A construction work scene adaptive speed matching signal visualization method

Abstract

The application discloses a kind of building with work scene adaptive speed matching signal visualization method, it is related to building construction management technical field, comprising: obtaining the process network model based on DAG of directed acyclic graph construction;Obtain labor demand and candidate worker real-time state data;When triggering speed matching, the matching degree is calculated by adaptive matching model, and is decoupled into multiple discrete decision factors, respectively quantifying its contribution value to score;In interface, the speed matching signal set connecting worker and task node is generated, and the signal set is formed by independent sub-signal parallel aggregation that visual form parameter is positively correlated with corresponding factor contribution value;When dispatching is locked or recommended, delay risk propagation calculation is carried out along link downstream along process network.The present application improves the decision explainability and man-machine trust degree, effectively avoids the global process delay conflict caused by local dispatching.

Description

Technical Field

[0001] This invention relates to the field of construction management technology, and more specifically to a method for visualizing adaptive quick-match signals in construction labor scenarios. Background Technology

[0002] In the construction process, labor scheduling is a key factor affecting project progress efficiency and resource utilization. Due to the highly dynamic construction site environment, workers' locations, skill levels, fatigue levels, and the urgency of each task are constantly changing, making traditional static scheduling difficult to meet management needs.

[0003] Existing intelligent work assignment systems are typically based on Building Information Modeling (BIM) combined with positioning technology, using adaptive matching algorithms to recommend suitable workers for the current task. These systems usually calculate a comprehensive matching score based on multiple dimensions such as distance and skill tags, and display it to dispatchers in a sorted list or simple icon-based format.

[0004] However, existing technologies have the following limitations in practical applications:

[0005] First, the decision-making logic lacks interpretability. Deep learning or complex adaptive matching models are often black boxes, outputting only a single recommendation or score. Construction scheduling relies heavily on on-site experience. When dispatchers cannot intuitively obtain the specific basis for the system's recommendations, they often distrust the assignment results, leading to a low actual adoption rate of the system.

[0006] Secondly, there is a lack of dynamic representation of global process constraints. Construction tasks do not exist in isolation, but are subject to strict time delays and pre-process requirements. Existing systems are usually limited to local matching of single tasks and cannot intuitively show the chain of delays that the work assignment decision may cause to the subsequent process network at the moment the decision is made. This can easily lead to management errors that result in local optima at the expense of the overall system.

[0007] Third, the interaction mode is one-way and lacks flexibility. The existing system's visual interface mainly serves as the output end. When unexpected situations on-site cause the model's preset weights to become inapplicable, the dispatcher finds it difficult to directly intervene in the underlying decision factor configuration through the intuitive graphical interface, resulting in low human-machine collaboration efficiency.

[0008] Therefore, it is necessary to propose an adaptive speed-matching signal visualization method for construction labor scenarios to solve the above problems. Summary of the Invention

[0009] The purpose of this invention is to solve the problems of low adoption rate, lack of global process delay prediction, and limited human-computer interaction methods in existing intelligent dispatching systems due to black-box decision-making.

[0010] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0011] An adaptive speed-matching signal visualization method for construction worker scenarios includes the following steps:

[0012] Obtain the process network model of the construction project. The process network model is constructed based on a directed acyclic graph and includes construction task nodes and directed edges representing the dependencies between tasks.

[0013] Acquire labor demand information and real-time status data of candidate workers. The real-time status data includes at least the worker's skill characteristics, spatial location information, and current fatigue status. When a quick match for the labor demand is triggered, calculate the total matching score of the candidate workers through an adaptive matching model. Use an attribution algorithm to decouple the calculation process into multiple discrete decision factors and quantify the contribution of each decision factor to the total matching score.

[0014] In the visualization interface, a quick-match signal set connecting candidate worker nodes and target task nodes is generated; the quick-match signal set is aggregated from independent sub-signals corresponding to each decision factor, wherein the visual morphological parameters of each independent sub-signal are positively correlated with the contribution value of the corresponding decision factor;

[0015] When the dispatch status of a task node is locked or the adaptive matching model outputs a recommendation result, the delay risk propagation calculation is performed along the dependency link to the downstream task node based on the process network model.

[0016] In the visualization interface, a process ripple signal is generated at the affected downstream task node, and the diffusion parameters of the process ripple signal are associated with the calculated expected delay duration and the total float of the downstream task node.

