Construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture

Abstract

The invention discloses a construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture. The method comprises the following steps: S1, collecting microscopic level data of a construction site; s2, obtaining mid-level data of the construction site; s3, acquiring macroscopic hierarchical data, and constructing a multi-scale spatio-temporal data set; s4, decomposing and processing the multi-scale spatio-temporal data set by using a spatio-temporal diagram convolutional network, and constructing a dynamic digital twin model; s5, designing a competition and cooperation mechanism between agents by using a game theory model, solving resource allocation conflicts through Nash equilibrium, and dynamically optimizing resource scheduling; s6, optimizing a scheduling strategy of the intelligent agent through a distributed algorithm based on near-end strategy optimization; s7, detecting errors in digital-to-analog conversion and scheduling processes by using a belief propagation algorithm; and S8, optimizing a scheduling strategy through a distributed adaptive feedback control mechanism. According to the invention, a distributed computing architecture, space-time diagram convolution, an intelligent scheduling technology and the like are adopted, and digital-to-analog conversion and resource management of the construction site are optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of construction site management and resource scheduling, and in particular to a method for digital-to-analog conversion and collaborative management of construction sites based on a distributed computing architecture. Background Art

[0002] Construction site management and scheduling is one of the most complex and critical links in modern construction projects. As the scale of construction projects increases, the construction process becomes more complicated, and the coordination and scheduling of equipment, materials, and personnel becomes more difficult. In order to improve the management efficiency of construction sites, many researchers and industry experts have proposed construction site management methods based on informatization and intelligence. Among them, the use of technologies such as building information modeling (BIM) and geographic information system (GIS) has become a basic tool for construction site management. These technologies provide managers with visual and digital management methods by converting the data of construction projects into digital models. However, existing technologies still have many problems in data integration, dynamic scheduling, and real-time feedback.

[0003] Most existing technologies rely on BIM systems to build digital models of construction projects. BIM models provide managers with information such as construction progress, equipment distribution, and resource allocation through digital integration of design drawings and data. However, BIM systems usually only support static data information updates and lack real-time dynamic feedback mechanisms, making it difficult to cope with the complex and changeable construction environment of construction sites. In actual operations, the construction progress, resource scheduling, and equipment status of construction sites often change dynamically, and it is difficult to accurately reflect the actual situation on site by relying on static BIM models for management. This lagging information transmission and response mechanism makes it impossible for managers to quickly identify and solve problems on the construction site, seriously affecting the efficiency and quality of construction.

[0004] On this basis, in recent years, technologies such as drones and 3D laser scanning have been gradually introduced to obtain real-time data on construction sites. For example, drones can quickly obtain the spatial status of a construction site through aerial photography, and 3D laser scanning technology can accurately capture the spatial position and progress information of building structures. However, the amount of data collected by these technologies is huge and most of them are meso-level data. When processing these multi-source heterogeneous data, existing technologies lack effective data integration methods, and it is difficult to organically combine micro, meso and macro-level data to achieve global dynamic digital model construction. In addition, drones and 3D laser scanning technologies also face the problems of low data processing efficiency and poor real-time performance, and cannot support the rapid changes and scheduling needs of construction sites.

[0005] In response to resource scheduling issues on construction sites, traditional scheduling systems mostly rely on manual operations or automated systems based on fixed rules, lacking intelligent scheduling optimization mechanisms. Traditional scheduling systems usually use simple scheduling algorithms when dealing with dynamic scheduling of equipment and resources, and fail to fully consider the complex relationships between different equipment and tasks. For example, scheduling between key equipment such as cranes, transport vehicles, and concrete pumps often involves resource conflicts and priority competition. Existing scheduling algorithms are difficult to effectively resolve these conflicts, which can easily lead to resource waste or scheduling errors, ultimately affecting construction progress and resource utilization efficiency.

[0006] Although optimization algorithms such as game theory have been gradually applied to resource scheduling on construction sites, the application of these algorithms is still in its early stages. Traditional game theory models have difficulty handling the scheduling of numerous equipment and resources on construction sites, especially when multiple devices are operating in coordination and multiple tasks are executed in parallel. The computational complexity of the game model increases significantly, and existing technologies are difficult to ensure the real-time nature of the scheduling process. In addition, game theory models require a large amount of prior knowledge and historical data support. Current construction sites often lack a complete data collection and feedback mechanism, which limits the optimization effect of the model.

[0007] In scheduling optimization, although the existing technology proposes to use reinforcement learning algorithms (such as deep Q networks) to adaptively optimize scheduling strategies, due to the complex and changeable environment of construction sites, traditional reinforcement learning algorithms are relatively slow in processing real-time data and have difficulty responding to emergencies quickly. In addition, when facing multi-agent systems, traditional reinforcement learning algorithms often find it difficult to coordinate the competition and cooperation between agents, resulting in poor global optimization results. There are multiple devices and resources on a construction site, and the status and task requirements of each device and resource are constantly changing. Single-agent scheduling often has difficulty solving the problem of coordinated operation of multiple devices, further limiting the application of existing scheduling systems.

[0008] Error detection and feedback mechanism is also an important issue in construction site management. Traditional construction site management systems usually rely on manual monitoring and manual feedback, and lack automated error detection and correction mechanisms. When the construction progress of the construction site deviates from the digital model, the existing technology cannot identify and correct these errors in time, which can easily lead to failure or delay in site scheduling. In addition, most of the existing error detection algorithms rely on simple threshold settings, lack dynamic adjustment mechanisms, and cannot adapt to the changing needs under different construction site environments, affecting the management efficiency of the construction site.

[0009] Therefore, how to provide a digital-to-analog conversion and collaborative management method for construction sites based on a distributed computing architecture is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0010] One purpose of the present invention is to propose a method for digital-to-analog conversion and collaborative management of construction sites based on a distributed computing architecture. The present invention comprehensively optimizes multi-level data integration, dynamic digital-to-analog conversion, intelligent resource scheduling and error correction in construction sites through a distributed computing architecture, a spatiotemporal graph convolutional network, a game theory model and a proximal strategy optimization algorithm. The method realizes the real-time collection and integration of micro, meso and macro data, and dynamically schedules resources through an intelligent agent system, using a belief propagation algorithm to quickly correct errors. Compared with traditional technologies, the present invention has higher scheduling efficiency, accuracy and real-time response capabilities, significantly improving the level of intelligent management of construction sites.

