Urban road sectional type construction method capable of reducing traffic influence

Abstract

The invention relates to the technical field of road construction, and discloses an urban road sectional construction method capable of reducing traffic influence, comprising the following steps: step 1, acquiring real-time traffic data representing the current traffic condition and preset construction process information; 2, calculating through an optimized objective function, and dynamically generating a construction form scheme aiming at balancing the traffic passing efficiency and the construction work efficiency based on the real-time traffic data and the construction process information; and step 3, according to the construction form scheme, generating an adjustment instruction for the physical boundary of the construction area and executing adjustment. By constructing a traffic digital twinborn model, dynamic self-adaption of a construction boundary to a real-time traffic condition is realized in a closed-loop management and control mode. The construction boundary can be automatically contracted or expanded according to the traffic flow, the construction requirement is guaranteed, meanwhile, traffic jam caused by a traditional fixed fence is effectively relieved, and the passing efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of road construction, in particular to a segmented construction method for urban roads that reduces traffic impact. Background Art

[0002] In recent years, with the acceleration of global urbanization, the need for the maintenance, renovation, and upgrading of urban infrastructure, particularly road networks, has become increasingly urgent. Urban road construction, an essential component for ensuring the healthy operation of road networks and improving traffic capacity, has seen a significant increase in both frequency and scale. However, road construction inevitably consumes limited road resources, significantly impacting normal urban traffic flow. Construction areas often become sources of traffic congestion, leading to increasingly prominent problems such as traffic delays, energy consumption, and environmental pollution.

[0003] Currently, to ensure construction safety and the integrity of the work area during urban road construction, a commonly used technical approach is to use physical barriers, such as water barriers, barriers, and guardrails, to physically isolate the construction area from the traffic lanes. This approach involves defining and deploying a construction boundary at the outset, based on the construction plan and relevant safety regulations. Once established, this static, fixed construction boundary typically remains unchanged throughout the construction cycle or over a longer construction phase.

[0004] However, this traditional construction management approach suffers from inherent technical flaws. The core issue lies in a profound and irreconcilable contradiction between the "static" nature of physical boundaries and the "dynamic" nature of urban traffic flows. Urban traffic exhibits significant tidal characteristics, with traffic volumes differing several times between weekday morning and evening rush hours and nighttime off-peak periods. Fixed construction boundaries are set based on a trade-off or worst-case scenario, a one-size-fits-all approach that results in suboptimal performance most of the time. During off-peak hours, the established construction boundaries may exceed actual operational requirements, unnecessarily occupying road space that could be freed up for public use. During peak hours, however, these same fixed boundaries can become traffic bottlenecks, severely limiting road capacity, leading to dramatically increased queue lengths, and exacerbating regional traffic congestion.

[0005] Furthermore, the decision-making process of existing technologies lacks the scientific nature of being driven by real-time data. The demarcation of construction boundaries relies more on engineering experience, historical statistics, or static design specifications, and is unable to make dynamic and refined adjustments based on the ever-changing real-world traffic situation at the construction site. At the same time, traditional construction management has also failed to effectively link the dynamic demands of the construction task itself with traffic management. Different construction processes have different requirements for workspace, but static fencing cannot reflect these differences, resulting in inefficient resource utilization. Although fencing can theoretically be adjusted manually, in practice, this process not only consumes a lot of manpower and material resources, and the work process is cumbersome, but the frequent changes in the work surface also bring additional safety risks to on-site construction personnel and traffic participants. Therefore, it is rarely used in actual projects. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the present invention provides a segmented construction method for urban roads that reduces traffic impact and solves the problem of traffic congestion caused by traditional fixed fences.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a segmented construction method for urban roads that reduces traffic impact, comprising the following steps:

[0008] Step 1: Acquire real-time traffic data representing current traffic conditions and preset construction process information;

[0009] Step 2: Calculating through an optimization objective function, based on the real-time traffic data and construction process information, dynamically generating a construction form plan that aims to balance traffic efficiency and construction work benefits;

[0010] Step 3: Generate an adjustment instruction for the physical boundary of the construction area according to the construction form plan and execute the adjustment;

[0011] Step 4: Execute steps 1 to 3 repeatedly in a preset decision cycle to achieve adaptive control of the construction area.

[0012] Preferably, step 1 further includes: constructing and calibrating in real time a traffic digital twin model synchronized with the physical world based on the real-time traffic data.