[0017] Furthermore, the decision factors include: skill matching factor, spatial accessibility factor, risk factor for delays in subsequent tasks, worker fatigue recovery factor, and historical quality factor.

[0018] Furthermore, the visual morphological parameters include line segment width, color threshold, and / or texture display status; multiple independent sub-signals are spatially aggregated into a tubular set of speed-dating signals.

[0019] Furthermore, the method also includes a two-way interactive feedback step: receiving a touch deformation gesture applied by the user on the visualization interface for a specific independent sub-signal; obtaining the displacement change of the touch deformation gesture, mapping the displacement change to a weight adjustment parameter through a preset nonlinear feedback function; updating the input weights of the corresponding decision factors in the adaptive matching model according to the weight adjustment parameter, recalculating the total matching score and contribution value, and redrawing the speed-matching signal set in real time.

[0020] Furthermore, the worker's skill characteristics are associated with a credibility decay coefficient based on time or quality records; when the credibility decay coefficient decreases, the contribution value of the corresponding skill matching factor decreases synchronously, and in the visualization interface, edge superimposed graphic noise representing the independent sub-signal of the skill matching factor is controlled to present a blurred or jittery visual effect.

[0021] Furthermore, the diffusion parameters of the process ripple signal are calculated by ratio. Confirmed, among which To propagate to downstream task nodes The estimated delay time, This represents the total float of the node in the process network model; when When it is greater than 0, ;when equal to 0 and If it is greater than 0, it is determined directly. Exceeding a preset extreme value threshold; when the ratio is detected. When the preset extreme value threshold is exceeded, the visual characteristic step of the process ripple signal is triggered, and the control signal coloring changes abruptly from the warning color to a bright alarm color with directional flow characteristics.

[0022] Furthermore, the preset nonlinear feedback function is a mapping model based on the hyperbolic tangent function: ;in, For weight adjustment parameters, The displacement change is... The preset weight adjustment sensitivity constant, This is the maximum allowed drag pixel threshold for the interface.

[0023] The beneficial effects of this invention are as follows:

[0024] 1. This invention introduces an attribution algorithm to deconstruct the single matching score black box into multi-dimensional visual sub-signals, enabling schedulers to intuitively obtain the specific reasons for machine decisions, establish human-machine trust, and improve the actual adoption rate of intelligent recommendations.

[0025] 2. This invention transforms hidden time constraint conflicts into visualized process ripple signals. Before confirming work assignments, schedulers can perceive the impact of the decision on the global critical path through these ripple signals, effectively avoiding global project delays caused by local suboptimal approaches.

[0026] 3. This invention supports direct touch operation on caustic signals to provide feedback for modifying the weights of the underlying model, breaking through the parameter form modification mode and realizing dynamic fine-tuning of model parameters based on graphical intuition. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method flow provided in an embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of the multi-factor decoupling and aggregation rendering effect of the speed-matching signal set.

[0029] Figure 3 This is a schematic diagram of process ripple signal propagation based on a process network model. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0031] This example uses the main construction phase of a large commercial complex project. The project dispatch center is equipped with a large visual screen and an explosion-proof tablet terminal that supports multi-touch. On-site, real-time monitoring of personnel location and work status is achieved through an IoT system, UWB positioning base stations, and smart safety helmets.

[0032] like Figure 1 As shown, the method provided in this embodiment mainly follows core steps such as data acquisition and network construction, triggering speed matching and factor decoupling, generating a visual signal set, and predicting the propagation of global delay risks.

[0033] First, data preparation and process network construction are performed. The system's data acquisition module extracts the current construction task list from the project management database and constructs a process network model based on a directed acyclic graph (DAG). Task A (reinforcement binding of the core tube on the 3rd floor) is defined in the model. Task B (Establishment of core tube formwork on the 3rd floor) The critical task Z (welding of high-difficulty load-bearing columns) and other processes have strict dependencies. Using the Critical Path Method (CPM), the system calculates the total float of each task node. The system continuously acquires the spatial coordinates of candidate workers, their continuous working hours for the day to calculate fatigue status, skill tags, and historical completion quality records.