[0011] The method for digital-to-analog conversion and collaborative management of a construction site based on a distributed computing architecture according to an embodiment of the present invention comprises the following steps:

[0012] S1. Collecting micro-level data of the construction site through a distributed sensor network, wherein the micro-level data includes equipment vibration, material changes, and temperature and humidity environmental parameters;

[0013] S2. Using drones and 3D laser scanning technology to obtain meso-level data of the construction site, the meso-level data includes equipment operation status, local construction progress and environmental changes;

[0014] S3. Obtain macro-level data from distributed building information models and geographic information systems, including overall construction progress, equipment distribution, and resource allocation, and integrate the macro-level data with meso-level data and micro-level data, pre-process using edge computing devices, and construct a multi-scale spatiotemporal data set;

[0015] S4. Use the spatiotemporal graph convolutional network to decompose and process multi-scale spatiotemporal datasets in a distributed computing environment, analyze the spatiotemporal correlation between data at different levels, and build a dynamic digital twin model of the construction site;

[0016] S5. Use intelligent agents to schedule key equipment or resources. Under the distributed computing architecture, use game theory models to design competition and cooperation mechanisms between intelligent agents, resolve resource allocation conflicts through Nash equilibrium, and dynamically optimize resource scheduling.

[0017] S6, optimize the agent's scheduling strategy through a distributed algorithm based on proximal policy optimization, combined with historical scheduling data and real-time feedback;

[0018] S7. Use the belief propagation algorithm to detect errors in the digital-to-analog conversion and scheduling process between distributed nodes and perform real-time correction. The correction information is quickly propagated in the distributed system to trigger the adjustment mechanism.

[0019] S8. Through the distributed adaptive feedback control mechanism, the scheduling strategy of the intelligent agent is continuously optimized based on historical scheduling data, resource utilization and real-time feedback.

[0020] Optionally, the S3 specifically includes:

[0021] S31. Obtain macro-level data on overall construction progress, equipment distribution, and resource allocation from distributed building information models and geographic information systems, and standardize the macro-level data into spatiotemporal coordinate representation;

[0022] S32, using a spatiotemporal alignment algorithm to process macro-level data, using a uniform time step and spatial resolution for alignment;

[0023] S33, integrating the macro-level data with the meso-level data to obtain macro-meso-level fusion data:

[0024]

[0025] Among them, M 宏中融合 (x, y, t) represents macro-meso fusion data, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial position x, y and time t, M 中观 (x, y, z, t) represents the data distribution of meso-level data at spatial position x, y, height z and time t, λ1 and λ2 represent weight factors, tanh represents the hyperbolic tangent function, α represents the regulation coefficient, β, γ and δ represent the adjustment factors, and ∈ represents the interaction adjustment coefficient between macro-level data and meso-level data;

[0026] S34. Integrate the macro-meso fusion data with the micro-level data to obtain multi-level data:

[0027]

[0028] Among them, M 融合 (x, y, z, t) represents the fused multi-level data, M 微观 (x, y, z, t) represents the data distribution of micro-level data at spatial position x, y, height z and time t, γ1 and γ2 represent weight factors, η, α1, α2 and ξ represent control factors;

[0029] S35. Preprocessing the integrated multi-level data using edge computing devices, wherein the preprocessing includes data cleaning, noise reduction, and removal of redundant data;

[0030] S36. Construct a multi-scale spatiotemporal dataset based on the preprocessed multi-level data.

[0031] Optionally, the S4 specifically includes:

[0032] S41, constructing a spatiotemporal graph structure G(V, E), wherein the node set V represents each key position of the construction site, the edge set E represents the spatiotemporal dependency relationship between the key positions, the characteristics of each node include macro-level data, meso-level data and micro-level data, and the spatiotemporal association between the nodes is defined by the adjacency matrix;

[0033] S42, initialize the feature matrix of each node, the feature matrix is:

[0034] X(t)=[M 宏观 (x,y,t),M 中观 (x, y, z, t), M 微观 (x, y, z, t)];

[0035] Among them, X(t) represents the node feature at time t, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial position x, y and time t, M 中观 (x, y, z, t) represents the data distribution of the meso-level data at spatial position x, y, height z and time t, M 微观 (x, y, z, t) represents the data distribution of micro-level data at spatial position x, y, height z and time t;

[0036] S43, decomposing and processing the multi-scale spatiotemporal dataset of the construction site through a spatiotemporal graph convolutional network, wherein the spatiotemporal graph convolutional network calculates the spatiotemporal dependencies between nodes through an adjacency matrix to extract high-dimensional features of the multi-scale spatiotemporal dataset;

[0037] S44, dynamically capturing the processed multi-scale spatiotemporal data set in the time dimension, using multi-layer temporal convolution operations to process feature changes in the time step, and forming a high-dimensional feature representation of spatiotemporal association;

[0038] S45. Construct a dynamic digital twin model of the construction site based on high-dimensional feature representation of spatiotemporal associations:

[0039]

[0040] Among them, M 动态孪生 (x, y, z, t) represents the dynamic digital twin model of the construction site, M 时空 (x, y, z, t) represents the spatial features extracted by spatiotemporal convolution, M 时间 (x, y, z, t) represents the temporal features extracted by temporal convolution, and β1, β2, β3, τ, ζ and κ represent the control parameters.

