Park traffic signal control method and system

Abstract

This invention provides a method and system for traffic signal control within a park, belonging to the field of traffic signal control within a park. The method includes: receiving logistics task demand data packets from vehicles; receiving real-time traffic status data packets from traffic lights; employing an algorithm model to minimize the priority-weighted total delay of all vehicles as the optimization objective, and solving for traffic signal control parameters within a future decision cycle based on the logistics task demand data packets and the real-time traffic status data packets; and sending the traffic signal control parameters for controlling the traffic lights. This invention can improve the overall logistics efficiency within a park.

Description

Technical Field

[0001] This invention relates to the field of traffic control technology within industrial parks, and more specifically, to a traffic signal control method and system for industrial parks. Background Technology

[0002] Within closed or semi-closed industrial parks and warehousing and logistics bases, efficient internal logistics transportation is crucial for ensuring production continuity and reducing operating costs. Currently, traffic management and logistics transportation systems within these parks are typically planned, constructed, and operated independently. Traffic signal control systems mostly rely on preset timing schemes (fixed timing) or simple vehicle detection methods (such as inductive loop detectors) to manage right-of-way at intersections; while logistics transportation systems (TMS) focus on order dispatch, vehicle scheduling, and static route planning, with their optimization goals often limited to locally minimizing the distance or time traveled by a single vehicle.

[0003] This systemic fragmentation leads to two core technical problems: First, high-priority emergency transportation tasks (such as urgent production line materials and emergency supplies) cannot obtain priority at the traffic signal control level, making it difficult to guarantee their critical timeliness and posing a risk of production interruption; Second, logistics vehicles can only passively adapt to fixed or delayed traffic signals, resulting in frequent starts and stops and long waiting times at intersections, increasing overall transportation time, operating costs, and energy emissions.

[0004] To alleviate the above problems, existing technologies have attempted optimizations from the perspectives of hardware linkage or software path planning, but none have achieved system-level collaboration. A detailed analysis follows: Existing technology 1: Hardware linkage device based on physical triggering A typical prior art technology, as described in patent CN115100877A, provides a physical trigger-based intersection warning device. This solution embeds a follow-up pressure plate at the road intersection. When a vehicle passes, a mechanical transmission activates the circuit, illuminating the corresponding warning light to alert oncoming vehicles from the intersecting direction. Essentially, this solution is an isolated hardware device with direct "vehicle-light" linkage.

[0005] The existing technology has the following drawbacks: (1) Information is one-way and has very low dimension: The device can only sense the single physical event of "a vehicle passing by", and cannot obtain any upper-level logistics information such as vehicle identification, transportation task, destination, priority, etc. Its decision-making logic is based on simple mechanical or circuit triggers and cannot perform complex optimization calculations.

[0006] (2) Lack of collaborative optimization capability: Its control logic is fixed in advance (such as first-come, first-served), and it cannot be dynamically and intelligently adjusted according to the real-time global traffic flow status and the scheduling instructions of the logistics system. For example, it cannot extend the green light time for specific emergency vehicles.

[0007] (3) System isolation: The device is an independent piece of hardware and has not been integrated with the logistics transportation management system (TMS) at the upper level of the park in any way. It is a typical "information island" and cannot participate in the global scheduling and optimization at the park level.

[0008] Existing technology 2: Path planning method based on software avoidance Another representative existing technology, as described in patents CN119600835A and CN119962782A, represents the mainstream software optimization approach. This type of technology collects information such as road vehicle density and average speed using devices like cameras and GPS, and combines this with historical data to predict traffic conditions (including traffic light delays). When planning routes, the estimated traffic light waiting time is used as a cost factor to select the route with the lowest total time or cost for logistics vehicles. This is a strategy that allows vehicles to proactively "avoid" congestion or red lights.

[0009] The existing technology has the following drawbacks: (1) Unidirectional Adaptive Logic: The optimization logic of this type of technology is to make logistics vehicles adapt to a given and unchanging traffic signal environment. The timing scheme of traffic signals is regarded as a fixed and external constraint. The route planning system can only passively predict and avoid, and cannot actively influence or optimize the signal control strategy.

[0010] (2) Lack of reverse interaction mechanism: The route planning system does not feed back its scheduling plan (e.g., which high-priority vehicles will pass through which intersections at what times) to the traffic signal control system. Therefore, the signal system cannot obtain the demand for important logistics tasks in advance and cannot prepare or provide passage convenience for them.

[0011] (3) Local optimization may lead to global inefficiency: When multiple vehicle systems adopt similar avoidance strategies, all vehicles may converge on a few "optimal paths" evaluated by the algorithm, thereby causing new and more severe congestion. This strategy cannot achieve overall load balancing of the park's road network and is a local and passive optimization method.

[0012] In summary, existing technologies have failed to break down the information barriers between logistics scheduling systems and traffic control systems, lacking a mechanism capable of achieving deep, intelligent, and two-way collaboration to fundamentally improve the overall efficiency of logistics transportation within the park. Therefore, a new method to solve these technical problems is urgently needed. Summary of the Invention

[0013] The purpose of this invention is to provide a traffic signal control method within a park to improve the overall logistics efficiency within the park.