[0034] When a sudden delay occurs on-site, the dispatcher selects task A on the interface and issues an instruction to urgently add one senior steelworker, triggering the system's adaptive quick-match logic. The cloud-based adaptive matching model calculates and assigns worker A a high matching score of 0.88. At this point, the decoupling and propagation calculation module calls the Shapley value attribution algorithm to transparently decouple the single black-box output of 0.88. The algorithm calculates the following... Figure 2 The discrete decision factor set on the left shows the contribution values ​​of the five discrete decision factors: f1 (skill matching factor) contributes 0.35; f2 (spatial accessibility factor) contributes 0.25; f3 (risk of delay for subsequent tasks) contributes 0.20; f4 (worker fatigue recovery factor) contributes 0.08; and f5 (historical quality factor) contributes 0.00. The sum of the five factor contributions gives a perfect match score of 0.88.

[0035] Then, signal bundle rendering and credibility degradation display are performed. For example... Figure 2 As shown in the process network visualization diagram and magnifying glass area on the right, the visualization rendering module generates a bundle-shaped quick-match signal set on the terminal interface, which is composed of the above 5 independent sub-signals aggregated in parallel, connecting the nodes of worker A and task A. The sub-signals representing each factor use different colors and textures, and their line segment widths are strictly positively correlated with their contribution values. The line width representing the skill factor (f1) is 35 pixels, while the line width representing the fatigue factor (f4) is only 8 pixels, and the line width representing the quality factor (f5) is 0 (invisible). At this time, the system detects that although worker A holds an advanced certificate, he has not performed this task for two months, triggering the skill credibility decay calculation. Its credibility decay coefficient is... The variance decreased from 1.0 to 0.65. The system reduces this by introducing variance into the rendering shader of the skill sub-signal (f1). Positively correlated Gaussian graphic noise, such as Figure 2 As shown in the illustration of edge jitter / blurring in the lower right corner, the thickest sub-signal edge exhibits significant jagged jitter and blurring. The dispatcher can instantly understand the recommendation reason by visually observing the line thickness and edge texture: Worker A is close by and has the appropriate skills on paper, but is extremely fatigued (f4 extremely thin), and the authenticity of their skill proficiency is questionable (f1 edge jitter).

[0036] In the global constraint verification and process ripple alarm phase, the dispatcher attempts to assign worker B, who possesses special welding qualifications, to a current non-critical cleanup task on the interface. After the work assignment status is pre-locked, combined with... Figure 3The delay risk propagation calculation logic shown in the upper part involves the system immediately performing delay risk propagation calculations downstream along the DAG process network. Specifically, the system extracts the estimated completion delay time of the current task, combines it with the standard operation time of each node in the process network model to perform forward deduction of the critical path, and calculates the difference between the earliest start time of the affected downstream node and the original earliest start time, which is used as the estimated delay duration of the downstream node. Calculations revealed that worker B's scheduled participation in the critical task Z—the welding of a high-difficulty load-bearing column—in the afternoon would face severe delays due to a manpower shortage. The system's forward calculations yielded the estimated delay duration. Since task Z is on the critical path, its total float in the CPM model is... According to the system's delay impact rules, if Then calculate the ratio. Because at this time and The system directly determines the current impact of the delay. Exceeding the system's set extreme threshold (e.g., the normally preset critical threshold) This effectively avoids system logic crashes caused by a zero denominator in conventional ratio calculations. The system detects that the extreme value threshold has been breached, such as... Figure 3 The lower half of the network ripple propagation diagram shows that the visualization interface immediately generates a rapidly expanding outward process ripple signal at the critical task Z node, and triggers a visual feature step: the coloring of the ripple signal is controlled by a change from the usual semi-transparent warning color ( Figure 3 The orange warning ripples in the image instantly changed to a brighter warning color with a strong alert effect. Figure 3 A bright red alarm ripple appeared in the lower right corner, accompanied by a pulse flow effect pointing to downstream nodes. The dispatcher detected the red visual step alarm and immediately realized that the current local dispatch would cause the paralysis of critical nodes globally in the afternoon, so he proactively canceled the operation.