[0041] Optionally, the S5 specifically includes:

[0042] S51, assigning an independent intelligent agent to each key equipment or resource in the construction site, each intelligent agent obtains the status information of the responsible equipment or resource through the real-time data acquisition system in the distributed computing architecture, the status information including the location, availability and current task load;

[0043] S52. The competition and cooperation relationship between agents is modeled through game model to make scheduling decisions:

[0044]

[0045] Among them, U i represents the utility function of the ith agent, D i (t) represents the scheduling strategy of the ith agent, μ i represents the scheduling weight of the ith agent, f i (D i (t)) represents the scheduling strategy function of the ith agent, g ij (D i (t), D j (t)) represents the interaction utility between the ith agent and the jth agent, v i , i , κ i , δ i and i represents the adjustment coefficient;

[0046] S53. When there is resource competition between agents, the utility function U is calculated through the game theory model. i Evaluate the benefits of each agent and dynamically allocate resources through competition and cooperation mechanisms. Each agent is assigned a certain amount of resources based on the utility function U. i The optimal solution to adjust the scheduling strategy;

[0047] S54. Resolve resource allocation conflicts between agents through Nash equilibrium and maximize the global utility of agents:

[0048]

[0049] Among them, D * represents the optimal scheduling strategy combination, N represents the total number of agents, represents the utility of the ith agent when the strategies of other agents remain unchanged, ρ i ,θ i , and represents the control parameters;

[0050] S55. Under the distributed computing architecture, the intelligent agent shares the current status information, task requirements and resource availability in real time through the network, dynamically responds to the real-time changes of the construction site, and adjusts the resource scheduling strategy;

[0051] S56. During the scheduling process, the intelligent agent continuously monitors the status information of the equipment and the progress of task execution, optimizes the scheduling strategy through feedback from the game theory model, and uses Nash equilibrium to achieve adaptive optimization of resource scheduling.

[0052] Optionally, the S6 specifically includes:

[0053] S61. Initialize the scheduling strategy π of each agent i , the agent collects historical scheduling data H through nodes in the distributed computing architecture i (t) and real-time feedback data F i (t), where H i (t) represents the historical scheduling strategy of the ith agent at time t, F i (t) represents the real-time feedback data of the i-th agent at time t;

[0054] S62. Using a distributed algorithm based on proximal policy optimization, the agent evaluates the performance of the current scheduling strategy, combines historical scheduling data and real-time feedback data to optimize the strategy, and performs deviation analysis and value assessment:

[0055]

[0056] Among them, V i (π i ) indicates that the i-th agent is in strategy π i The value evaluation function under , t represents the time step, T represents the total time step, A represents the discount factor, R i (t) represents the immediate reward obtained by the i-th agent at time t, α v , β v , η v , v , δ v and v represents the adjustment parameter;

[0057] S63, optimize the current scheduling strategy of the agent through the distributed algorithm of proximal strategy optimization:

[0058]

[0059] in, represents the optimized agent scheduling strategy, π i (a i |s i) represents the state s of the i-th agent under the new strategy i Select action a i The probability of Represents the state s of the i-th agent under the old strategy i Select action a i The probability of N is the total number of agents, α π , β π , π , δ π and η π represents the adjustment parameter of the strategy update, U i (s i , a i ) represents the utility function of the ith agent, U j (s j , a j ) represents the utility function of the jth agent;

[0060] S64. In the process of optimizing the strategy, the intelligent agents communicate and share data in real time through the distributed computing architecture. The intelligent agents transmit the current scheduling status and feedback information through the network, coordinate the scheduling behavior of other intelligent agents, and optimize the global allocation;

[0061] S65, according to the optimized intelligent agent scheduling strategy The scheduling strategy is adjusted based on real-time feedback. Each agent re-evaluates equipment availability, task priority, and resource allocation requirements through updated strategies.

[0062] S66. After completing the scheduling task, the intelligent agent continuously updates the scheduling strategy parameters through the accumulation of historical data and real-time feedback monitoring, and gradually approaches the global optimal solution through multiple scheduling strategy iterations.

[0063] Optionally, the S7 specifically includes:

[0064] S71. Through the distributed sensor network and feedback system, the real-time data of each node is obtained, including the equipment status, task progress and resource allocation. The feedback data of each node is recorded as D i (t), where D i (t) represents the state data of node i at time t;

[0065] S72, based on node feedback data D i (t) and the expected value P of the global digital twin model i (t), calculate the initial error of each node:

[0066]

[0067] Among them, E i(t) represents the error value of node i at time t, α e , β e , e and e represents the adjustment parameter;

[0068] S73. Use the belief propagation algorithm to transmit error information between distributed nodes, and each node updates its own belief value according to the error value of the neighboring node:

[0069]

[0070] Among them, B i (t+1) represents the belief value of node i at time t+1, B i (t) represents the belief value of node i at time t, N(i) represents the set of neighbor nodes of node i, β ij represents the belief propagation coefficient between node i and node j, E j (t) represents the error value of node j at time t, α b , γ b , b and δ b represents the adjustment parameter, β ik represents the belief propagation coefficient between node i and node k;

[0071] S74, according to the updated belief value B i (t+1), each node adjusts the scheduling parameters and the digital-to-analog conversion control signal, and the node reduces the feedback data D according to the result of belief propagation. i (t) and the expected value P of the global digital twin model i (t) Deviation between;

[0072] S75. During the iterative process of belief propagation, each node continuously exchanges and updates error information with neighboring nodes until the error value of the global system converges and eventually reaches global optimization.

[0073] S76. When the system detects that the error exceeds a predetermined threshold, a distributed error correction mechanism is triggered. The error correction mechanism adjusts the scheduling decision and parameter configuration of the node according to the result of belief propagation, and corrects the error in the digital-to-analog conversion in real time.

[0074] The beneficial effects of the present invention are:

[0075] First, the present invention realizes the comprehensive capture of construction site data from micro to macro levels through multi-level data integration and processing. Different from the practice of relying on static BIM models or single data sources in the prior art, the present invention can collect dynamic data of the construction site in real time, including multi-dimensional information such as equipment vibration, construction progress, and environmental parameters, by using distributed sensor networks, drones, and 3D laser scanning technology, and fuse and pre-process these data. Through the application of spatiotemporal alignment algorithms and spatiotemporal graph convolutional networks, data at different levels can be effectively integrated to construct a global multi-scale spatiotemporal data set to ensure that the physical state of the construction site is always consistent with the digital twin model. This dynamic digital-to-analog conversion mechanism enables changes in the construction site to be captured and fed back in real time, greatly improving the accuracy and precision of management.