[0014] To achieve the above objectives, a first aspect of the present invention provides a traffic signal control method for a park, the method comprising: Receive data packets containing logistics task requirements for vehicles; Receive real-time traffic status data packets from traffic lights; An algorithmic model is adopted to minimize the priority-weighted total delay of all vehicles as the optimization objective. Based on the logistics task demand data packet and the real-time traffic status data packet, the traffic signal control parameters for the next decision cycle are solved. Send out traffic signal control parameters for controlling the traffic lights.

[0015] Preferably, the logistics task requirement data package includes: the vehicle's unique identifier, the vehicle's real-time location coordinates, the vehicle's preset path node ID array, the vehicle's task priority quantification value, and the vehicle's category.

[0016] Preferably, the task priority quantification value of the vehicle is calculated by a function related to the vehicle's urgency level characterization value and the cargo value coefficient.

[0017] Preferably, the task priority quantification value of the vehicle is calculated using the following formula: priority_score=α*Turgency+β*Vvalue; Among them, priority_score is the quantification value of task priority; Turgency is a value representing the urgency level of a vehicle. Turgency = (Promised delivery time - Current time) / Standard transit time; Vvalue is a preset commodity value coefficient, and α and β are adjustable weights.

[0018] Preferably, the real-time traffic status data packet includes: real-time traffic flow and the current phase and remaining time of the traffic lights.

[0019] Preferably, the algorithm model is an improved genetic algorithm or a decomposition coordination algorithm based on Lagrange relaxation.

[0020] Preferably, the improved genetic algorithm includes the following steps: S1, Construct a global traffic and logistics fusion situation map based on the logistics task demand data packet and the real-time traffic status data packet; S2, based on the global traffic and logistics fusion situation map, the problem is encoded, and a possible solution is encoded as an individual. The possible solution represents the phase sequence scheme of all relevant intersections in the future time T1. S3, randomly generate a number of individuals as the initial population, and the green light duration of each individual is generated within a preset feasible range; S4, calculate the fitness value for each individual; S5, determine whether the termination condition is met: the number of iterations t≥N or the improvement of the optimal fitness value for 3 consecutive generations is less than the preset value m; if it is met, go to step S8, otherwise go to step S6, where the initial value of t is 0; S6, Perform genetic operation calculations to obtain the next generation population. The genetic operation calculations include selection, crossover, mutation and elite preservation steps. S7, set the iteration count t=t+1, and return to step S4; S8, Decode the optimal individual based on the current population, where the optimal individual is the individual with the highest fitness value in the current population; S9, Generate the traffic signal control parameters based on the optimal individual.

[0021] Preferably, the step after S8 and before S9 further includes: Simulation verification steps: Perform micro-traffic flow simulation verification on the optimal individual obtained by decoding, calculate the total weighted delay corresponding to the optimal individual, and compare it with the total weighted delay of the currently executed signal scheme; If the improvement exceeds the preset threshold M, then proceed to step S9; If the improvement does not exceed the preset threshold M, the current signal scheme remains unchanged, and the process jumps to step S1 to reacquire the logistics task demand data packet and the real-time traffic status data packet for optimization. The preset threshold M is between 5% and 10%.

[0022] Preferably, after step S9, the method further includes: recording the traffic signal control parameters and comparing the micro-traffic flow simulation verification results with the measured traffic conditions to obtain difference data. The difference data is used to periodically calibrate the parameters of the micro-traffic flow simulation model to achieve self-learning and continuous optimization.

[0023] Preferably, step S4, calculating the fitness value of each individual, includes: Individual i is decoded into a signal timing scheme. Based on the global traffic and logistics situation map, a microscopic traffic flow simulation model is used to extrapolate the vehicle flow movement within the next T2 seconds. In the extrapolation, the additional travel time Delay for each vehicle is calculated. i ; Delay based on the additional travel time for each vehicle i And the task priority quantification value priority_score i Calculate the total weighted delay: Total_Weighted_Delay= (Delay i *priority_score i ), where n is the total number of vehicles involved in the calculation; The fitness value (Fitness) is calculated based on the total weighted delay. .

[0024] Preferably, the values ​​of T1 and T2 are both 120 seconds; the value of N is 10; and the preset value m is 1%.

[0025] Preferably, step S6, performing genetic operations to obtain the next generation population, includes: Selection steps: Using the tournament selection method, a preset number of k individuals are randomly selected from the current population. The fitness values ​​of these k individuals are compared. The individual with the highest fitness value is copied as the parent individual, and the k individuals are returned to the original population. The above process is repeated until the number of parent individuals in the parent individual pool reaches a preset size. Crossover step: Randomly pair up the parent individuals in the parent individual pool, and perform an arithmetic crossover operation on each pair of parent individuals to generate two child individuals; all child individuals constitute the crossover child pool; each parent individual in the parent individual pool participates in pairing only once. Mutation step: Non-uniform mutation is performed on the offspring individuals in the crossover offspring pool. The mutation amplitude decreases as the number of iterations increases. Random perturbation is introduced, and the mutated offspring pool is output. Elite retention step: Before the genetic operation begins, select several individuals with the highest fitness values ​​from the current population as elite individuals, and directly copy the elite individuals into the next generation population; The next generation population consists of elite individuals and mutated offspring individuals, and the sum of the number of elite individuals and the number of mutated offspring individuals is equal to the number of individuals in the initial population.

[0026] Based on the traffic signal control method for parks provided in the first aspect of the present invention, the second aspect of the present invention provides a traffic signal control system for parks, including a collaborative control platform, wherein the collaborative control platform is used to execute the traffic signal control method for parks according to the first aspect of the present invention.

