of the central traffic phase is obtained; wherein, each central traffic phase represents: a signal combination mode corresponding to each traffic light; The estimated traffic flow of each lane corresponding to each central traffic phase is sent to at least one adjacent first agent. The system receives first interaction messages iteratively sent by each first agent using a message passing method. Each first agent's first round of first interaction messages includes: predicted congestion information generated when selecting a matching first traffic phase for each central traffic phase, based on the estimated traffic flow of the corresponding lane from the first agent to the central agent and the estimated traffic flow of the corresponding lane from the central agent to the first agent. Each subsequent round of iteratively updated first interaction messages includes: updated predicted congestion information obtained by the first agent based on the received second interaction messages iteratively sent by at least one adjacent second agent using a message passing method, and predicted congestion information between the first agent and each second agent obtained based on each second interaction message. The type of the second interaction message is the same as the type of the first interaction message. When the preset reception termination condition is met, a target central traffic phase is selected from the central traffic phases based on the latest round of received first interaction messages, and the traffic lights at the target intersection are controlled.

2. The method according to claim 1, characterized in that, The step of selecting a target central traffic phase from the central traffic phases based on the latest round of received first interactive messages includes: Based on the latest round of received first interaction messages, the predicted congestion information in each of the first interaction messages corresponding to each central traffic phase is determined; From the various central traffic phases, select a target central traffic phase whose predicted congestion information meets the preset conditions.

3. The method according to claim 1, characterized in that, The method further includes: Receive the estimated traffic flow for each first traffic phase sent by each of the first intelligent agents; For each of the first intelligent agents, perform the following operations: Based on the estimated traffic flow under each first traffic phase of the first intelligent agent and the estimated traffic flow under the target center traffic phase, the predicted congestion information under each first traffic phase is obtained. A third interaction message is sent to the first intelligent agent, so that the first intelligent agent selects a first traffic phase from the first traffic phases based on the third interaction message and controls the traffic lights under its jurisdiction; wherein, the third interaction message includes predicted congestion information for each first traffic phase.

4. The method according to claim 1, characterized in that, Obtain the estimated traffic flow for each lane passing through the target intersection within the execution cycle of the central traffic phase, including: Based on the current traffic flow, preset vehicle speed, and traffic turning information at the target intersection, the estimated traffic flow of each lane passing through the target intersection is obtained within the execution cycle of each central traffic phase.

5. A traffic signal control method, characterized in that, A non-central agent is applied to a plurality of agents within a defined area, wherein the central agent is located at the center of the network connecting the agents, and each agent controls at least one traffic light at an intersection within the defined area, including: For each third traffic phase corresponding to each traffic light in the target intersection, the estimated traffic flow of each lane passing through the target intersection within the execution cycle of the third traffic phase is obtained; wherein, each third traffic phase represents: a signal combination mode corresponding to each traffic light; The estimated traffic flow for each lane corresponding to each of the third traffic phases is sent to at least one adjacent fourth agent. The system iteratively sends fourth interaction messages to each adjacent fourth agent using a message passing method until a preset reception termination condition is met. The first round of fourth interaction messages sent to each fourth agent includes: predicted congestion information generated when selecting a matching third traffic phase for each fourth traffic phase of the fourth agent, based on the estimated traffic flow of the corresponding lane from the non-central agent to the fourth agent and the estimated traffic flow of the corresponding lane from the fourth agent to the non-central agent. Each subsequent round of iteratively updated fourth interaction messages includes: updated predicted congestion information obtained by the non-central agent combining sixth interaction messages iteratively sent using a message passing method with at least one adjacent fifth agent, and predicted congestion information between the non-central agent and each fifth agent obtained based on each sixth interaction message. The type of the sixth interaction message is the same as the type of the fourth interaction message. It continuously receives fifth interaction messages iteratively sent by each of the fourth intelligent agents using a message passing method; wherein the type of the fifth interaction message is the same as the type of the fourth interaction message; When the receiving end condition is met, based on the latest round of received fifth interaction messages, a target third traffic phase is selected from the third traffic phases, and the traffic lights at the target intersection are controlled.