[0013] Preferably, the dynamic generation of the construction form scheme in step 2 is specifically as follows: in the traffic digital twin model, a plurality of candidate construction form schemes are forward-looking simulated to predict the impact of each candidate scheme on future traffic conditions, and the construction form scheme is selected based on merit.

[0014] Preferably, the optimization objective function at least includes:

[0015] Traffic delay cost item, which is used to characterize the impact of the construction scheme on traffic efficiency; construction benefit item, which is used to characterize the size of the working space provided by the construction scheme;

[0016] And the form switching cost term is used to represent the physical cost of transforming from the previous construction form to the current construction form.

[0017] Preferably, the weight of each cost item in the optimization objective function is dynamically adjusted according to the current construction process information obtained in step 1.

[0018] Preferably, the method further comprises: determining the weight of each cost item in the optimization objective function from a preset process-weight mapping relationship according to the type of the current construction process.

[0019] Preferably, the process of dynamically generating a solution by optimizing the objective function in step 2 is implemented by using a hybrid intelligent algorithm;

[0020] The hybrid intelligent algorithm includes: using a genetic algorithm to perform a global search on the solution space of the construction form scheme; and using a reinforcement learning algorithm to guide the mutation direction of the genetic algorithm according to real-time traffic data.

[0021] Preferably, the adjustment of the physical boundary of the construction area in step 3 is specifically: completing the contraction, expansion or translation of the physical boundary of the construction area by controlling the movement of multiple modular intelligent enclosure units.

[0022] Preferably, the calculation of the morphology switching cost item is based on the physical state information of the modular intelligent enclosure unit itself obtained in real time after the adjustment is performed.

[0023] Preferably, the self-physical status information includes at least: the remaining power and / or health status of the modular intelligent enclosure unit.

[0024] The present invention provides a segmented construction method for urban roads that reduces traffic impacts. It has the following beneficial effects:

[0025] 1. This invention achieves dynamic adaptation of construction zone boundaries to real-world traffic conditions by constructing a high-fidelity digital twin model of traffic and executing a complete closed loop of "perception-decision-action-feedback" within a preset decision cycle. When traffic increases, the system automatically decides to shrink the boundaries, freeing up more space for non-local vehicles; when traffic calms, the boundaries are expanded to ensure smooth construction. This continuous, automated fine-tuning capability effectively alleviates the severe congestion inevitably caused by traditional fixed barriers during peak traffic hours, significantly improving traffic efficiency along construction routes.

[0026] 2. This invention incorporates construction efficiency as a clear, quantifiable optimization objective into the core decision-making algorithm and establishes a context-aware mechanism whose weighting coefficients dynamically adjust with the construction process. This means that rather than simply sacrificing construction to ensure traffic flow, this method scientifically seeks the optimal balance between maximizing construction space and minimizing traffic impact at every decision moment. During critical processes, the system intelligently strives for a larger work surface within the permitted traffic carrying capacity, effectively ensuring the continuity and progress of segmented construction in a complex urban traffic environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0028] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0029] Please see the attached Figure 1 The embodiment of the present invention provides a segmented construction method for urban roads that reduces traffic impact, comprising:

[0030] In one specific embodiment of the present invention, step S1 serves as the initial and foundational stage of the method. Its core objective is to construct a digital replica of the physical world's transportation-construction system, a high-fidelity, predictive transportation digital twin model, that accurately maps and dynamically deduces it. This step, through the deep integration and processing of multi-source heterogeneous information, provides reliable, comprehensive, and forward-looking data support and simulation environment for dynamic optimization decisions in the subsequent step S2.

[0031] In order to achieve the above purpose, the specific technical process of step S1 can be further decomposed into static basic data loading, dynamic real-time data acquisition, data fusion preprocessing, and digital twin modeling and calibration.

[0032] Regarding static basic data loading: This embodiment first loads and parses two key static data sources to build the static geographic and engineering background skeleton of the digital twin environment.

[0033] Specifically, the system first loads and deeply analyzes the construction project's Building Information Model (BIM), commonly known as the BIM. It's important to note that the use of the BIM here goes beyond simply acquiring three-dimensional geometric information about the construction area. Its deeper purpose is to automatically extract structured construction process information and form a process database. In this database, each construction process, such as "roadbed excavation," "asphalt paving," or "line marking," is associated with a set of attributes that characterize its inherent engineering requirements, including but not limited to its requirements for workspace dimensions, such as minimum construction width and minimum continuous work length; its time constraints, such as the required minimum continuous work time or concrete curing rest time; and its pre- and post-process logical dependencies with other processes. The precise acquisition and structuring of this process information is the fundamental prerequisite and technical basis for implementing context-aware decision-making in the subsequent step S2, ensuring that the optimization objectives closely align with the actual engineering needs on site.