[0027] Preferably, it also includes a logistics transportation management system for sending data packets containing vehicle logistics task requirements to the collaborative control platform.

[0028] Preferably, it also includes traffic signal control equipment for sending real-time traffic status data packets of traffic lights to the collaborative control platform.

[0029] Preferably, the traffic signal control equipment includes a traffic signal controller, which is used to receive traffic signal control parameters sent by the collaborative control platform and control the traffic lights at intersections within the park according to the traffic signal control parameters.

[0030] This invention uses an algorithm model to solve traffic signal control parameters based on logistics task demand data packets and real-time traffic status data packets. This minimizes the priority-weighted total delay of all vehicles, breaks down the "information silos" between the logistics and transportation systems within the park, establishes two-way, real-time, and structured data interaction, transforms "vehicles adapting to lights" into "lights and vehicles coordinating," and dynamically adjusts signal control strategies according to quantified logistics task demands, thereby improving the overall logistics efficiency within the park.

[0031] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0032] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a hardware architecture diagram of the traffic signal control method within a park provided in an embodiment of the present invention.

[0033] Figure 2 This is a flowchart of a traffic signal control method for a park provided in an embodiment of the present invention.

[0034] Figure 3 This is a flowchart of the decision-making process of the improved genetic algorithm provided in this embodiment of the invention. Detailed Implementation

[0035] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0036] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0037] Method Implementation Examples See Figures 1-2 The first aspect of this invention provides a traffic signal control method within a park, the method comprising: receiving a logistics task demand data packet from a vehicle; receiving a real-time traffic status data packet from a traffic light; employing an algorithm model to minimize the priority-weighted total delay of all vehicles as the optimization objective, and solving for traffic signal control parameters within a future decision cycle based on the logistics task demand data packet and the real-time traffic status data packet; and sending the traffic signal control parameters for controlling the traffic light.

[0038] This invention uses an algorithm model to solve traffic signal control parameters based on logistics task demand data packets and real-time traffic status data packets. This minimizes the priority-weighted total delay of all vehicles, breaks down the "information silos" between the logistics and transportation systems within the park, establishes two-way, real-time, and structured data interaction, transforms "vehicles adapting to lights" into "lights and vehicles coordinating," and dynamically adjusts signal control strategies according to quantified logistics task demands, thereby improving the overall logistics efficiency within the park.

[0039] In one specific embodiment, the park can be an industrial park or a warehousing and logistics park. A Collaborative Control Platform (CCP) can be deployed within the park, preferably on an edge computing server, such as the Huawei Atlas 800 intelligent edge server. Alternatively, the CCP can adopt a cloud-edge two-tier architecture, where edge CCPs are deployed at intersections in each zone, responsible for millisecond-level emergency response based on simple rules, such as absolute priority for fire trucks; and cloud CCPs are deployed on the park's central cloud server, responsible for minute-level global policy optimization based on more complex models (such as deep learning). The cloud CCP can be deployed internally within the park, such as in the park's data center, or externally, such as in a public cloud or hosting facility.

[0040] The collaborative control platform receives vehicle logistics task requirement data packets, which may specifically include a vehicle unique identifier (Vehicle_id), the vehicle's real-time coordinates (position), a vehicle path node ID array (planned_route), a vehicle task priority quantification value (Task_priority), and the vehicle category (vehicle). type .

[0041] In one specific embodiment, the task priority quantification value of the vehicle, Task_priority, adopts a three-level quantification system: emergency tasks = 1.0 (such as fire fighting, production line emergency material stoppage risk), high priority tasks = 0.7 (such as fresh food delivery, VIP customer orders), and ordinary tasks = 0.3 (routine replenishment).

[0042] In one specific embodiment, logistics vehicles are equipped with UWB (Ultra-Wideband) positioning tags. These tags report the vehicle's unique identifier and real-time location (accuracy ±0.3m) to the Transportation Management System (TMS) via the campus positioning base station network. The TMS then reports these signals to the collaborative control platform. Additionally, the TMS stores an array of path node IDs for each vehicle and a quantified task priority value. These signals, along with the vehicle's unique identifier and real-time location, are also sent to the collaborative control platform.

[0043] In one specific embodiment, the vehicle's task priority quantification value is calculated using a function related to the vehicle's urgency level and the cargo value coefficient, for example, using the following formula: priority_score=α*Turgency+β*Vvalue; Among them, priority_score is the quantification value of task priority; Turgency is a value representing the urgency level of a vehicle. Turgency = (promised delivery time - current time) / standard transit time (the smaller the value, the more urgent the situation). Vvalue is the preset value coefficient of goods (preset by TMS according to order type, such as urgent materials = 1.0, ordinary materials = 0.3), α and β are adjustable weights, with the default α = 0.7 and β = 0.3.

[0044] By using the above methods, subjective "priority" can be transformed into objective and calculable technical parameters, thereby enabling better vehicle logistics management.

[0045] To calculate the required traffic signal control parameters, the collaborative control platform receives not only logistics task demand data packets from the Transportation Management System (TMS), but also real-time traffic status data packets, specifically including real-time traffic flow, the current phase of traffic lights, and the remaining time of the current phase. Real-time traffic flow is detected by multi-target radar detectors deployed on the roads and sent to the traffic signal controller, which then uploads the real-time traffic flow signal to the collaborative control platform. Simultaneously, the traffic signal controller also uploads the current phase of traffic lights and the remaining time of the current phase to the collaborative control platform. To ensure low-latency and highly reliable data transmission, the network transmission between the traffic signal controller and the collaborative control platform uses a hybrid network of industrial-grade wireless access points (APs) and gigabit industrial Ethernet.

