Intelligent factory logistics path dynamic optimization method based on digital twin space-time trajectory clustering

Abstract

The invention discloses an intelligent factory logistics path dynamic optimization method based on digital twin spatial-temporal trajectory clustering. According to the method, trajectory decoupling is carried out by introducing a graph neural network and a topological data analysis technology, inherent path preference and instantaneous space coupling components in the trajectory are effectively separated, the problem of pseudo cluster generation caused by trajectory space coupling is fundamentally solved, and more reasonable path planning can be provided for logistics tasks. The new task AGV is no longer compulsively compiled into the existing queue, the standby channel can be started according to the actual situation, the operation flexibility of the logistics equipment is improved, the optimized logistics path dynamic optimization method enables the efficiency of the logistics system to be remarkably improved, the operation efficiency of the logistics equipment is improved, and the operation efficiency of the logistics equipment is improved. The waiting time and the congestion condition of goods in the transportation process are reduced, logistics tasks can be completed more quickly and accurately, and powerful guarantee is provided for efficient production of intelligent factories.

Description

Technical Field

[0001] The present invention belongs to the field of logistics technology, and specifically provides a dynamic optimization method for smart factory logistics paths based on digital twin spatiotemporal trajectory clustering. Background Art

[0002] With the rapid development of intelligent manufacturing, smart factories have become a key area of ​​transformation and upgrading for the manufacturing industry. In smart factories, efficient logistics systems are key factors in ensuring production continuity, improving efficiency, and reducing costs. Logistics path optimization, as a core component of the logistics system, directly impacts the operational efficiency of the entire smart factory. Current general approaches to logistics path optimization include: Using a distributed sensor network to collect real-time location, velocity, and environmental data of logistics entities, these data are synchronously mapped to a digital twin platform at preset intervals to generate a dynamic dataset. Based on the completion of logistics tasks or time window triggers, structured spatiotemporal trajectories are reconstructed and trajectory vectors containing spatial, temporal, and dynamic interaction features are extracted. A spatiotemporal density clustering algorithm is used to incrementally analyze trajectory vectors, identifying high-frequency and anomalous path patterns and constructing a path pattern library containing core node sequences, spatiotemporal constraints, and confidence levels. In response to new tasks or pattern library updates, the task requirements are matched with the pattern library, and the optimal path is solved based on a spatiotemporal cost function and distributed to logistics equipment. Finally, based on the deviation between the actual and planned paths, incremental updates to the pattern library and cost function parameter adjustments are triggered.

[0003] However, existing methods only focus on indicators such as inflection point density and area coverage when extracting trajectory spatial features, and are unable to distinguish the trajectory differences between active path selection and passive spatial coupling of logistics entities. In areas with high spatial constraints (such as narrow passages), multiple AGVs form fixed queues due to spatial limitations. The strong overlap of trajectories causes traditional clustering algorithms to misjudge them as high-confidence pseudo-clusters, which not only generates inefficient "car-following" paths that limit equipment flexibility, but also pollutes the path pattern library with high-confidence pseudo-patterns, ultimately leading to the failure of dynamic optimization and a decrease in factory throughput. Summary of the Invention

[0004] The purpose of the present invention is to provide a dynamic optimization method for smart factory logistics paths based on digital twin spatiotemporal trajectory clustering in order to solve the above-mentioned problems.

[0005] The technical solution adopted by the present invention is as follows: a method for dynamic optimization of logistics paths in smart factories based on spatiotemporal trajectory clustering of digital twins, the method comprising the following steps: S1: Data synchronization is performed first: the distributed sensor network collects the location, speed and environmental data of the logistics entity at a preset period, and maps it to the digital twin platform in real time to generate a dynamic twin dataset; S2: Trajectory reconstruction: Responding to the logistics task completion event or time window triggering condition, reconstructing the structured spatiotemporal trajectory of the entity, and extracting the trajectory feature vector containing spatial distribution features, dynamic interaction features, and correlation features; S3: Decouple the trajectory: In response to the trajectory feature vector generation event in step S2, construct the factory space topology graph G = (V, E), where the vertex V is the key node of the path and the edge E is assigned attributes. ; Calculate the spatial freedom factor based on the channel width ; Extract the spatial constraint component of the trajectory through the graph neural network encoder, press Separate the coupled components to generate decoupled trajectories; S4: Perform cluster analysis: Input the decoupled trajectory into the improved spatiotemporal density clustering algorithm, and use the composite density threshold and adaptive threshold Identify high-frequency path patterns and build a path pattern library containing core node sequences, spatiotemporal constraints, and confidence levels; S5: Path optimization: respond to new logistics tasks or pattern library update events, match task requirements with the path pattern library, solve the optimal path based on the time-space cost function, and send it to the logistics equipment; S6: Perform the final dynamic adjustment step: according to the deviation entropy between the actual path and the planned path Trigger incremental update of path pattern library and adjustment of cost function parameters.