[0076] Secondly, the present invention proposes a highly intelligent solution in resource scheduling. Unlike traditional scheduling systems that rely on regularized and manual operations, the present invention uses intelligent agents to be responsible for the scheduling of key equipment and resources, and can automatically manage and allocate resources. Between intelligent agents, game theory models are used to deal with competition and cooperation relationships, and the resource allocation strategy is dynamically adjusted through the Nash equilibrium mechanism to ensure that resources can still be allocated efficiently and reasonably in the case of resource shortages or task conflicts. The intelligent agent uses the proximal policy optimization algorithm (PPO) combined with historical scheduling data and real-time feedback to continuously optimize its scheduling decisions, so that the scheduling efficiency is gradually improved, and it can quickly respond to sudden changes in the construction site. This dynamic optimization and adaptive learning mechanism makes resource scheduling more flexible and intelligent, and solves the problem of slow response and scheduling errors of traditional scheduling methods in the face of complex and changeable construction site environments.

[0077] In addition, the distributed belief propagation algorithm of the present invention significantly improves the error detection and correction capabilities in construction site management. In the prior art, error detection at construction sites relies on manual monitoring or simple threshold settings, which makes it difficult to cope with complex and changeable construction environments. The present invention uses a belief propagation algorithm to transmit error information between distributed nodes, and is able to detect deviations in the digital-to-analog conversion and scheduling process in real time and perform rapid corrections. The belief propagation algorithm rapidly propagates error information in a distributed system, triggers dynamic adjustment of nodes, and ensures global optimization of construction site management. This real-time correction mechanism not only improves the system's sensitivity to changes in the construction site, but also effectively avoids global management failures caused by local errors, greatly improving the reliability and accuracy of construction site scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0079] Figure 1 A flowchart of the construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture proposed by the present invention;

[0080] Figure 2 This is a schematic diagram of the multi-scale spatiotemporal data integration and processing structure of the construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture proposed in the present invention. DETAILED DESCRIPTION

[0081] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0082] refer to Figure 1 and Figure 2 , a construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture, comprising the following steps:

[0083] S1. Collecting micro-level data of the construction site through a distributed sensor network, wherein the micro-level data includes equipment vibration, material changes, and temperature and humidity environmental parameters;

[0084] S2. Using drones and 3D laser scanning technology to obtain meso-level data of the construction site, the meso-level data includes equipment operation status, local construction progress and environmental changes;

[0085] S3. Obtain macro-level data from distributed building information models and geographic information systems, including overall construction progress, equipment distribution, and resource allocation, and integrate the macro-level data with meso-level data and micro-level data, pre-process using edge computing devices, and construct a multi-scale spatiotemporal data set;

[0086] S4. Use the spatiotemporal graph convolutional network to decompose and process multi-scale spatiotemporal datasets in a distributed computing environment, analyze the spatiotemporal correlation between data at different levels, and build a dynamic digital twin model of the construction site;

[0087] S5. Use intelligent agents to schedule key equipment or resources. Under the distributed computing architecture, use game theory models to design competition and cooperation mechanisms between intelligent agents, resolve resource allocation conflicts through Nash equilibrium, and dynamically optimize resource scheduling.

[0088] S6, optimize the agent's scheduling strategy through a distributed algorithm based on proximal policy optimization, combined with historical scheduling data and real-time feedback;

[0089] S7. Use the belief propagation algorithm to detect errors in the digital-to-analog conversion and scheduling process between distributed nodes and perform real-time correction. The correction information is quickly propagated in the distributed system to trigger the adjustment mechanism.

[0090] S8. Through the distributed adaptive feedback control mechanism, the scheduling strategy of the intelligent agent is continuously optimized based on historical scheduling data, resource utilization and real-time feedback.

[0091] In this implementation, S3 specifically includes:

[0092] S31. Obtain macro-level data on overall construction progress, equipment distribution, and resource allocation from distributed building information models and geographic information systems, and standardize the macro-level data into spatiotemporal coordinate representation;

[0093] S32, using a spatiotemporal alignment algorithm to process macro-level data, using a uniform time step and spatial resolution for alignment;

[0094] S33, integrating the macro-level data with the meso-level data to obtain macro-meso-level fusion data:

[0095]

[0096] Among them, M 宏中融合 (x, y, t) represents macro-meso fusion data, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial position x, y and time t, M 中观 (x, y, z, t) represents the data distribution of meso-level data at spatial position x, y, height z and time t, λ1 and λ2 represent weight factors, tanh represents the hyperbolic tangent function, α represents the regulation coefficient, β, γ and δ represent the adjustment factors, and ∈ represents the interaction adjustment coefficient between macro-level data and meso-level data;

[0097] S34. Integrate the macro-meso fusion data with the micro-level data to obtain multi-level data:

[0098]

[0099] Among them, M 融合 (x, y, z, t) represents the fused multi-level data, M 微观 (x, y, z, t) represents the data distribution of micro-level data at spatial position x, y, height z and time t, γ1 and γ2 represent weight factors, η, α1, α2 and ξ represent control factors;

[0100] S35. Preprocessing the integrated multi-level data using edge computing devices, wherein the preprocessing includes data cleaning, noise reduction, and removal of redundant data;

[0101] S36. Construct a multi-scale spatiotemporal dataset based on the preprocessed multi-level data.

[0102] In this implementation, S4 specifically includes:

[0103] S41, constructing a spatiotemporal graph structure G(V, E), wherein the node set V represents each key position of the construction site, the edge set E represents the spatiotemporal dependency relationship between the key positions, the characteristics of each node include macro-level data, meso-level data and micro-level data, and the spatiotemporal association between the nodes is defined by the adjacency matrix;

[0104] S42, initialize the feature matrix of each node, the feature matrix is:

[0105] X(t)=[M 宏观 (x,y,t),M 中观 (x, y, z, t), M 微观 (x, y, z, t)];

[0106] Among them, X(t) represents the node feature at time t, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial position x, y and time t, M 中观 (x, y, z, t) represents the data distribution of the meso-level data at spatial position x, y, height z and time t, M 微观 (x, y, z, t) represents the data distribution of micro-level data at spatial position x, y, height z and time t;

[0107] S43, decomposing and processing the multi-scale spatiotemporal dataset of the construction site through a spatiotemporal graph convolutional network, wherein the spatiotemporal graph convolutional network calculates the spatiotemporal dependencies between nodes through an adjacency matrix to extract high-dimensional features of the multi-scale spatiotemporal dataset;

[0108] S44, dynamically capturing the processed multi-scale spatiotemporal data set in the time dimension, using multi-layer temporal convolution operations to process feature changes in the time step, and forming a high-dimensional feature representation of spatiotemporal association;

[0109] S45. Construct a dynamic digital twin model of the construction site based on high-dimensional feature representation of spatiotemporal associations:

[0110]

[0111] Among them, M 动态孪生 (x, y, z, t) represents the dynamic digital twin model of the construction site, M 时空 (x, y, z, t) represents the spatial features extracted by spatiotemporal convolution, M 时间 (x, y, z, t) represents the temporal features extracted by temporal convolution, and β1, β2, β3, τ, ζ and κ represent the control parameters.