[0046] After receiving logistics task demand data packets and real-time traffic status data packets, the collaborative control platform calculates traffic signal control parameters using a preset algorithm model. The preset algorithm model can be an improved genetic algorithm or a decomposition and coordination algorithm based on Lagrange relaxation.

[0047] In a preferred embodiment, the present invention uses an improved genetic algorithm to calculate traffic signal control parameters. Its core objective is to find a set of signal control parameters within a future decision period T, for example, T = 120 seconds, that minimizes the priority-weighted total delay for all vehicles.

[0048] Specifically, such as Figure 3 As shown, The improved genetic algorithm includes the following steps: S1. Construct a global traffic and logistics fusion situation map based on the logistics task demand data package and the real-time traffic status data package; wherein, the global traffic and logistics fusion situation map uses a Geographic Information System (GIS) as the base map, and overlays icons of all vehicles (with priority color codes: red = emergency, yellow = high, green = normal) and real-time trajectories, which is equivalent to a snapshot of the current road conditions, telling the algorithm where the vehicles are now and what color the lights are.

[0049] S2, based on the global traffic and logistics fusion situation map, problem encoding is performed, encoding a possible solution as an individual. This possible solution represents the phase sequence scheme of all relevant intersections within the next time period T1. For example, controlling 3 intersections, taking 2 key phases at each intersection, with a decision cycle of 2 signal cycles (assuming a cycle of 60 seconds), then the individual encoding is: a real number vector of length 3*2*2=12, for example, individual encoding = [18, 12, 20, 10, 15, 15, 18, 12, 20, 10, 15,

[15] , where the first and second bits represent the green light duration (in seconds) for phase 1 of intersection 1 in the first signal cycle (18 seconds for phase 1 and 12 seconds for phase 2); the third and fourth bits represent the green light duration (20 seconds for phase 1 and 10 seconds for phase 2) for intersection 1 in the second signal cycle; the fifth to twelfth bits represent the phase sequence of the two key phases of intersection 2 and intersection 3 in the next two cycles, similar to intersection 1; phase 1 and phase 2 represent two traffic flows in different directions at the same intersection within one signal cycle, and are mutually exclusive in time (they will not have green lights at the same time). Within one signal cycle, each key phase is executed in a predetermined order, and only one phase can have the right of way at any given time. Each bit in the problem code represents the green light duration (in seconds) of the corresponding key phase at the corresponding intersection within the corresponding cycle. The sum of the green light durations for each phase, plus the necessary yellow and all-red clearing times, equals the total duration of the cycle. For example, for the first signal cycle of intersection 1, the first bit in the code is 18 seconds, representing the green light duration for phase 1; the second bit is 12 seconds, representing the green light duration for phase 2. Within this cycle, phase 1 executes for 18 seconds first, followed by a 3-second yellow / all-red time, and then phase 2 executes for 12 seconds. The remaining time is allocated to other fixed phases or used as a buffer.

[0050] S3, randomly generate a number of individuals (e.g., 100) as the initial population, and generate the green light duration for each individual within a preset feasible range (e.g., within 60 seconds).

[0051] S4, calculate the fitness value for each individual; Specifically, by decoding individual i into a signal timing scheme, and based on a global traffic and logistics situation map, a microscopic traffic flow simulation model (such as the Cellular Transport Model, CTM) can be used to extrapolate the vehicle flow movement within the next T2 seconds. In the extrapolation, the additional travel time Delay for each vehicle is calculated. i ; Delay based on the additional travel time for each vehicle i And the task priority quantification value priority_score i Calculate the total weighted delay: Total_Weighted_Delay= (Delay i priority_score i ), where n is the total number of vehicles involved in the calculation; delays are not treated equally, and delays of vehicles with high priority_scores are given a higher cost by the algorithm and are therefore eliminated first.

[0052] The fitness value (Fitness) is calculated based on the total weighted delay. .

[0053] S5, determine whether the termination condition is met: the number of iterations t≥N or the improvement of the optimal fitness value for 3 consecutive generations is less than the preset value m; if it is met, go to step S8, otherwise go to step S6, where the initial value of t is 0, N can be 10 for example, and m can be 1 for example.

[0054] S6, Perform genetic operations to obtain the next generation population. These genetic operations include selection, crossover, mutation, and elite retention steps. In one specific embodiment, the genetic operations include the following steps: Selection steps: Using a tournament selection method, randomly select a preset number of k individuals (e.g., 5 individuals) from the current population. Compare the fitness values ​​of these k individuals, and duplicate the individual with the highest fitness value as the parent individual. Then, return the k individuals to the original population. Repeat the above process until the number of parent individuals in the parent individual pool reaches a preset size (e.g., if the initial population includes 100 individuals and the elite retention number is 2, then the preset size is 98). Crossover step: Randomly pair up the parent individuals in the parent individual pool, and perform an arithmetic crossover operation on each pair of parent individuals to generate two child individuals; all child individuals constitute the crossover child pool; each parent individual in the parent individual pool participates in pairing only once. Mutation step: Non-uniform mutation is performed on the offspring individuals in the crossover offspring pool. The mutation amplitude decreases as the number of iterations increases. Random perturbation is introduced, and the mutated offspring pool is output. Elite preservation step: Before the genetic operation begins, select several individuals (e.g., 2) with the highest fitness values ​​from the current population as elite individuals, and directly copy the elite individuals into the next generation population; The next generation population consists of elite individuals and mutated offspring individuals, and the sum of the number of elite individuals and the number of mutated offspring individuals is equal to the number of individuals in the initial population.