[0006] In a preferred embodiment, in step S2, the extraction of trajectory feature vectors includes three types of information: spatial distribution features, dynamic interaction features, and correlation features, which respectively characterize the behavior patterns of logistics entities in the spatiotemporal process from different dimensions.

[0007] The spatial distribution characteristics analyze the trajectory's stay behavior in key bottleneck areas within the factory and quantify its time proportion within a specified spatial range to reflect potential operational congestion and path efficiency issues. At the same time, the inflection point density index is formed by calculating the ratio of the number of inflection points in the trajectory to the overall path length. The higher the index, the more frequent the path changes direction, the greater the potential mechanical wear and energy consumption, which helps to reduce complex motion trajectories during the path optimization process.

[0008] The dynamic interaction feature uses a neural network model based on time series input to predict the speed of a trajectory over a continuous period of time, output its speed change trend, and generate a trajectory stability score based on this prediction. A high score indicates a stable trajectory with minimal environmental interference, while a low score indicates that the trajectory is frequently disturbed by dynamic events. By training the relationship between historical trajectories and their speed changes, the model adaptively extracts potential influencing factors, improving its ability to express the stability of dynamic behavior.

[0009] The association feature focuses on the spatial relationship between a trajectory and other surrounding logistics entities. Within the time window of a trajectory point, if the minimum Euclidean distance to another entity falls below a safe distance threshold, the system records the interaction event and associates the corresponding entity identifiers, constructing an interaction map between trajectories to identify potential conflicts, collaborations, or resource competition. Combining these three types of features creates a multidimensional trajectory feature vector, which serves as the basic input for cluster analysis, improving the accuracy of depicting the complexity of logistics behavior and the spatial organization patterns.

[0010] In a preferred embodiment, in step S3, trajectory decoupling removes the trajectory constraint components caused by factory structure restrictions by introducing a spatial topology perception mechanism, thereby more accurately extracting the autonomous navigation behavior of logistics entities. First, in the spatial topology graph construction stage, the internal structure of the factory is abstracted into a graph G = (V, E), where the vertex V represents the intersection node and the work station, the edge E represents the logistics channel, and each edge is bound to three types of attributes: length l length , channel width w width With the maximum capacity c capacity , constructing a navigation graph structure with physical constraints. Secondly, in the degree of freedom factor calculation stage, in order to quantify the degree of restriction of each channel space on the trajectory, the spatial degree of freedom factor is defined: ; Among them, w width is the current channel width, w min is the minimum reference value of the channel width in the system. The output value range of this function is (0,1). When the channel width is close to the minimum value, Approaching 0 indicates low channel freedom and strong constraints; when the channel width is significantly greater than the minimum value, When it approaches 1, it means that the path segment has high degrees of freedom and the trajectory variability is large.

[0011] In a preferred embodiment, in step S4, the improved spatiotemporal density clustering algorithm enhances the ability to identify high-frequency logistics paths and abnormal trajectories by introducing a composite density threshold and an adaptive trajectory number threshold mechanism. First, define the composite density threshold: , Among them, εs pace Represents the basic spatial distance threshold, which controls the clustering range of trajectory points in the spatial dimension; W zone is the spatial weight factor, which is dynamically adjusted according to the real-time congestion index of the current area. The more congested the area, the higher the W zone tends to increase, thereby increasing the local spatial density threshold, reflecting the impact of the environment on clustering sensitivity; ε time It represents the time distance threshold, which limits the proximity of trajectory points in the time dimension and is used to ensure the temporal consistency of clustering.