[0112] In this implementation manner, S5 specifically includes:

[0113] S51, assigning an independent intelligent agent to each key equipment or resource in the construction site, each intelligent agent obtains the status information of the responsible equipment or resource through the real-time data acquisition system in the distributed computing architecture, the status information including the location, availability and current task load;

[0114] S52. The competition and cooperation relationship between agents is modeled through game model to make scheduling decisions:

[0115]

[0116] Among them, U i represents the utility function of the ith agent, D i (t) represents the scheduling strategy of the ith agent, μ i represents the scheduling weight of the ith agent, f i (D i (t)) represents the scheduling strategy function of the ith agent, g ij (D i (t), D j (t)) represents the interaction utility between the ith agent and the jth agent, v i , i , κ i ,δ i and i represents the adjustment coefficient;

[0117] S53. When there is resource competition between agents, the utility function U is calculated through the game theory model. i Evaluate the benefits of each agent and dynamically allocate resources through competition and cooperation mechanisms. Each agent is assigned a certain amount of resources based on the utility function U. i The optimal solution to adjust the scheduling strategy;

[0118] S54. Resolve resource allocation conflicts between agents through Nash equilibrium and maximize the global utility of agents:

[0119]

[0120] Among them, D * represents the optimal scheduling strategy combination, N represents the total number of agents, represents the utility of the ith agent when the strategies of other agents remain unchanged, ρ i ,θ i , and represents the control parameters;

[0121] S55. Under the distributed computing architecture, the intelligent agent shares the current status information, task requirements and resource availability in real time through the network, dynamically responds to the real-time changes of the construction site, and adjusts the resource scheduling strategy;

[0122] S56. During the scheduling process, the intelligent agent continuously monitors the status information of the equipment and the progress of task execution, optimizes the scheduling strategy through feedback from the game theory model, and uses Nash equilibrium to achieve adaptive optimization of resource scheduling.

[0123] In this implementation manner, S6 specifically includes:

[0124] S61. Initialize the scheduling strategy π of each agent i , the agent collects historical scheduling data H through nodes in the distributed computing architecture i (t) and real-time feedback data F i (t), where H i (t) represents the historical scheduling strategy of the ith agent at time t, F i (t) represents the real-time feedback data of the i-th agent at time t;

[0125] S62. Using a distributed algorithm based on proximal policy optimization, the agent evaluates the performance of the current scheduling strategy, combines historical scheduling data and real-time feedback data to optimize the strategy, and performs deviation analysis and value assessment:

[0126]

[0127] Among them, V i (π i ) indicates that the i-th agent is in strategy π i The value evaluation function under , t represents the time step, T represents the total time step, A represents the discount factor, R i (t) represents the immediate reward obtained by the i-th agent at time t, α v , β v , η v , v , δ v and v represents the adjustment parameter;

[0128] S63, optimize the current scheduling strategy of the agent through the distributed algorithm of proximal strategy optimization:

[0129]

[0130] in, represents the optimized agent scheduling strategy, π i (a i |s i) represents the state s of the i-th agent under the new strategy i Select action a i The probability of Represents the state s of the i-th agent under the old strategy i Select action a i The probability of N is the total number of agents, α π , β π , π , δ π and η π represents the adjustment parameter of the strategy update, U i (s i , a i ) represents the utility function of the ith agent, U j (s j , a j ) represents the utility function of the jth agent;

[0131] S64. In the process of optimizing the strategy, the intelligent agents communicate and share data in real time through the distributed computing architecture. The intelligent agents transmit the current scheduling status and feedback information through the network, coordinate the scheduling behavior of other intelligent agents, and optimize the global allocation;

[0132] S65, according to the optimized intelligent agent scheduling strategy The scheduling strategy is adjusted based on real-time feedback. Each agent re-evaluates equipment availability, task priority, and resource allocation requirements through updated strategies.

[0133] S66. After completing the scheduling task, the intelligent agent continuously updates the scheduling strategy parameters through the accumulation of historical data and real-time feedback monitoring, and gradually approaches the global optimal solution through multiple scheduling strategy iterations.

[0134] In this implementation manner, the S7 specifically includes:

[0135] S71. Through the distributed sensor network and feedback system, the real-time data of each node is obtained, including the equipment status, task progress and resource allocation. The feedback data of each node is recorded as D i (t), where D i (t) represents the state data of node i at time t;

[0136] S72, based on node feedback data D i (t) and the expected value P of the global digital twin model i (t), calculate the initial error of each node:

[0137]

[0138] Among them, E i(t) represents the error value of node i at time t, α e , β e , e and e represents the adjustment parameter;

[0139] S73. Use the belief propagation algorithm to transmit error information between distributed nodes, and each node updates its own belief value according to the error value of the neighboring node:

[0140]

[0141] Among them, B i (t+1) represents the belief value of node i at time t+1, B i (t) represents the belief value of node i at time t, N(i) represents the set of neighbor nodes of node i, β ij represents the belief propagation coefficient between node i and node j, E j (t) represents the error value of node j at time t, α b , γ b , b and δ b represents the adjustment parameter, β ik represents the belief propagation coefficient between node i and node k;

[0142] S74, according to the updated belief value B i (t+1), each node adjusts the scheduling parameters and the digital-to-analog conversion control signal, and the node reduces the feedback data D according to the result of belief propagation. i (t) and the expected value P of the global digital twin model i (t) Deviation between;

[0143] S75. During the iterative process of belief propagation, each node continuously exchanges and updates error information with neighboring nodes until the error value of the global system converges and eventually reaches global optimization.