[0055] S7, set the iteration count t=t+1, and return to step S4; S8. Decode the optimal individual based on the current population. The optimal individual is the one with the highest fitness value in the current population. For example, in the current population, traverse all individuals, compare their fitness values, and find the individual with the highest fitness value. For example: Individual A: fitness value = 0.85, Individual B: fitness value = 0.92 (maximum value), Individual C: fitness value = 0.78, ... Among them, Individual B is the optimal individual.

[0056] During decoding, according to the preset encoding rules, each gene value in the real number vector of the best individual is mapped sequentially to the green light duration (seconds) of the corresponding intersection, the corresponding signal cycle, and the corresponding key phase.

[0057] For example, for an individual code = [18, 12, 20, 10, 15, 15, 18, 12, 20, 10, 15, 15], the decoded signal timing scheme is as follows: Intersection 1: In the first cycle, phase 1 (e.g., east-west direction) has a green light of 18 seconds, and phase 2 (north-south direction) has a green light of 12 seconds; in the second cycle, phase 1 has a green light of 20 seconds, and phase 2 has a green light of 10 seconds. Intersection 2: In the first cycle, phase 1 has a 15-second green light and phase 2 has a 15-second green light; in the second cycle, phase 1 has an 18-second green light and phase 2 has a 12-second green light. Intersection 3: In the first cycle, phase 1 has a 20-second green light and phase 2 has a 10-second green light; in the second cycle, phase 1 has a 15-second green light and phase 2 has a 15-second green light.

[0058] S9, Generate the traffic signal control parameters based on the optimal individual.

[0059] Specifically, after decoding, a complete signal timing scheme is obtained, including the green light start time and duration for each phase at each intersection within each cycle. This scheme is then converted into an executable instruction format for the traffic signal controller (such as an OPC UA protocol instruction packet), which generates the traffic signal control parameters, including: signal controller ID, phase number, green light start time / duration, cycle length, etc.

[0060] In another preferred embodiment, to avoid the optimal individual output by the genetic algorithm performing poorly in actual execution due to simulation model deviations or sudden changes in traffic flow, the collaborative control platform first performs micro-traffic flow simulation verification on the optimal individual before outputting signal control parameters. Specifically, the total weighted delay corresponding to the optimal individual is calculated and compared with the total weighted delay of the currently executing signal scheme. If the improvement exceeds a preset threshold M (e.g., 5%-10%), signal control parameters are generated based on the optimal individual and issued for execution; if the improvement does not exceed the threshold M, the current signal scheme remains unchanged, and the logistics task demand data packet (reflecting the latest vehicle location, task priority, and route planning information at the current moment) and real-time traffic status data packet are re-collected to trigger the next round of optimization. This mechanism effectively avoids frequent invalid switching of signal schemes, ensuring the stability and robustness of traffic control.

[0061] To continuously improve the accuracy of the simulation model and the overall optimization effect of the system, in another preferred embodiment, after step S9, the method further includes: recording the traffic signal control parameters and comparing the micro-traffic flow simulation verification results with the measured traffic conditions to obtain difference data. This difference data is used to periodically calibrate the micro-traffic flow simulation model parameters, enabling self-learning and continuous optimization.

[0062] Specifically, after the Collaborative Control Platform (CCP) sends the signal control parameters to the traffic signal controller for execution, the CCP does not stop working but continues to collect traffic status data during the actual execution process, including: (1) Data collection During the execution of the signal scheme (e.g., the next 120 seconds), the roadside multi-target radar detectors continuously collect the following measured data: actual traffic flow in each direction (number of vehicles / minute); actual passage time and delay time of each vehicle; actual queue length of each phase; and actual execution status of each intersection signal phase.

[0063] (2) Comparison between prediction and actual measurement CCP compares the predicted values ​​obtained from the micro-traffic flow simulation model before execution with the measured values ​​collected after execution, calculating the difference data. The difference data can be expressed as an absolute difference or a relative percentage, for example: Difference = (Measured value - Predicted value) / Predicted value × 100% (3) Model parameter calibration CCP records the aforementioned discrepancies and periodically (e.g., hourly, daily, or weekly) uses the accumulated discrepancies to calibrate the parameters of the micro-traffic flow simulation model. Calibrable parameters include, but are not limited to: vehicle following model parameters (e.g., reaction time, safe distance); saturation flow rate (the maximum number of vehicles that can pass per unit time); start-up loss time (the time from the green light turning on to the first vehicle passing); and traffic demand parameters (e.g., OD matrix, traffic flow distribution).

[0064] (4) Self-learning and continuous optimization After multiple calibrations, the prediction accuracy of the simulation model gradually improved. This enabled the genetic algorithm to more accurately evaluate the actual effect of each candidate signal scheme when calculating the fitness value, thereby outputting better signal control parameters. As the system's running time increased, the model became increasingly closer to the real traffic environment, and the optimization effect continued to improve.