[0034] The system then loads and processes geographic information system data. This effort aims to establish a precise local road network topology model covering the construction site and key nodes upstream and downstream. This model not only defines the road's geometric centerline, width, and number of lanes, but also includes detailed information on intersection topology, the precise location of traffic lights, entrance and exit ramps, variable lanes, bus lanes, and other key infrastructure that significantly impacts traffic flow. This road network topology model constitutes the fundamental geographic environment for traffic flow simulation in the digital twin, ensuring consistency between the simulated geometry and the physical world.

[0035] Regarding dynamic real-time data acquisition: On top of the static skeleton, the system continuously collects real-time traffic data through a heterogeneous sensor network that is preferably designed as a three-layer structure, injecting dynamic information reflecting the rapid changes in the physical world into the digital twin model.

[0036] At the micro level, high-precision sensors such as millimeter-wave radar, lidar, or high-definition video surveillance combined with image recognition algorithms can be deployed in sections immediately upstream and downstream of the construction area, as well as at key bottlenecks. These devices can provide detailed lane-level traffic parameters such as instantaneous vehicle speed, headway, real-time queue length, and traffic density at high frequency and low latency. This high-resolution micro data is the core basis for subsequent precise calibration of the digital twin model and verification of the fidelity of simulation results.

[0037] At the meso-level, the method of this embodiment can be connected to the existing infrastructure of urban traffic management departments, such as video surveillance systems and geomagnetic coil detectors, to obtain data such as the directional turning flow at key intersections, the average travel speed and time occupancy rate of major road sections, etc., in order to grasp the traffic operation characteristics of local areas.

[0038] At the macro level, the system preferably interacts with urban traffic information centers or mainstream internet navigation service platforms via an application programming interface (API). This aims to obtain macro-level traffic situation information, such as the traffic congestion index within a larger region and the average travel time between regions (i.e., the travel time between origin and destination). This macro-level data helps to understand the overall impact of construction activities on the regional road network from a broader perspective and provides a more practical reference for setting boundary conditions for the digital twin model.

[0039] Regarding data fusion and preprocessing: Raw data acquired from multi-source, heterogeneous sensor networks has varying formats, varying accuracy, and may contain noise and missing data, making it difficult to directly use for modeling. Therefore, before being incorporated into the digital twin model, it requires a series of preprocessing and fusion operations to form a unified, reliable, and spatiotemporally aligned view of the traffic situation.

[0040] This process specifically includes: data cleaning, which is used to eliminate obvious outliers caused by equipment failure or communication interference; data conversion, which processes different types of data through corresponding algorithms and converts them into unified quantitative traffic parameters; and spatiotemporal alignment, which uses the timestamp and geographic coordinate information of each data source to unify all data into the same spatiotemporal reference.

[0041] Subsequently, this embodiment uses a data fusion algorithm to fuse multiple observations of the same traffic parameter from different sensors into a single, more confident estimate. The fused data can be organized into a traffic situation state vector x t , which represents the key state of the road network at time t, such as x t =[q t ,v t ,ρ t ,…], where q t ,v t ,ρ t Represent the fused flow, velocity, and density respectively. This vector is the direct input for subsequent model calibration and prediction.

[0042] Regarding digital twin modeling, calibration, and prediction: This is the final output of step S1, which integrates all the aforementioned information into an interactive and deducible dynamic model.

[0043] In a preferred embodiment of the present invention, the digital twin model is developed and constructed based on a microscopic traffic simulation platform. After loading static geographic and engineering information, the system has a basic virtual scene that corresponds to the physical world in terms of geometry and engineering properties.