6. The method according to claim 5, characterized in that, The step of selecting a target third traffic phase from the third traffic phases based on the latest round of received fifth interaction messages includes: Based on the latest round of received fifth interaction messages, determine the predicted congestion information in the fifth interaction messages corresponding to each third traffic phase; From the aforementioned third traffic phases, select a target third traffic phase whose predicted congestion information meets preset conditions.

7. A traffic signal control device, characterized in that, A central agent is applied to multiple agents within a defined area, the central agent being located at the center of a network connecting the multiple agents, each agent controlling at least one traffic light at an intersection within the defined area, including: The first acquisition module is used to obtain the estimated traffic flow of each lane passing through the target intersection within the execution cycle of each central traffic phase corresponding to each traffic light at the target intersection within the set area; wherein, each central traffic phase represents a signal combination mode corresponding to each traffic light. The first sending module is used to send the estimated traffic flow of each lane corresponding to each central traffic phase to at least one adjacent first intelligent agent. The first receiving module is configured to continuously receive first interaction messages sent by each first intelligent agent using a message passing method. The first interaction message sent by each first intelligent agent in the first round includes: predicted congestion information generated when the first intelligent agent selects a matching first traffic phase for each central traffic phase, based on the estimated traffic flow of the corresponding lane from the first intelligent agent to the central intelligent agent and the estimated traffic flow of the corresponding lane from the central intelligent agent to the first intelligent agent. The first interaction message updated in each subsequent round includes: updated predicted congestion information obtained by the first intelligent agent based on the received second interaction messages iteratively sent by at least one adjacent second intelligent agent using a message passing method, and predicted congestion information between the first intelligent agent and each second intelligent agent obtained based on each second interaction message. The type of the second interaction message is the same as the type of the first interaction message. The first selection module is used to select a target central traffic phase from the central traffic phases based on the latest round of received first interaction messages when the preset reception end conditions are met, and to control the traffic lights at the target intersection.

8. A traffic signal control device, characterized in that, A non-central agent is applied to a plurality of agents within a defined area, wherein the central agent is located at the center of the network connecting the agents, and each agent controls at least one traffic light at an intersection within the defined area, including: The second acquisition module is used to obtain the estimated traffic flow of each lane passing through the target intersection within the execution cycle of each third traffic phase corresponding to each traffic light in the target intersection; wherein, each third traffic phase represents: a signal combination mode corresponding to each traffic light; The second sending module is used to send the estimated traffic flow of each lane corresponding to each of the third traffic phases to at least one adjacent fourth agent. The third sending module is used to iteratively send fourth interaction messages to each adjacent fourth agent using a message passing method until a preset reception termination condition is met. The first round of fourth interaction messages sent to each fourth agent includes: predicted congestion information generated when selecting a matching third traffic phase for each fourth traffic phase of the fourth agent, based on the estimated traffic flow of the corresponding lane from the non-central agent to the fourth agent and the estimated traffic flow of the corresponding lane from the fourth agent to the non-central agent. Each subsequent round of iteratively updated fourth interaction messages includes: updated predicted congestion information obtained by the non-central agent combining sixth interaction messages iteratively sent using a message passing method with at least one adjacent fifth agent, and predicted congestion information between the non-central agent and each fifth agent obtained based on each sixth interaction message. The type of the sixth interaction message is the same as the type of the fourth interaction message. The second receiving module is used to continuously receive the fifth interaction message sent iteratively by each of the fourth intelligent agents using a message passing method; wherein the type of the fifth interaction message is the same as the type of the fourth interaction message; The second selection module is used to select a target third traffic phase from the third traffic phases based on the latest round of received fifth interaction messages when the receiving end condition is met, and to control the traffic lights at the target intersection.

9. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of any of the methods described in claims 1 to 6.

10. A computer-readable storage medium, characterized in that, It includes a computer program that, when run on an electronic device, causes the electronic device to perform the steps of any of the methods described in claims 1 to 6.

11. A computer program product, characterized in that, It includes a computer program stored in a computer-readable storage medium; when the processor of the electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, causing the electronic device to perform the steps of any one of claims 1 to 6.

Images