[0065] Through the above mechanism, the embodiments of the present invention realize a complete closed loop of "perception → decision-making → execution → verification → learning → optimization", enabling the system to have the ability to self-evolve and continuously improve, and overcoming the shortcomings of traditional open-loop control systems that cannot adapt to environmental changes.

[0066] The above-mentioned traffic signal control method for parks provided by the present invention will be described below with reference to a specific embodiment.

[0067] Example 1: Urgent Material Transportation on the Production Line This embodiment uses an actual production scenario in a logistics park (pilot project in the first quarter of 2024) as an example to illustrate the implementation process of the present invention in detail.

[0068] (1) Application Background Within a logistics park, a production line faced the risk of shutdown due to a material shortage, with the downtime expected to exceed two hours, resulting in significant production losses. Upon detecting the shortage, the park's Transportation Management System (TMS) immediately triggered an emergency transport task.

[0069] (2) Hardware deployment and data acquisition Within the park, the Collaborative Control Platform (CCP) is deployed on Huawei Atlas 800 intelligent edge servers, connecting to traffic signal controllers at various intersections via industrial Ethernet and to the TMS via a RESTful API. All logistics vehicles (including AGVs, forklifts, and transport vehicles) are equipped with UWB (Ultra-Wideband) positioning tags, achieving real-time positioning accuracy of ±0.3m through the park's positioning base station network.

[0070] When a material shortage occurs on the production line, the TMS generates an emergency transport task with a task priority quantization value (task_priority) set to 1.0 (highest priority, corresponding to the "emergency task" level). The TMS dispatches an AGV from the warehouse to perform this emergency transport task. The AGV's UWB positioning tag reports its location information in real time (accuracy ±0.3m). The TMS sends a logistics task requirement data packet containing the vehicle's unique identifier (Vehicle_id), real-time coordinates (position), a path node ID array (planned_route), the task priority quantization value (Task_priority = 1.0), and the vehicle type (vehicle_type) to the CCP via a RESTful API.

[0071] (3) Collaborative decision-making CCP's decision engine is triggered. CCP first integrates logistics task demand data packets from TMS and real-time traffic status data packets from traffic signal controllers to construct a global traffic and logistics fusion situation map. Within this situation map, CCP identifies the following key information: The AGV is currently located at the warehouse exit and is traveling along its preset route (planned_route). It will arrive at intersection 3 in 75 seconds.

[0072] The current signal phase configuration at intersection 3 is as follows: the green light for the AGV's direction of travel will end in 60 seconds (i.e., the remaining green light time is 60 seconds), after which it will switch to the green light for the intersecting direction.

[0073] This means that if the current signal timing scheme is followed, the AGV will encounter a red light when it arrives at intersection 3 and will have to wait for at least one signal cycle to pass through, causing unnecessary delays.

[0074] (4) Algorithm solution CCP's improved genetic algorithm (IGA) decision engine was triggered. The algorithm takes the next decision cycle (T=120 seconds) as the optimization horizon and minimizes the priority-weighted total delay of all vehicles as the optimization objective, and iteratively solves the problem.

[0075] After five iterations, the algorithm converged and found the optimal solution. The specific content of this optimal solution includes: Extend the green light phase in the direction the AGV is traveling at intersection No. 3 by 15 seconds, so that the AGV can arrive just in time to catch the green light without having to slow down and wait.

[0076] At the same time, the downstream No. 4 intersection is optimized by adjusting its signal phase timing to form a continuous green wave with No. 3 intersection, so that the AGV can smoothly continue to pass through No. 4 intersection after passing through No. 3 intersection.

[0077] Algorithm calculations show that this scheme causes the least increase in weighted delays in other directions (such as intersecting social vehicles), ensuring the passage of high-priority tasks while keeping the impact on other traffic participants within an acceptable range.

[0078] (5) Hardware execution The CCP transforms the obtained signal optimization strategy into specific "signal optimization instruction packages," which are then sent in real-time to the traffic signal controllers at intersections 3 and 4 via the OPC UA protocol. Each instruction package contains the signal controller ID, phase number, green light extension time (15 seconds), and coordination timing information. Upon receiving the instruction, the signal controller immediately adjusts its internal timer and executes the new timing scheme. The system latency for the entire decision-making and instruction delivery process is controlled within 100ms, meeting the real-time requirements of industrial scenarios.

[0079] (6) Implementation results Under the aforementioned collaborative optimization scheme, when the AGV travels at a constant speed to intersection 3, the traffic light at the intersection happens to be green. The AGV does not need to slow down or wait and passes through intersection 3 smoothly. Then it continues to pass through intersection 4 smoothly without stopping or waiting throughout the entire process.

[0080] Compared with the traditional method (where vehicles can only passively adapt to fixed signal timing), the collaborative optimization scheme of the present invention can shorten the completion time of this emergency task from 4.2 minutes to 2.5 minutes, a reduction of about 40.5%, significantly reducing the risk of production line downtime.

[0081] In addition, CCP recorded all data during this collaborative process (including AGV trajectories, signal adjustment records, actual delays, etc.) for subsequent model self-learning optimization. This data will be compared with simulation predictions, and the discrepancies will be used to periodically calibrate the parameters of the micro-traffic flow simulation model, enabling the system to continuously improve during operation.

[0082] Example 2: Effect Verification Based on Simulation This embodiment uses a digital twin simulation platform to quantitatively verify the technical effects of the present invention in a typical rectangular park road network.