[0012] Then, set the adaptive minimum number of tracks threshold: , Among them, N total is the total number of trajectories currently being processed, and MinTrajs controls the minimum number of trajectories required to be included in each cluster. This strategy uses a fixed lower limit when the amount of trajectory data is small to prevent clustering instability caused by too few samples, and raises the threshold to enhance clustering confidence when the amount of data is large.

[0013] Noisy trajectories identified during the initial clustering process undergo a secondary analysis. The system continuously monitors the temporal and spatial distances between these trajectories and the cluster center. If a trajectory consistently remains within a preset distance interval for multiple consecutive periods, it indicates a potential pattern in its behavior. Based on this information, the system constructs a temporary path pattern and incorporates it into the path pattern library. This mechanism strengthens the identification of atypical repetitive behaviors, improving the integrity of the logistics path pattern library and the system's adaptive optimization capabilities.

[0014] In a preferred embodiment, in step S4, during the coupling component stripping phase, the original trajectory point sequence Ti and the topology graph G are simultaneously input into the graph neural network encoder, and a three-layer graph convolution operation is used to extract the spatial deformation features induced by the structure in the trajectory. The system dynamically adjusts the decoupling strength λ to: The lower the degree of freedom, The smaller , the larger λ is, thus enhancing the ability of the decoupling operation to correct the impact of topological constraints. Finally, trajectory decoupling is completed through the following formula: Among them, GNN Encoder (Ti,G) represents the spatial constraint component extracted by the graph neural network encoder. This operation removes path features caused by physical structure from the original trajectory, retaining the active navigation intention of the logistics entity. This makes the generated decoupled trajectory more versatile and recognizable in terms of structural diversity, helping subsequent cluster analysis to accurately identify high-frequency path patterns with similar characteristics.

[0015] In a preferred embodiment, in step S4, cluster analysis further improves the accuracy and adaptability of path pattern recognition by introducing a factor that decouples trajectory and spatial freedom. First, in terms of distance measurement, in order to more effectively measure the spatial and structural similarity between trajectories, an improved composite distance measurement formula is used: ; in, represents the Euclidean distance between the decoupled trajectories of trajectory i and j, which is used to reflect the geometric differences in trajectory morphology. and are the average spatial degrees of freedom factors of the path segments traversed by the trajectories, which are used to measure the difference in the degree of physical structural constraints between two trajectories. μ is the structural weight factor, which is used to adjust the proportion of the influence of structural differences on the overall distance. This distance metric combines behavioral path similarity with spatial structural similarity, making clustering more sensitive to actual path patterns that recur in similar structures.

[0016] In a preferred embodiment, in step S4, in density clustering, the spatial weight factor in the composite density threshold is no longer based on the static congestion index, but directly adopts the spatial freedom factor calculated in the decoupling stage. ,Right now ,in The larger the value, the higher the spatial tolerance of trajectory distribution in the clustering process. space With ε time is the basic distance threshold in the spatial and temporal dimensions. This setting reflects that trajectories are allowed to have higher morphological diversity in areas with small spatial constraints.

[0017] In a preferred embodiment, in step S4, trajectories that are not classified into any cluster in the clustering results are regarded as noise trajectories, and the system performs a secondary analysis on such trajectories. The method is to calculate the distance distribution between them and the decoupled trajectories of each identified cluster center. If the distance between a noise trajectory and a cluster center continues to fall outside the preset interval in multiple time periods, and the deviation trend remains continuous within a certain period, then it is considered that the trajectory may represent a new path behavior intention. After the continuous overlimit condition is met, the system generates a temporary path pattern for the trajectory and adds it to the pattern library, enhancing the sensitivity of path identification to the evolution of heterogeneous paths and providing a basis for subsequent pattern library updates. This mechanism improves the ability to respond to edge behaviors or early abnormal trends, ensuring that the path pattern library can dynamically adapt to micro-changes in the logistics environment.