[0144] S76. When the system detects that the error exceeds a predetermined threshold, a distributed error correction mechanism is triggered. The error correction mechanism adjusts the scheduling decision and parameter configuration of the node according to the result of belief propagation, and corrects the error in the digital-to-analog conversion in real time.

[0145] Embodiment 1:

[0146] In order to verify the feasibility of the present invention in implementation, the present invention is applied to a large-scale construction project. The project is large-scale, covers an area of ​​about 50,000 square meters, involves the simultaneous construction of multiple high-rise buildings, and the construction period is estimated to be three years. The project faced the following difficulties during the construction process: due to the complex site and numerous tasks, the equipment and resource scheduling efficiency on site is low, and competition between equipment often leads to task delays; in addition, data is not updated in time during the construction process, and it is difficult for managers to grasp the real-time construction progress in time, resulting in a lag in the overall progress of the project. In order to cope with these problems, the project management party decided to adopt a construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture to improve construction efficiency and ensure the optimal allocation of construction resources.

[0147] In this project, the construction management deployed a variety of data collection equipment, including distributed sensor networks, drones, and 3D laser scanners. Through the distributed sensor network, micro-level data, such as vibration of construction equipment, changes in construction materials, temperature and humidity, and other parameters, are collected in real time. These data are pre-processed by edge computing devices, such as data cleaning, noise reduction, and redundancy removal. At the same time, the management also uses drones to regularly take aerial photos of the construction site, and combines 3D laser scanning technology to capture meso-level data, including the operating status of equipment, construction progress, and changes in the construction environment. These meso-level data can help managers understand the construction progress of each building and the coordination between equipment. Macro-level data comes from distributed building information models (BIM) and geographic information systems (GIS), which provide the overall construction progress of the building, equipment distribution, and resource allocation. These data are collected and processed through a distributed computing architecture to form multi-scale spatiotemporal data integration.

[0148] At the construction site, each key equipment, such as cranes, concrete pumps, and transport vehicles, is assigned an independent agent. The agent continuously monitors the status, task load, and availability of the equipment through a real-time data acquisition system. Since the project is simultaneously constructing multiple buildings, resource scheduling conflicts are particularly prominent. For example, when multiple high-rise buildings need to use cranes at the same time, traditional scheduling methods are difficult to reasonably allocate crane resources, which can easily lead to resource waste or delays. In this embodiment, the agent competes and cooperates with other equipment agents through a game theory model. The agent calculates based on the utility function in the game model and dynamically adjusts its scheduling strategy through the Nash equilibrium mechanism. When a building has a higher construction priority, crane resources will be allocated to that building first, while other low-priority tasks will be automatically scheduled and optimized to avoid idleness and waste of equipment.

[0149] In addition, the present invention also applies the belief propagation algorithm to monitor the errors in digital-to-analog conversion in real time. During the construction process on the construction site, the management personnel monitor the construction progress through the digital twin model of the building. However, due to the complexity of the construction environment on the construction site, there is often a certain deviation between the digital model and the actual progress. Through the belief propagation algorithm, the data error of each node can be propagated and corrected in the distributed system in a timely manner. In order to verify the overall effect of the present invention, the project management party conducted comparative tests before and after adopting the present invention, and the test results are shown in the following table:

[0150] Table 1 Performance comparison test data table of not using the present invention and using the present invention

[0151] Comparison Items Not using the present invention Using the Invention Average dispatch delay time 180 minutes / day 25 minutes / day Equipment utilization 68％ 92％ Construction progress is delayed 2 days / month 0.5 days / month Scheduling errors caused by resource conflicts 20 times / week 4 times / week Digital-to-analog conversion error correction time Unable to correct 45 minutes Error detection times (3 months) 0 times (no detection mechanism) 6 times Error correction accuracy Unable to evaluate Deviation rate within 1% Project progress control accuracy Approx. ±5% ±1％ Resource scheduling optimization times (3 months) None (manual scheduling) 35 times (automatic optimization) Project completed ahead of schedule None (Completed on schedule) 5 days in advance

[0152] Through the comparative test data in Table 1 above, we can clearly see the remarkable effect of the present invention in construction site scheduling and management. Without using the present invention, the average scheduling delay time is 180 minutes / day, which means that the equipment scheduling of the construction site often has a delay of about three hours in a day's work. After using the present invention, the scheduling delay time is significantly reduced to 25 minutes / day. This shows that the intelligent scheduling system can greatly improve the scheduling efficiency of equipment, quickly respond to the real-time needs of the construction site, and avoid unnecessary delays.

[0153] The data on equipment utilization also shows a significant improvement. When the present invention is not used, the equipment utilization is only 68%, which means that a considerable part of the equipment is idle or inefficiently used. However, through the optimized scheduling mechanism of the present invention, the equipment utilization rate is increased to 92%. This improvement not only reduces the idle time of equipment, but also maximizes the utilization efficiency of resources, thereby shortening the construction period and reducing costs.

[0154] In terms of construction progress control, when the present invention was not used, the monthly construction progress of the project lagged by an average of 2 days, which was mainly caused by improper equipment scheduling and resource conflicts. After using the present invention, the monthly construction progress lag time was shortened to 0.5 days, greatly reducing construction delays and ensuring that the project can be completed on time or even ahead of schedule.

[0155] Scheduling errors caused by resource conflicts are a common problem in construction site management. Before the present invention was used, an average of 20 scheduling errors due to resource conflicts occurred in the project each week, but through the optimization and adjustment of the intelligent scheduling system, scheduling errors were reduced to 4 times per week. This data shows that game theory and Nash equilibrium mechanism can effectively solve the problem of resource competition and optimize resource allocation, thereby greatly reducing scheduling confusion caused by resource competition.

[0156] The improvement in the error correction time of digital-to-analog conversion is also quite significant. When the present invention is not used, errors in the construction process cannot be corrected in time, resulting in a large deviation between the actual construction and the digital model. When the belief propagation algorithm is used, the error correction time is shortened to less than 45 minutes, and the system can quickly correct the deviation after the deviation occurs, ensuring that the actual progress of the construction site is consistent with the digital twin model. In addition, the system detected 6 errors within 3 months, and the deviation rate of all errors after correction was controlled within 1%, which further proves the high efficiency of the present invention in real-time correction and progress control.