[0083] (1) Simulation environment settings The simulated road network is a typical rectangular park road network, containing a road layout of 4 north-south and 3 east-west roads, with a total of 12 signal-controlled intersections. The road network includes two types of traffic participants: private vehicles and logistics vehicles.

[0084] (2) Comparison of scene settings To verify the technical effects of the embodiments of the present invention, the following two comparison scenarios are set up:

[0085] (3) Test conditions Simulation duration: Simulates 1 hour of park operation Vehicle composition: 300 private vehicles + 50 logistics vehicles Emergency task setting: Randomly generate 5 emergency task vehicles with priority_score>0.8 from 50 logistics vehicles (corresponding to high priority transportation tasks, such as urgent materials for the production line, emergency supplies, etc.). Decision cycle: T=120 seconds, meaning that CCP triggers a collaborative optimization every 120 seconds.

[0086] (4) Comparison of simulation results

[0087] (5) Results Analysis The above experimental data show that: First, high-priority tasks receive a deterministic guarantee of timeliness. Under traditional methods, the average transit time for high-priority logistics vehicles is 425 seconds (approximately 7 minutes); under the solution of this invention, this time is reduced to 253 seconds (approximately 4.2 minutes), a reduction of 40.5%. This verifies the effectiveness of the "two-way value transfer mechanism" of this invention—by incorporating the quantified value of task priority into the optimization objective, the transit needs of high-priority vehicles are prioritized.

[0088] Second, the overall system efficiency is significantly improved. The average delay of all logistics vehicles has been reduced from 118 seconds to 95 seconds, a reduction of 19.5%. This shows that the present invention not only serves high-priority tasks but also brings about an improvement in overall logistics efficiency, overcoming the defect in the prior art where "local optimization may lead to global inefficiency".

[0089] Third, the impact on other vehicles is within a controllable range. The average delay for other vehicles only increased from 65 seconds to 68 seconds, an increase of +4.6%, far less than the improvement in high-priority tasks and overall logistics efficiency. This shows that while prioritizing logistics tasks, the invention did not significantly sacrifice the right of passage for other vehicles, achieving the goal of "precise coordination and overall optimization."

[0090] Fourth, it meets the real-time requirements of industrial scenarios. The decision-making cycle calculation time of this invention is less than 3 seconds, far lower than the decision-making cycle itself (120 seconds), fully meeting the second-level calculation requirements of real-time collaboration in industrial parks. This verifies that this invention, through targeted improvements to the genetic algorithm (limiting the number of iterations, using a fast simulation model, etc.), makes the theoretical "collaboration" technically "feasible" in engineering.

[0091] Fifth, the algorithm's optimization objective was directly achieved. The total weighted delay (Total_Weighted_Delay) decreased from the baseline value of 100 to 71.2, a reduction of 28.8%. Since the total weighted delay is the core optimization objective of this invention (i.e., minimizing the priority-weighted total delay of all vehicles), this result directly proves that the optimization algorithm of this invention is effective.

[0092] In summary, the embodiments of the present invention significantly reduce the transit time of high-priority logistics tasks (by 40.5%) and overall logistics delays (by 19.5%) while having almost no impact on the passage of social vehicles (delays increase by only 4.6%). Simulation experimental data fully verify the technical effect of the "precise coordination and overall optimization" of the present invention, proving the effectiveness and superiority of the traffic signal control method for parks provided by the embodiments of the present invention.

[0093] A second aspect of the present invention provides a traffic signal control system for a park, including a collaborative control platform, which is used to execute the traffic signal control method for a park according to a first aspect of the present invention.

[0094] Furthermore, it also includes a logistics transportation management system for sending data packets containing vehicle logistics task requirements to the collaborative control platform.

[0095] Furthermore, it also includes traffic signal control equipment for sending real-time traffic status data packets of traffic lights to the collaborative control platform.

[0096] The traffic signal control equipment includes a traffic signal controller, which is used to receive traffic signal control parameters sent by the collaborative control platform and control the traffic lights at intersections within the park according to the traffic signal control parameters.

[0097] The functions and effects of the traffic signal control system within the park provided in the second aspect of the present invention are the same as those of the traffic signal control method within the park provided in the first aspect of the present invention, and will not be repeated here.

[0098] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A traffic signal control method within a park, characterized in that, The method includes: Receive data packets containing logistics task requirements for vehicles; Receive real-time traffic status data packets from traffic lights; An algorithmic model is adopted to minimize the priority-weighted total delay of all vehicles as the optimization objective. Based on the logistics task demand data packet and the real-time traffic status data packet, the traffic signal control parameters for the next decision cycle are solved. Send out traffic signal control parameters for controlling the traffic lights.

2. The traffic signal control method within the park according to claim 1, characterized in that: The logistics task requirement data package includes: the vehicle's unique identifier, the vehicle's real-time location coordinates, the vehicle's preset path node ID array, the vehicle's task priority quantification value, and the vehicle's category.

3. The traffic signal control method within the park according to claim 1, characterized in that: The vehicle's task priority quantification value is calculated using a function related to the vehicle's urgency level and cargo value coefficient.

4. The traffic signal control method within the park according to claim 3, characterized in that: The task priority quantification value of the vehicle is calculated using the following formula: priority_score=α*Turgency+β*Vvalue; Among them, priority_score is the quantification value of task priority; Turgency is a value representing the urgency level of a vehicle. Turgency = (Promised delivery time - Current time) / Standard transit time; Vvalue is a preset commodity value coefficient, and α and β are adjustable weights.