[0018] In a preferred embodiment, in step S5, the core of the route optimization is a spatiotemporal cost function constructed based on multi-dimensional factors to quantify the comprehensive cost of the candidate route in terms of transportation efficiency, safety risk, and time deviation. The specific cost function is in the form of: , Among them, T trave l is the estimated travel time of the route, which measures the cost of the route in terms of transportation efficiency; d t It represents the spatial distance between the logistics entity and the nearest obstacle or other mobile unit at time t, d safe is the safety distance threshold, k is the growth coefficient of the distance penalty, which makes the risk term increase exponentially when approaching the safety distance. The second term comprehensively reflects the level of safety hazards that may occur in the entire path; the last term |T actual -T expected∣ represents the degree of deviation between the actual completion time of the task and the expected time, and evaluates the time accuracy of the task.

[0019] The three costs are dynamically controlled by weight coefficients α, β, and γ, where α represents the priority of efficiency, β represents the sensitivity to safety distance, and γ controls the weight of the time deviation penalty.

[0020] During peak operating hours, the system increases β to 2.5 times the baseline value to enhance obstacle and congestion avoidance capabilities. During emergency missions, γ is increased to 3 times the baseline value to prioritize time-based meeting requirements. The adjustment strategy for these weight coefficients is iteratively updated in step S6 based on real-time feedback from path execution deviation data. This allows the optimization objective to dynamically adapt to different business scenarios and operating environments, improving the accuracy and robustness of path allocation. Finally, based on this cost function, the system searches for the path with the lowest total cost within the candidate path set and issues it as an execution task to the logistics equipment, ensuring that the path planning results strike a balance between time efficiency, safety, and timeliness.

[0021] In a preferred embodiment, in step S6, the deviation entropy processing measures the randomness and uncertainty of the path execution deviation to ensure the stability and adaptability of the path optimization system. First, the entropy index of the path deviation is calculated: ; Among them, p i represents the probability distribution within the i-th type deviation interval, E dev Measures the degree of discreteness and unpredictability of path execution deviations. When this entropy value exceeds a set threshold, it indicates that the path deviation has become highly unstable and difficult to predict using the existing pattern library. At this point, the system freezes the current path pattern library and stops adding new patterns to prevent erroneous information contamination. Subsequently, the system activates a reinforcement learning module to autonomously generate an emergency path planning strategy based on real-time environmental feedback and historical experience. This module interacts with the sensor network, issuing calibration instructions to the sensors and adjusting the sensor data acquisition accuracy and transmission frequency in real time to improve the accuracy and response speed of environmental perception. If the deviation persists and the difference between the decoupled trajectory and the planned path continues to exceed the tolerance range, the system triggers parameter retraining of the graph neural network encoder, using newly collected trajectory data and spatial topology information to update the model weights, thereby optimizing the encoding of spatial constraints and path behavior. This mechanism ensures that the model can adapt to environmental changes and path behavior evolution, achieving closed-loop adaptive optimization of path planning and execution.

[0022] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. In this invention, by introducing graph neural networks and topological data analysis technology, an innovative "trajectory decoupling" step is proposed: first, a factory space topology map is constructed, and the "spatial freedom factor" reflecting the degree of channel space restriction is calculated; then, the spatial constraint component of the trajectory is extracted through the graph neural network encoder, and combined with the dynamically adjusted decoupling strength, the "instantaneous spatial coupling" component caused by physical structure restrictions is stripped away, retaining the "inherent path preference" of the logistics entity. By introducing graph neural networks and topological data analysis technology for trajectory decoupling, the "inherent path preference" and "instantaneous spatial coupling" components in the trajectory are effectively separated, fundamentally solving the problem of pseudo-cluster generation caused by trajectory spatial coupling. After actual testing, in areas with high spatial constraints, the pseudo-cluster generation rate was reduced from more than 30% to less than 5%, greatly improving the accuracy of path pattern recognition.

[0023] 2. This invention fundamentally improves the rationality of path planning by improving the spatiotemporal density clustering algorithm and the path planning method based on an accurate pattern library. First, the improved clustering algorithm introduces a "composite density threshold" and an "adaptive trajectory number threshold" mechanism: the composite density threshold is dynamically adjusted in conjunction with the spatial degree of freedom factor, while the adaptive trajectory number threshold changes dynamically based on the amount of trajectory data, thereby more accurately identifying high-frequency path patterns. Second, the path optimization phase solves the optimal path based on a multidimensional spatiotemporal cost function, where the weight coefficients of the cost function can be dynamically adjusted based on the business scenario. This eliminates the need for AGVs assigned to new tasks to be forced into existing queues, enabling the use of alternative channels based on real-time environmental conditions and task requirements. Actual application data shows that the average travel distance of logistics equipment has been shortened by 10% to 15%, and transportation time has been reduced by 15% to 20%, significantly enhancing the flexibility of equipment operation and the rationality of path planning.