[0044] In order to ensure that the virtual scene can truly reflect the dynamic operating laws of the physical world, real-time calibration is an indispensable and ongoing key link. Its basic principle is to continuously compare the observation values ​​of the real world with the output values ​​of the simulated world, and dynamically and recursively adjust the internal parameters of the model to eliminate the differences between the two. Preferably, this process can be implemented using Kalman filtering or its nonlinear variants, such as extended Kalman filtering. Specifically, the state of the simulation model, such as the average speed of a certain road section in the model, can be used as the system state variable, and the real-time observation value of the sensor network can be used as the measurement value. Through the two steps of prediction and update of Kalman filtering, the system can optimally estimate and continuously correct the parameters within the simulation model that are usually difficult to observe directly, such as the reaction time in the driver following model and the degree of aggressiveness in the lane changing behavior model, so that the macro and micro behavioral characteristics of the simulation model continue to approach the real traffic flow.

[0045] Furthermore, in order to support the forward-looking optimization decision in step S2, this embodiment gives the digital twin model a key predictive capability. This is not a simple linear extrapolation, but is achieved by embedding a pre-trained time series prediction module in the model. Preferably, the module can be constructed using a recurrent neural network such as a long short-term memory network that can effectively capture long-term dependencies in time series data. The long short-term memory network uses a large amount of historical traffic data for offline training to learn the traffic flow evolution pattern of a specific road section at different times, such as weekdays, holidays, and different weather conditions. During actual operation, it will fuse the processed traffic situation state vector x for the most recent period of time. t As input, the output is a prediction of the traffic flow at the simulation model boundary within one or more future decision cycles. This predicted flow serves as the dynamic boundary input of the simulation model, allowing the digital twin system to scientifically deduce the possible evolution of traffic conditions over a period of time under the current decision.

[0046] At this point, step S1 completes the entire process. The final output is no longer a static map or isolated set of data, but a dynamic, high-fidelity digital twin system that synchronizes with the physical world in real time and can scientifically predict future traffic conditions. This system provides a solid foundation for the effectiveness and scientific nature of all subsequent optimization decisions, ensuring that the method of the present invention can truly respond intelligently based on a deep understanding of current and future traffic conditions.

[0047] In this embodiment, step S2 serves as the core decision-making hub of the method. Its fundamental purpose is to dynamically generate a construction plan that achieves the optimal balance between traffic efficiency and construction benefits through a systematic optimization process within the high-fidelity digital twin environment constructed by step S1. In other words, this step embodies the intelligent nature of the method, transforming a complex, multi-objective engineering trade-off problem into a computable, iteratively solvable mathematical optimization problem.

[0048] The implementation of this step mainly relies on two complementary core technical parts: one is to build an optimization objective function that can perceive the construction context and dynamically adjust; the other is to solve the optimization problem through an efficient hybrid intelligent algorithm.

[0049] Regarding the context-aware optimization objective function: To scientifically and quantitatively evaluate the pros and cons of any candidate construction configuration, this embodiment first defines a comprehensive optimization objective function. This function aims to unify the various impacts of a configuration into a single, comparable evaluation dimension.

[0050] Preferably, the optimization objective function J(S t ) In a decision cycle t, it can be expressed as a given construction form scheme S t The comprehensive cost assessment is as follows:

[0051] J(S t )=w d ·T delay (S t )-w a ·A work (S t )+w c ·C switch (S t-1 ,S t );

[0052] It should be noted that this function is designed to find the cost J(S t ). Each parameter and function term here has a clear physical meaning and calculation basis, which is explained in detail as follows:

[0053] Construction form plan S t : This parameter represents a specific candidate construction area layout scheme. Technically, it is a set of half-plane coordinates that contain the locations of all modular intelligent enclosure units under this scheme.

[0054] Traffic delay cost item T delay (S t ): This item is used to characterize the selected scheme St The negative impact on social traffic operation is a quantitative prediction result. Specifically, this value is not estimated based on static rules or historical experience, but is obtained by conducting a complete forward-looking simulation in the digital twin model constructed in step S1. t The boundary information is loaded into the simulation environment, and the future traffic flow predicted in step S1 is used as input to run a simulation for one decision cycle. After the simulation is completed, the system calculates the total delay time accumulated by all affected vehicles during that cycle, for example, in units of "total vehicle-hours," and uses this as the value of the cost item. This ensures that the assessment of traffic impacts is dynamic, predictive, and highly contextualized.

[0055] Construction benefit item A work (S t ): This item is used to characterize the scheme S t The core of the support for construction activities is to ensure the necessary working space to improve construction efficiency and safety. t The area of ​​the effective construction zone enclosed by the determined coordinates. It's important to note that this term appears as a negative cost term in the optimization objective function, which is mathematically equivalent to maximizing its own value. Therefore, in the process of minimizing overall cost, the system will naturally tend to choose the solution that provides a larger working area.