5. The traffic signal control method within the park according to claim 1, characterized in that: The real-time traffic status data packet includes: real-time traffic flow and the current phase and remaining time of the traffic lights.

6. The traffic signal control method within the park according to claim 1, characterized in that: The algorithm model is an improved genetic algorithm or a decomposition coordination algorithm based on Lagrange relaxation.

7. The traffic signal control method within the park according to claim 6, characterized in that: The improved genetic algorithm includes the following steps: S1, Construct a global traffic and logistics fusion situation map based on the logistics task demand data packet and the real-time traffic status data packet; S2, based on the global traffic and logistics fusion situation map, the problem is encoded, and a possible solution is encoded as an individual. The possible solution represents the phase sequence scheme of all relevant intersections in the future time T1. S3, randomly generate a number of individuals as the initial population, and the green light duration of each individual is generated within a preset feasible range; S4, calculate the fitness value for each individual; S5, determine whether the termination condition is met: the number of iterations t≥N or the improvement of the optimal fitness value for 3 consecutive generations is less than the preset value m; if it is met, go to step S8, otherwise go to step S6, where the initial value of t is 0; S6, Perform genetic operation calculations to obtain the next generation population. The genetic operation calculations include selection, crossover, mutation and elite preservation steps. S7, set the iteration count t=t+1, and return to step S4; S8, Decode the optimal individual based on the current population, where the optimal individual is the individual with the highest fitness value in the current population; S9, Generate the traffic signal control parameters based on the optimal individual.

8. The traffic signal control method within the park according to claim 7, characterized in that: The process after step S8 and before step S9 also includes: Simulation verification steps: Perform micro-traffic flow simulation verification on the optimal individual obtained by decoding, calculate the total weighted delay corresponding to the optimal individual, and compare it with the total weighted delay of the currently executed signal scheme; If the improvement exceeds the preset threshold M, then proceed to step S9; If the improvement does not exceed the preset threshold M, the current signal scheme remains unchanged, and the process jumps to step S1 to reacquire the logistics task demand data packet and the real-time traffic status data packet for optimization. The preset threshold M is between 5% and 10%.

9. The traffic signal control method within the park according to claim 7, characterized in that: After step S9, the method further includes: recording the traffic signal control parameters and comparing the micro-traffic flow simulation verification results with the measured traffic conditions to obtain difference data. The difference data is used to periodically calibrate the parameters of the micro-traffic flow simulation model to achieve self-learning and continuous optimization.

10. The traffic signal control method within the park according to claim 7, characterized in that: Step S4, calculating the fitness value for each individual, includes: Individual i is decoded into a signal timing scheme. Based on the global traffic and logistics situation map, a microscopic traffic flow simulation model is used to extrapolate the vehicle flow movement within the next T2 seconds. In the extrapolation, the additional travel time Delay for each vehicle is calculated. i ; Delay based on the additional travel time for each vehicle i And the task priority quantification value priority_score i Calculate the total weighted delay: Total_Weighted_Delay= (Delay i *priority_score i ), where n is the total number of vehicles involved in the calculation; The fitness value (Fitness) is calculated based on the total weighted delay. .

11. The traffic signal control method within the park according to claim 10, characterized in that: The values ​​of T1 and T2 are both 120 seconds; the value of N is 10; and the preset value m is 1%.

12. The traffic signal control method within the park according to claim 7, characterized in that: Step S6 involves performing genetic operations to obtain the next generation population, including: Selection steps: Using the tournament selection method, a preset number of k individuals are randomly selected from the current population. The fitness values ​​of these k individuals are compared. The individual with the highest fitness value is copied as the parent individual, and the k individuals are returned to the original population. The above process is repeated until the number of parent individuals in the parent individual pool reaches a preset size. Crossover step: Randomly pair up the parent individuals in the parent individual pool, and perform an arithmetic crossover operation on each pair of parent individuals to generate two child individuals; all child individuals constitute the crossover child pool; each parent individual in the parent individual pool participates in pairing only once. Mutation step: Non-uniform mutation is performed on the offspring individuals in the crossover offspring pool. The mutation amplitude decreases as the number of iterations increases. Random perturbation is introduced, and the mutated offspring pool is output. Elite retention step: Before the genetic operation begins, select several individuals with the highest fitness values ​​from the current population as elite individuals, and directly copy the elite individuals into the next generation population; The next generation population consists of elite individuals and mutated offspring individuals, and the sum of the number of elite individuals and the number of mutated offspring individuals is equal to the number of individuals in the initial population.

13. A traffic signal control system for a park, characterized in that: It includes a collaborative control platform, which is used to execute the traffic signal control method within the park as described in any one of claims 1-12.

14. The traffic signal control system within the park according to claim 13, characterized in that: It also includes a logistics transportation management system for sending vehicle logistics task demand data packets to the collaborative control platform.

15. The traffic signal control system within the park according to claim 13, characterized in that: It also includes traffic signal control equipment for sending real-time traffic status data packets of traffic lights to the collaborative control platform.

16. The traffic signal control system within the park according to claim 15, characterized in that: The traffic signal control equipment includes a traffic signal controller, which is used to receive traffic signal control parameters sent by the collaborative control platform and control the traffic lights at intersections within the park according to the traffic signal control parameters.

Images