[0157] The accuracy of project progress control has also been significantly improved. Before the present invention was used, the control accuracy of the project progress fluctuated greatly, usually with a deviation of about ±5%. After the present invention was used, the control accuracy of the project progress was greatly improved, and the deviation was reduced to ±1%. This shows that the present invention can more accurately predict and control the construction progress and ensure that various tasks are completed on time.

[0158] In terms of resource scheduling optimization, before the present invention was used, scheduling was mostly manual, and it was difficult to make real-time optimization adjustments. After the present invention was used, the system performed 35 automatic optimizations in 3 months, and was able to adjust resource allocation in real time according to on-site conditions. This function not only reduced the burden on management personnel, but also improved the flexibility and responsiveness of the entire project.

[0159] In general, after using the present invention, the overall management efficiency of the project is greatly improved. In terms of resource scheduling, equipment utilization, construction progress control and error correction, the present invention has shown excellent performance, enabling the project to complete construction 5 days ahead of schedule, significantly improving the construction quality and economic benefits. Through the application of the present invention, the project management party can effectively solve the real-time integration of multi-source data in the construction site, the conflict optimization of equipment scheduling, and the error correction problem in digital-to-analog conversion, significantly improving the resource utilization and scheduling efficiency during the construction process, and ensuring that the construction progress of the site can be completed on time or even ahead of schedule. The present invention is not only applicable to large-scale complex construction sites, but also to other types of construction scenes, and has broad application prospects.

[0160] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A construction site digital-to-analog conversion and collaborative management method based on a distributed computing architecture, characterized in that: The steps include: S1. Collecting micro-level data of the construction site through a distributed sensor network, wherein the micro-level data includes equipment vibration, material changes, and temperature and humidity environmental parameters; S2. Using drones and 3D laser scanning technology to obtain meso-level data of the construction site, the meso-level data includes equipment operation status, local construction progress and environmental changes; S3. Obtain macro-level data from distributed building information models and geographic information systems, including overall construction progress, equipment distribution, and resource allocation, and integrate the macro-level data with meso-level data and micro-level data, pre-process using edge computing devices, and construct a multi-scale spatiotemporal data set; S4. Use the spatiotemporal graph convolutional network to decompose and process multi-scale spatiotemporal datasets in a distributed computing environment, analyze the spatiotemporal correlation between data at different levels, and build a dynamic digital twin model of the construction site; S5. Use intelligent agents to schedule key equipment or resources. Under the distributed computing architecture, use game theory models to design competition and cooperation mechanisms between intelligent agents, resolve resource allocation conflicts through Nash equilibrium, and dynamically optimize resource scheduling. S6, optimize the agent's scheduling strategy through a distributed algorithm based on proximal policy optimization, combined with historical scheduling data and real-time feedback; S7. Use the belief propagation algorithm to detect errors in the digital-to-analog conversion and scheduling process between distributed nodes and perform real-time correction. The correction information is quickly propagated in the distributed system to trigger the adjustment mechanism. S8. Through the distributed adaptive feedback control mechanism, the scheduling strategy of the intelligent agent is continuously optimized based on historical scheduling data, resource utilization and real-time feedback.

2. The construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture according to claim 1 is characterized in that: The S3 specifically includes: S31. Obtain macro-level data on overall construction progress, equipment distribution, and resource allocation from distributed building information models and geographic information systems, and standardize the macro-level data into spatiotemporal coordinate representation; S32, using a spatiotemporal alignment algorithm to process macro-level data, using a uniform time step and spatial resolution for alignment; S33, integrating the macro-level data with the meso-level data to obtain macro-meso-level fusion data: Among them, M 宏中融合 (x, y, t) represents macro-meso fusion data, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial location x, y and time t, M 中观 (x, y, z, t) represents the data distribution of meso-level data at spatial position x, y, height z and time t, λ1 and λ2 represent weight factors, tanh represents the hyperbolic tangent function, α represents the regulation coefficient, β, γ and δ represent the adjustment factors, and ∈ represents the interaction adjustment coefficient between macro-level data and meso-level data; S34. Integrate the macro-meso fusion data with the micro-level data to obtain multi-level data: Among them, M 融合 (x, y, z, t) represents the fused multi-level data, M 微观 (x, y, z, t) represents the data distribution of micro-level data at spatial position x, y, height z and time t, γ1 and γ2 represent weight factors, η, α1, α2 and ξ represent control factors; S35. Preprocessing the integrated multi-level data using edge computing devices, wherein the preprocessing includes data cleaning, noise reduction, and removal of redundant data; S36. Construct a multi-scale spatiotemporal dataset based on the preprocessed multi-level data.

3. The construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture according to claim 1 is characterized in that: The S4 specifically includes: S41. Construct a spatiotemporal graph structure G(V,E), where the node set V represents each key position of the construction site, the edge set E represents the spatiotemporal dependency between the key positions, the characteristics of each node include macro-level data, meso-level data and micro-level data, and the spatiotemporal association between nodes is defined by the adjacency matrix; S42, initialize the feature matrix of each node, the feature matrix is: X(t)=[M 宏观 (x,y,t),M 中观 (x,y,z,t),M 微观 (x,y,z,t)]; Among them, X(t) represents the node feature at time t, M 宏观 (x, y, t) represents the data distribution of macro-level data at spatial location x, y and time t, M 中观 ( x,y,z,t ) represents the data distribution of meso-level data at spatial position x, y, height z and time t, M 微观 ( x,y,z,t ) Represents the data distribution of micro-level data at spatial position x, y, height z and time t; S43, decomposing and processing the multi-scale spatiotemporal dataset of the construction site through a spatiotemporal graph convolutional network, wherein the spatiotemporal graph convolutional network calculates the spatiotemporal dependencies between nodes through an adjacency matrix to extract high-dimensional features of the multi-scale spatiotemporal dataset; S44, dynamically capturing the processed multi-scale spatiotemporal data set in the time dimension, using multi-layer temporal convolution operations to process feature changes in the time step, and forming a high-dimensional feature representation of spatiotemporal association; S45. Construct a dynamic digital twin model of the construction site based on high-dimensional feature representation of spatiotemporal associations: Among them, M 动态孪生 (x, y, z, t) represents the dynamic digital twin model of the construction site, M 时空 ( x,y,z,t ) represents the spatial features extracted by spatiotemporal convolution, M 时间 ( x,y,z,t ) represents the temporal features extracted by temporal convolution, and β1, β2, β3, τ, ζ and κ represent the control parameters.