[0056] Form switching cost C switch (S t-1 ,S t ): This item is used to characterize the optimal form S from the previous cycle. t-1 Transform to the current candidate form S t The physical cost required is to add a certain amount of "inertia" to the system's decision-making, avoiding overly frequent or drastic, uneconomical morphological adjustments due to minor traffic fluctuations. In preliminary calculations that do not consider physical feedback, this cost term can be modeled as a function of the total displacement distance of all enclosure units that need to be moved. Furthermore, as described in step S3, the calculation of this cost term will be refined based on real-time feedback of the physical state information of the enclosure units, thereby forming a complete physical information closed loop.

[0057] Weight coefficient w d ,w a ,w c: These three parameters are non-negative weight coefficients, which are used to adjust the relative importance of the above three cost items in the total cost function. This set of weight coefficients is not fixed, but can be dynamically adjusted according to the context of construction. Specifically, the system introduces a dynamic weight adjustment mechanism. At the beginning of each decision cycle, the system will identify the core construction tasks that are currently in progress or about to begin from the process database obtained in step S1. Subsequently, the system will query a pre-configured process-weight mapping relationship knowledge base that is part of the present invention. The knowledge base associates different process types with a set of optimal weight coefficient combinations. For example, when the process is a task such as "asphalt paving" that requires a large area and continuous operation, the system will determine a higher construction benefit weight w from the mapping relationship. a When the process changes to a task such as "concrete curing" that only requires static work and has no dynamic demand for work space, the system will determine a higher traffic delay weight w d This context-aware dynamic adjustment mechanism enables the decision-making of this method to be more closely aligned with engineering practice, which is an important manifestation of its intelligence.

[0058] Faced with the vast and complex solution space composed of numerous construction enclosure unit locations, employing an efficient optimization algorithm is crucial. In this embodiment, a genetic algorithm modified with reinforcement learning, namely the RLGA hybrid algorithm, is preferably used to solve the above optimization problem. This algorithm combines the global search capabilities of the genetic algorithm with the adaptive learning capabilities of reinforcement learning.

[0059] The genetic algorithm part constitutes the main framework of the solution algorithm. t The algorithm starts from an initial population and iteratively performs operations such as selection, crossover, and mutation to continuously evolve a population with higher fitness, i.e., the comprehensive cost J(S t ) A lower population. The fitness evaluation of each individual needs to be completed by calling the digital twin model in step S1 to perform a complete forward-looking simulation, as described above.

[0060] The innovation of this embodiment lies in the deep improvement of the mutation link of the traditional genetic algorithm, and the introduction of reinforcement learning agents for guidance. Its purpose is to upgrade the mutation operation from random exploration to guided, experience-based intelligent exploration, thereby significantly improving the optimization efficiency. Specifically, a reinforcement learning agent is embedded in the mutation operator. The state (State) of the agent is defined as the traffic situation feature vector of the current road network, which is directly derived from the fusion result of the real-time traffic data in step S1. Its action (Action) space consists of a set of predefined mutation strategies with clear physical meanings, such as "contracting mutation in the direction of the most congested lane", "expanding mutation along the idle lane" or "performing small-scale Gaussian perturbations". Its reward (Reward) is set as the amount of fitness improvement of the offspring individual relative to the parent individual after a certain mutation action is performed on the parent individual. By adopting reinforcement learning algorithms such as Q-Learning or its deep variants, the intelligent agent continuously conducts trial and error and learns during the solution process, and eventually understands which traffic situation and which mutation strategy are most likely to quickly find a better solution.

[0061] In summary, step S2 constructs a context-aware optimization objective function and uses an efficient hybrid intelligent algorithm guided by reinforcement learning to solve it. Ultimately, it is able to find an optimal construction form solution from a large number of possibilities within a decision cycle. This plan represents the best trade-off between traffic impact and construction benefits under current conditions, and will be passed to step S3 as a clear execution instruction.

[0062] In this embodiment, step S3 constitutes the concrete embodiment of the process from digital decision-making to execution in the physical world in the method of the present invention, and is the key hub connecting the information space and the physical space. The core purposes of this step are twofold: first, to accurately and flexibly implement the optimal construction form plan generated in step S2 through a group of physical execution units; second, to establish a closed-loop information feedback channel from the physical execution unit to the central decision-making system about the physical state of the unit itself. This step is not a simple instruction execution, but an indispensable link in the entire adaptive control closed loop. By introducing the real constraints of the physical world, it enables the decision-making process of the present invention to get rid of idealized assumptions, thereby having higher robustness and practical feasibility.