[0024] 3. In the present invention, by eliminating pseudo-clusters and improving the rationality of path planning, the operating efficiency of the logistics system is fundamentally improved. On the one hand, the congestion and waiting time of logistics equipment caused by unreasonable path planning are greatly reduced, and the timeliness and accuracy of cargo transportation are significantly improved; on the other hand, the dynamic adjustment step monitors the deviation between the actual path and the planned path through the "deviation entropy" indicator, triggering the incremental update of the path pattern library and the adjustment of the cost function parameters, forming a closed-loop optimization mechanism of "data collection-trajectory decoupling-clustering analysis-path optimization-deviation feedback", ensuring that the system can continuously adapt to environmental changes. In actual applications, the factory throughput has increased by 15% to 25% compared with traditional methods, and the operating efficiency of logistics equipment has been significantly improved, effectively ensuring the continuity of smart factory production and providing strong support for the efficient operation of the manufacturing industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is the logic diagram of the improved spatiotemporal density clustering algorithm of the present invention; Figure 2 This is a schematic diagram of the trajectory decoupling process in the present invention; Figure 3 This is a schematic diagram of constructing the spatial topology map in the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0027] Example: Reference Figure 1-3 , a dynamic optimization method for smart factory logistics paths based on digital twin spatiotemporal trajectory clustering, the method includes the following steps: S1: Data synchronization is performed first: the distributed sensor network collects the location, speed and environmental data of the logistics entity at a preset period, and maps it to the digital twin platform in real time to generate a dynamic twin dataset; S2: Trajectory reconstruction: Responding to the logistics task completion event or time window triggering condition, reconstructing the structured spatiotemporal trajectory of the entity, and extracting the trajectory feature vector containing spatial distribution features, dynamic interaction features, and correlation features; S3: Decouple the trajectory: In response to the trajectory feature vector generation event in step S2, construct the factory space topology graph G = (V, E), where the vertex V is the key node of the path and the edge E is assigned attributes. ; Calculate the spatial freedom factor based on the channel width ; Extract the spatial constraint component of the trajectory through the graph neural network encoder, press Separate the coupled components to generate decoupled trajectories; S4: Perform cluster analysis: Input the decoupled trajectory into the improved spatiotemporal density clustering algorithm, and use the composite density threshold and adaptive threshold Identify high-frequency path patterns and build a path pattern library containing core node sequences, spatiotemporal constraints, and confidence levels; S5: Path optimization: respond to new logistics tasks or pattern library update events, match task requirements with the path pattern library, solve the optimal path based on the time-space cost function, and send it to the logistics equipment; S6: Perform the final dynamic adjustment step: according to the deviation entropy between the actual path and the planned path Trigger incremental update of path pattern library and adjustment of cost function parameters.

[0028] In step S2, the extraction of trajectory feature vectors includes three types of information: spatial distribution features, dynamic interaction features, and correlation features, which respectively characterize the behavior patterns of logistics entities in the spatiotemporal process from different dimensions.

[0029] The spatial distribution characteristics analyze the trajectory's stay behavior in key bottleneck areas within the factory and quantify its time proportion within a specified spatial range to reflect potential operational congestion and path efficiency issues. At the same time, the inflection point density index is formed by calculating the ratio of the number of inflection points in the trajectory to the overall path length. The higher the index, the more frequent the path changes direction, the greater the potential mechanical wear and energy consumption, which helps to reduce complex motion trajectories during the path optimization process.

[0030] The dynamic interaction feature uses a neural network model based on time series input to predict the speed of a trajectory over a continuous period of time, output its speed change trend, and generate a trajectory stability score based on this prediction. A high score indicates a stable trajectory with minimal environmental interference, while a low score indicates that the trajectory is frequently disturbed by dynamic events. By training the relationship between historical trajectories and their speed changes, the model adaptively extracts potential influencing factors, improving its ability to express the stability of dynamic behavior.