4. The construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture according to claim 1 is characterized in that: The S5 specifically includes: S51, assigning an independent intelligent agent to each key equipment or resource in the construction site, each intelligent agent obtains the status information of the responsible equipment or resource through the real-time data acquisition system in the distributed computing architecture, the status information including the location, availability and current task load; S52. The competition and cooperation relationship between agents is modeled through game model to make scheduling decisions: Among them, U i represents the utility function of the ith agent, D i (t) represents the scheduling strategy of the ith agent, μ i represents the scheduling weight of the ith agent, f i (D i (t)) represents the scheduling strategy function of the ith agent, g ij (D i (t),D j (t)) represents the interaction utility between the ith agent and the jth agent, ν i , i , κ i , δ i and i represents the adjustment coefficient; S53. When there is resource competition between agents, the utility function U is calculated through the game theory model. i Evaluate the benefits of each agent and dynamically allocate resources through competition and cooperation mechanisms. Each agent is assigned a certain amount of resources based on the utility function U. i The optimal solution to adjust the scheduling strategy; S54. Resolve resource allocation conflicts between agents through Nash equilibrium and maximize the global utility of agents: Among them, D * represents the optimal scheduling strategy combination, N represents the total number of agents, represents the utility of the ith agent when the strategies of other agents remain unchanged, ρ i ,θ i , and represents the control parameters; S55. Under the distributed computing architecture, the intelligent agent shares the current status information, task requirements and resource availability in real time through the network, dynamically responds to the real-time changes of the construction site, and adjusts the resource scheduling strategy; S56. During the scheduling process, the intelligent agent continuously monitors the status information of the equipment and the progress of task execution, optimizes the scheduling strategy through feedback from the game theory model, and uses Nash equilibrium to achieve adaptive optimization of resource scheduling.

5. The construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture according to claim 1 is characterized in that: The S6 specifically includes: S61. Initialize the scheduling strategy π of each agent i , the agent collects historical scheduling data H through nodes in the distributed computing architecture i (t) and real-time feedback data F i (t), where H i (t) represents the historical scheduling strategy of the ith agent at time t, F i (t) represents the real-time feedback data of the i-th agent at time t; S62, using a distributed algorithm based on proximal policy optimization, the agent evaluates the current scheduling strategy π i The performance of the system is analyzed by combining historical scheduling data and real-time feedback data to optimize the strategy, and deviation analysis and value assessment are performed: Among them, V i (π i ) indicates that the i-th agent is in strategy π i The value evaluation function under , t represents the time step, T represents the total time step, Α represents the discount factor, R i (t) represents the immediate reward obtained by the i-th agent at time t, α v , β v , η v , v , δ v and v represents the adjustment parameter; S63, optimize the current scheduling strategy of the agent through the distributed algorithm of proximal strategy optimization: in, represents the optimized agent scheduling strategy, π i (a i |s i ) represents the state s of the i-th agent under the new strategy i Select action a i The probability of Represents the state s of the i-th agent under the old strategy i Select action a i The probability of N is the total number of agents, α π , β π , π , δ π and η π represents the adjustment parameter of the strategy update, U i (s i ,a i ) represents the utility function of the ith agent, U j (s j ,a j ) represents the utility function of the jth agent; S64. In the process of optimizing the strategy, the intelligent agents communicate and share data in real time through the distributed computing architecture. The intelligent agents transmit the current scheduling status and feedback information through the network, coordinate the scheduling behavior of other intelligent agents, and optimize the global allocation; S65, according to the optimized intelligent agent scheduling strategy The scheduling strategy is adjusted based on real-time feedback. Each agent re-evaluates equipment availability, task priority, and resource allocation requirements through updated strategies. S66. After completing the scheduling task, the intelligent agent continuously updates the scheduling strategy parameters through the accumulation of historical data and real-time feedback monitoring, and gradually approaches the global optimal solution through multiple scheduling strategy iterations.

6. The construction site digital-to-analog conversion and collaborative management method based on distributed computing architecture according to claim 1 is characterized in that: The S7 specifically includes: S71. Through the distributed sensor network and feedback system, the real-time data of each node is obtained, including the equipment status, task progress and resource allocation. The feedback data of each node is recorded as D i (t), where D i (t) represents the state data of node i at time t; S72, based on node feedback data D i (t) and the expected value P of the global digital twin model i (t), calculate the initial error of each node: Among them, E i (t) represents the error value of node i at time t, α e , β e , e and e represents the adjustment parameter; S73. Use the belief propagation algorithm to transmit error information between distributed nodes, and each node updates its own belief value according to the error value of the neighboring node: Among them, B i (t+1) represents the belief value of node i at time t+1, B i (t) represents the belief value of node i at time t, N ( i ) represents the set of neighbor nodes of node i, β ij represents the belief propagation coefficient between node i and node j, E j ( t ) represents the error value of node j at time t, α b , γ b , b and δ b represents the adjustment parameter, β ik represents the belief propagation coefficient between node i and node k; S74, according to the updated belief value B i (t+1), each node adjusts the scheduling parameters and the digital-to-analog conversion control signal, and the node reduces the feedback data D according to the result of belief propagation. i (t) and the expected value P of the global digital twin model i (t) Deviation between; S75. During the iterative process of belief propagation, each node continuously exchanges and updates error information with neighboring nodes until the error value of the global system converges and eventually reaches global optimization. S76. When the system detects that the error exceeds a predetermined threshold, a distributed error correction mechanism is triggered. The error correction mechanism adjusts the scheduling decision and parameter configuration of the node according to the result of belief propagation, and corrects the error in the digital-to-analog conversion in real time.

Images