[0063] The physical execution of the method of this embodiment is preferably achieved through a set of modular intelligent enclosure units that can be centrally controlled. These units work together to form a physical construction boundary that can be dynamically reconfigured according to instructions.

[0064] When the decision core of step S2 calculates the optimal construction form plan for this cycle After that, the solution will first be parsed by the system into specific movement instructions for each independent modular intelligent enclosure unit. This instruction is a data structure containing the target three-dimensional coordinates of each unit (or the two-dimensional coordinates and attitude angles in the site coordinate system). Subsequently, these instructions are accurately sent to the corresponding enclosure units through an industrial-grade wireless communication network with high reliability and appropriate latency characteristics. Depending on the different requirements for real-time performance, the network can be a low-power wide area network optimized for long-distance, low-power status reporting, or a dedicated wireless LAN or 5G network slice deployed for high-density instruction interaction during morphological reconstruction.

[0065] Each modular intelligent enclosure unit is designed as an intelligent robotic terminal integrating perception, decision-making, and execution capabilities. Its hardware components include at least: a drive system for autonomous mobility, preferably using omnidirectional wheels or differential drive for greater maneuverability on crowded construction sites; a high-precision positioning system, a prerequisite for precise deployment, preferably utilizing ultra-wideband (UWB) technology that provides stable centimeter-level accuracy even in obscured environments, or real-time dynamic differential (RTK)-GPS technology in open areas; and an onboard microcontroller, serving as its "local brain."

[0066] After receiving instructions from the central system, the onboard microcontroller is responsible for executing low-level tasks. It uses its own positioning system to determine its current position and compare it with the target position in the instructions. From there, it runs a simple path planning algorithm to move to the target point in a collision-free manner, performing closed-loop motion control during movement using its motor encoders and inertial measurement unit (IMU). Through the coordinated actions of all units within the site, the physical boundaries of the construction area can be reconstructed to shrink, expand, or translate according to the requirements of the optimal solution.

[0067] Simple open-loop instruction execution cannot cope with the uncertainties inherent in the physical world, such as friction, wear, and power consumption. Therefore, a key technical feature of the present invention is the establishment of an upward information feedback loop from the physical execution unit to the central decision-making core after physical execution is completed, centered on the unit's own physical state.

[0068] Specifically, after each modular intelligent enclosure unit completes its movement task, reaches the designated location, or times out, its onboard microcontroller immediately collects and packages a series of key information representing its own physical state. It should be noted that this information is not about the external traffic environment, but about the health status and continued working ability of the execution unit itself. Preferably, this status information includes at least the following categories:

[0069] Final pose confirmation: This is the unit's final actual position and pose at the end of the mission, as reported by its localization system. This information is crucial because it allows the central system to determine whether instructions were fully executed and whether the unit became stuck, strayed from its target, or moved unexpectedly. The central system uses this information to update the physical boundaries in the digital twin, ensuring absolute realism in the simulated environment.

[0070] Remaining energy information: This information is provided by the battery management system (BMS) within the unit and is typically expressed as a precise percentage of remaining power (State of Charge). It is a key indicator of whether the unit can continue to perform high-intensity tasks.

[0071] A quantitative health index: This index aims to translate the fuzzy concept of unit "health" into a calculable and comparable value. It is a comprehensive assessment derived from the integration of data from multiple internal sensors. For example, this can be achieved by continuously monitoring the drive motor's operating current for abnormalities, operating temperature limits for violations, and the vibration spectrum measured by the onboard accelerometer for fault frequency signatures. After normalizing this raw data, a health index between 0 and 1 is calculated using a weighted model or a simple fuzzy logic inference engine. The closer the index is to 1, the better the unit's condition.

[0072] This packaged status information is then proactively uploaded to the central control system. The fundamental purpose of this is to provide accurate input about the actual status of the physical execution layer for subsequent decision-making cycles, thus forming a complete, intelligent closed loop from information decision-making to physical execution, and then from physical execution status to information decision-making.