[0037] In the reverse gesture interaction and weight reconstruction stage, regarding the work assignment decision for worker A, due to the high temperature that day, the dispatcher, based on on-site experience, believed that the fatigue recovery factor (f4) should be forcibly placed first. The dispatcher directly made an outward stretching gesture on the touch screen. The image was stretched outward by 150 pixels, and then substituted into a preset hyperbolic tangent nonlinear feedback function to calculate the weight adjustment parameters of the fatigue factor. .in, This is a preset weight adjustment sensitivity constant used to control the maximum impact of a single drag on the weight. The value is 0.8; The maximum effective drag pixel threshold set for the system interface is used to normalize the input displacement. Based on the above-determined function parameter constraints, the system then safely and smoothly increases the input weight of the fatigue factor (f4) in the underlying adaptive matching model and triggers a recalculation. After the recalculation, the signal set of the interface is instantly redrawn: worker A, who was originally ranked first but was very fatigued, has a total score that drops to 0.86; while another candidate worker C, who is resting well in the shade but is a little far away, has a fatigue factor (f4) sub-signal line that rapidly thickens and expands, and his total score jumps to 0.91, with his overall signal bundle ranking jumping to the top of the recommendation list. The dispatcher then confirms the assignment of worker C in the optimal state, completing the dynamic fine-tuning of model parameters based on graphical intuition and the closed loop of human-machine collaboration.

[0038] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the content of the present invention's specification shall also be included within the scope of protection of the present invention.

Claims

1. A method for visualizing adaptive speed-matching signals in construction worker scenarios, characterized in that, Includes the following steps: Obtain the process network model of the construction project. The process network model is constructed based on a directed acyclic graph and includes construction task nodes and directed edges representing the dependencies between tasks. Acquire labor demand information and real-time status data of candidate workers. The real-time status data includes at least the worker's skill characteristics, spatial location information, and current fatigue status. When a quick match for the labor demand is triggered, the total matching score of the candidate workers is calculated through an adaptive matching model. The calculation process is decoupled into multiple discrete decision factors using an attribution algorithm, and the contribution of each decision factor to the total matching score is quantified. In the visualization interface, a quick-match signal set connecting candidate worker nodes and target task nodes is generated; the quick-match signal set is aggregated from independent sub-signals corresponding to each decision factor, wherein the visual morphological parameters of each independent sub-signal are positively correlated with the contribution value of the corresponding decision factor; When the dispatch status of a task node is locked or the adaptive matching model outputs a recommendation result, the delay risk propagation calculation is performed along the dependency link to the downstream task node based on the process network model. In the visualization interface, a process ripple signal is generated at the affected downstream task node, and the diffusion parameters of the process ripple signal are associated with the calculated expected delay duration and the total float of the downstream task node.

2. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 1, characterized in that, The decision factors include: skill matching factor, spatial accessibility factor, risk factor for delays in subsequent tasks, worker fatigue recovery factor, and historical quality factor.

3. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 1, characterized in that, The visual morphological parameters include line segment width, color threshold, and / or texture display status; multiple independent sub-signals are spatially aggregated into a tubular set of speed-dating signals.

4. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 1, characterized in that, The method further includes a two-way interactive feedback step: receiving a touch deformation gesture applied by the user on the visualization interface for a specific independent sub-signal; obtaining the displacement change of the touch deformation gesture, mapping the displacement change to a weight adjustment parameter through a preset nonlinear feedback function; updating the input weights of the corresponding decision factors in the adaptive matching model according to the weight adjustment parameter, recalculating the total matching score and contribution value, and redrawing the speed-matching signal set in real time.

5. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 2, characterized in that, The worker's skill characteristics are associated with a reliability decay coefficient based on time or quality records; when the reliability decay coefficient decreases, the contribution value of the corresponding skill matching factor decreases synchronously; and in the visualization interface, edge superimposed graphic noise representing the independent sub-signals of the skill matching factor is controlled to present a blurred or jittery visual effect.

6. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 1, characterized in that, The diffusion parameters of the ripple signal in the process are calculated by ratio. Confirmed, among which To propagate to downstream task nodes The estimated delay time, This represents the total float of the node in the process network model; when When greater than 0, ;when equal to 0 and If it is greater than 0, it is directly judged. Exceeding a preset extreme value threshold; when the ratio is detected. When the preset extreme value threshold is exceeded, the visual characteristic step of the process ripple signal is triggered, and the control signal coloring changes abruptly from the warning color to a bright alarm color with directional flow characteristics.

7. The adaptive speed-matching signal visualization method for construction worker scenarios according to claim 4, characterized in that, The preset nonlinear feedback function is a mapping model based on the hyperbolic tangent function: ;in, For weight adjustment parameters, The displacement change is... The preset weight adjustment sensitivity constant, This is the maximum allowed drag pixel threshold for the interface.

Images