[0031] The association feature focuses on the spatial relationship between a trajectory and other surrounding logistics entities. Within the time window of a trajectory point, if the minimum Euclidean distance to another entity falls below a safe distance threshold, the system records the interaction event and associates the corresponding entity identifiers, constructing an interaction map between trajectories to identify potential conflicts, collaborations, or resource competition. Combining these three types of features creates a multidimensional trajectory feature vector, which serves as the basic input for cluster analysis, improving the accuracy of depicting the complexity of logistics behavior and the spatial organization patterns.

[0032] In step S3, trajectory decoupling removes the trajectory constraint components caused by factory structure restrictions by introducing a spatial topology perception mechanism, thereby more accurately extracting the autonomous navigation behavior of logistics entities. First, in the spatial topology graph construction phase, the internal structure of the factory is abstracted into a graph G = (V, E), where the vertices V represent intersection nodes and workstations, and the edges E represent logistics channels. Three attributes are bound to each edge: length l length , channel width w width With the maximum capacity c capacity , constructing a navigation graph structure with physical constraints. Secondly, in the degree of freedom factor calculation stage, in order to quantify the degree of restriction of each channel space on the trajectory, the spatial degree of freedom factor is defined: ; Among them, w width is the current channel width, w min is the minimum reference value of the channel width in the system. The output value range of this function is (0,1). When the channel width is close to the minimum value, Approaching 0 indicates low channel freedom and strong constraints; when the channel width is significantly greater than the minimum value, When it approaches 1, it means that the path segment has high degrees of freedom and the trajectory variability is large.

[0033] In step S4, the improved spatiotemporal density clustering algorithm enhances the ability to identify high-frequency logistics paths and abnormal trajectories by introducing a composite density threshold and an adaptive trajectory number threshold mechanism. First, define the composite density threshold: , Among them, εs pace Represents the basic spatial distance threshold, which controls the clustering range of trajectory points in the spatial dimension; W zone is the spatial weight factor, which is dynamically adjusted according to the real-time congestion index of the current area. The more congested the area, the higher the W zone tends to increase, thereby increasing the local spatial density threshold, reflecting the impact of the environment on clustering sensitivity; ε time It represents the time distance threshold, which limits the proximity of trajectory points in the time dimension and is used to ensure the temporal consistency of clustering.

[0034] Then, set the adaptive minimum number of tracks threshold: , Among them, N total is the total number of trajectories currently being processed, and MinTrajs controls the minimum number of trajectories required to be included in each cluster. This strategy uses a fixed lower limit when the amount of trajectory data is small to prevent clustering instability caused by too few samples, and raises the threshold to enhance clustering confidence when the amount of data is large.

[0035] Noisy trajectories identified during the initial clustering process undergo a secondary analysis. The system continuously monitors the temporal and spatial distances between these trajectories and the cluster center. If a trajectory consistently remains within a preset distance interval for multiple consecutive periods, it indicates a potential pattern in its behavior. Based on this information, the system constructs a temporary path pattern and incorporates it into the path pattern library. This mechanism strengthens the identification of atypical repetitive behaviors, improving the integrity of the logistics path pattern library and the system's adaptive optimization capabilities.

[0036] In step S4, during the coupling component stripping phase, the original trajectory point sequence Ti and the topology graph G are simultaneously input into the graph neural network encoder, and a three-layer graph convolution operation is used to extract the spatial deformation features induced by the structure in the trajectory. The system dynamically adjusts the decoupling strength λ to: The lower the degree of freedom, The smaller , the larger λ is, thus enhancing the ability of the decoupling operation to correct the impact of topological constraints. Finally, trajectory decoupling is completed through the following formula: , Among them, GNN Encoder(Ti,G) represents the spatial constraint component extracted by the graph neural network encoder. This operation removes path features caused by physical structure from the original trajectory, retaining the active navigation intention of the logistics entity. This makes the generated decoupled trajectory more versatile and recognizable in terms of structural diversity, helping subsequent cluster analysis to accurately identify high-frequency path patterns with similar characteristics.