[0073] The core application of this feedback information is to optimize the morphological switching cost term C in the objective function in step S2. switch (S t-1 ,S t ) for more accurate and dynamic quantification. With real feedback from the physical layer, the cost calculation is no longer simply a cumulative sum of displacement distances, but can be modeled as a function that deeply reflects the physical execution cost. For example, it can be rigorously specified as follows:

[0074]

[0075] in:

[0076] M moved Represents the S from the previous form t-1 Transform to the current candidate form S t The collection of all enclosure units that need to move in the process.

[0077] dist(p i,t-1 ,p i,t ) represents the Euclidean distance that the i-th enclosure unit in the set needs to move, which is the basic physical power consumption cost.

[0078] b i,t and h i,t They represent the remaining power and health status index of the i-th unit at the time of task execution, which are fed back by itself in real time.

[0079] f e (·) and f h (·) are two preset penalty functions, which are the key to integrating physical state into decision making in this invention. Both functions are monotonically decreasing functions. For example, f e (b i,t ) can be designed as 1 / (b i,t +∈) or exponential form where ∈ is a small positive number that prevents the denominator from being zero, k e is a coefficient. Its physical meaning is that the lower the power of a unit, the more expensive it is to move it. Similarly, f h (h i,t ) can be designed as 1 / (h i,t +∈). Its physical meaning is that the worse the health of a unit is, the higher the risk cost of moving it (e.g. the risk of complete damage during movement).

[0080] c1 is a basic cost coefficient used to calibrate the basic cost of moving per unit distance and to balance the dimensions of this switching cost with the traffic delay cost and construction benefit item.

[0081] By introducing this cost calculation method based on physical feedback, the decision-making system S2 of the present invention is able to perceive the state of the physical execution layer during the next round of optimization and make more robust decisions. For example, the system will automatically and inherently tend to avoid dispatching a containment unit that is low on battery or showing signs of early failure, even if moving it would bring traffic benefits from a pure traffic flow simulation. This mechanism greatly improves the long-term operational reliability and robustness of the entire system, ensuring that optimal decisions in the digital world can be economically and reliably executed in the physical world, thereby achieving a truly intelligent closed-loop control system.

[0082] In this embodiment, step S4 defines the overall operational framework and core working mode of the method of the present invention. Its purpose is to organically connect the aforementioned steps S1 (information acquisition and modeling), S2 (dynamic optimization decision-making), and S3 (flexible physical execution) to form a continuous, repetitive closed-loop control process. This step is the methodological guarantee for the present invention to achieve continuous self-adaptation. It ensures that the system is not a one-time static optimization, but can, like a living organism, dynamically synchronize with the ever-changing traffic environment and the progressive construction needs, continuously seeking and executing the optimal solution under the current conditions.

[0083] The method of the present invention operates in a cycle with a preset, fixed "decision-making cycle" as the basic time unit. The duration of this decision-making cycle is a key operating parameter of the present method. Its setting requires a scientific balance of two factors: on the one hand, the duration of the cycle needs to be long enough to ensure that the physical execution unit in step S3 has sufficient time to complete the morphological reconstruction task; on the other hand, the duration of the cycle needs to be short enough to ensure that the system can perceive significant changes in traffic flow status in a timely manner and respond to them, avoiding decision-making lags. Therefore, the setting of this cycle is an engineering parameter that needs to be determined based on the specific construction scenario and the physical properties of the execution unit used.

[0084] Within a single decision cycle, the system executes steps S1, S2, and S3 in strict sequence. At the end of a cycle, namely step S3, the system does not terminate but seamlessly and automatically initiates the next new decision cycle. This cycle begins when the method is applied to a construction project and continues until completion, enabling dynamic control of the entire construction process over time.

[0085] The core of this invention lies in its closed loop nature. The key to achieving this closed loop is the seamless and lossless handover of states and information between two adjacent decision cycles. The following uses the transition from decision cycle t to decision cycle t+1 as an example to illustrate this closed loop information flow:

[0086] At the end of decision cycle t, step S3 completes the physical form reconstruction and feeds back the final actual status of all execution units (including confirmed location, remaining power, health status, etc.) to the central system. Then, decision cycle t+1 starts:

[0087] In the S1 phase of cycle t+1:

[0088] The system begins collecting the latest real-time traffic data. It's important to emphasize that the traffic flow data collected at this point, such as vehicle speeds and queue lengths, has already been impacted by the morphological adjustments performed in cycle t. In other words, the system is observing the real-world consequences of its previous actions. Step S1 uses this latest traffic data, reflecting the "post-action" reality, to calibrate the digital twin model and predict traffic flow for the next cycle. This process ensures that the model's evolution is based on a continuous, verified reality foundation.

[0089] In the S2 phase of cycle t+1:

[0090] The system begins to build and solve the optimization problem of this cycle. At this time, the "initial state" or "reference form" it is based on is the optimal form after the execution and confirmation of step S3 in cycle t. At the same time, when calculating the morphological switching cost The system uses the latest, most accurate data from step S3 of cycle t to determine the remaining battery life and health status of each enclosure unit. This ensures that the next round of decisions is based on the final results of the previous cycle's actions and a comprehensive understanding of the latest capabilities of the physical execution units.

[0091] In the S3 phase of cycle t+1: the system calculates the new optimal morphological solution calculated in the S2 phase and applicable to cycle t+1 It is then issued and executed. After execution, the latest physical state is fed back to the system again, providing a basis for decision-making in the upcoming decision cycle t+2. Through this seamless cross-cycle state handover mechanism, the present invention continuously advances the classic cybernetic closed loop of "perception-decision-action-feedback" in one decision cycle after another. Each cycle is a complete adaptive adjustment process based on the latest reality. This cyclical operation mode gives the method proposed in this invention strong environmental adaptability and robustness. Whether facing the inherent and predictable tidal characteristics of road traffic, such as the alternation of morning and evening peaks and off-peaks, or facing occasional and unpredictable traffic events, such as sudden congestion caused by a traffic accident upstream, the system can automatically perceive changes within one or more subsequent decision cycles through its inherent cyclical mechanism and autonomously adjust the construction structure without human intervention to adapt to the new boundary conditions, continuously maintaining a dynamic balance between traffic impact and construction benefits.

[0092] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A segmented construction method for urban roads that reduces traffic impact, characterized in that: The following steps are involved: Step 1: Acquire real-time traffic data representing current traffic conditions and preset construction process information; Step 2: Calculating through an optimization objective function, based on the real-time traffic data and construction process information, dynamically generating a construction form plan that aims to balance traffic efficiency and construction work benefits; Step 3: Generate an adjustment instruction for the physical boundary of the construction area according to the construction form plan and execute the adjustment; Step 4: Execute steps 1 to 3 repeatedly in a preset decision cycle to achieve adaptive control of the construction area.

2. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: Step 1 further includes: building and calibrating in real time a traffic digital twin model synchronized with the physical world based on the real-time traffic data.

3. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: The dynamic generation of the construction form scheme in step 2 is specifically as follows: in the traffic digital twin model, a plurality of candidate construction form schemes are forward-looking simulated to predict the impact of each candidate scheme on future traffic conditions, and the construction form scheme is selected based on merit.

4. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: The optimization objective function at least includes: Traffic delay cost item, which is used to characterize the impact of the construction scheme on traffic efficiency; construction benefit item, which is used to characterize the size of the working space provided by the construction scheme; And the form switching cost term is used to represent the physical cost of transforming from the previous construction form to the current construction form.

5. The urban road segmented construction method for reducing traffic impact according to claim 4 is characterized in that: The weight of each cost item in the optimization objective function is dynamically adjusted according to the current construction process information obtained in step 1.

6. The urban road segmented construction method for reducing traffic impact according to claim 5 is characterized in that: The method further includes: determining the weight of each cost item in the optimization objective function from a preset process-weight mapping relationship according to the type of the current construction process.

7. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: The process of dynamically generating a solution by optimizing the objective function in step 2 is implemented by using a hybrid intelligent algorithm; The hybrid intelligent algorithm includes: using a genetic algorithm to perform a global search on the solution space of the construction form scheme; and using a reinforcement learning algorithm to guide the mutation direction of the genetic algorithm according to real-time traffic data.

8. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: The adjustment of the physical boundary of the construction area in step 3 is specifically: completing the contraction, expansion or translation of the physical boundary of the construction area by controlling the movement of multiple modular intelligent enclosure units.

9. The urban road segmented construction method for reducing traffic impact according to claim 1 is characterized in that: The calculation of the morphology switching cost item is based on the physical state information of the modular intelligent enclosure unit itself obtained in real time after the adjustment is performed.

10. The urban road segmented construction method for reducing traffic impact according to claim 9, characterized in that: The self-physical status information includes at least: the remaining power and / or health status of the modular intelligent enclosure unit.

Images