[0037] In step S4, cluster analysis further improves the accuracy and adaptability of path pattern recognition by introducing decoupling trajectory and spatial freedom factors. First, in terms of distance measurement, in order to more effectively measure the spatial and structural similarities between trajectories, an improved composite distance measurement formula is used: ; in, represents the Euclidean distance between the decoupled trajectories of trajectory i and j, which is used to reflect the geometric differences in trajectory morphology. and are the average spatial degrees of freedom factors of the path segments traversed by the trajectories, which are used to measure the difference in the degree of physical structural constraints between two trajectories. μ is the structural weight factor, which is used to adjust the proportion of the influence of structural differences on the overall distance. This distance metric combines behavioral path similarity with spatial structural similarity, making clustering more sensitive to actual path patterns that recur in similar structures.

[0038] In step S4, in density clustering, the spatial weight factor in the composite density threshold is no longer based on the static congestion index, but directly adopts the spatial freedom factor calculated in the decoupling stage. ,Right now ,in The larger the value, the higher the spatial tolerance of trajectory distribution in the clustering process. space With ε time is the basic distance threshold in the spatial and temporal dimensions. This setting reflects that trajectories are allowed to have higher morphological diversity in areas with small spatial constraints.

[0039] In step S4, trajectories that are not classified into any clusters in the clustering results are considered noise trajectories, and the system performs a secondary analysis on these trajectories. This method calculates the distance distribution between them and the decoupled trajectories of each identified cluster center. If the distance of a noise trajectory from a cluster center consistently falls outside a preset interval over multiple time periods, and this deviation trend remains continuous over a certain period, the trajectory is considered to represent a new path behavior intention. After the continuous overrun condition is met, the system generates a temporary path pattern for the trajectory and adds it to the pattern library. This enhances the sensitivity of path identification to heterogeneous path evolution and provides a foundation for subsequent pattern library updates. This mechanism improves the ability to respond to edge behaviors or early abnormal trends, ensuring that the path pattern library can dynamically adapt to subtle changes in the logistics environment.

[0040] In step S5, the core of route optimization is a spatiotemporal cost function constructed based on multi-dimensional factors to quantify the comprehensive cost of candidate routes in terms of transportation efficiency, safety risks, and time deviation. The specific cost function form is: ,

[0041] Among them, T trave l is the estimated travel time of the route, which measures the cost of the route in terms of transportation efficiency; d t It represents the spatial distance between the logistics entity and the nearest obstacle or other mobile unit at time t, d safe is the safety distance threshold, k is the growth coefficient of the distance penalty, which makes the risk term increase exponentially when approaching the safety distance. The second term comprehensively reflects the level of safety hazards that may occur in the entire path; the last term |T actual -T expected ∣ represents the degree of deviation between the actual completion time of the task and the expected time, and evaluates the time accuracy of the task.

[0042] The three costs are dynamically controlled by weight coefficients α, β, and γ, where α represents the priority of efficiency, β represents the sensitivity to safety distance, and γ controls the weight of the time deviation penalty.

[0043] During peak operating hours, the system increases β to 2.5 times the baseline value to enhance obstacle and congestion avoidance capabilities. During emergency missions, γ is increased to 3 times the baseline value to prioritize time-based meeting requirements. The adjustment strategy for these weight coefficients is iteratively updated in step S6 based on real-time feedback from path execution deviation data. This allows the optimization objective to dynamically adapt to different business scenarios and operating environments, improving the accuracy and robustness of path allocation. Finally, based on this cost function, the system searches for the path with the lowest total cost within the candidate path set and issues it as an execution task to the logistics equipment, ensuring that the path planning results strike a balance between time efficiency, safety, and timeliness.

[0044] In step S6, the deviation entropy processing measures the randomness and uncertainty of the path execution deviation to ensure the stability and adaptability of the path optimization system. First, the entropy index of the path deviation is calculated: ; Among them, p i represents the probability distribution within the i-th type deviation interval, E devMeasures the degree of discreteness and unpredictability of path execution deviations. When this entropy value exceeds a set threshold, it indicates that the path deviation has become highly unstable and difficult to predict using the existing pattern library. At this point, the system freezes the current path pattern library and stops adding new patterns to prevent erroneous information contamination. Subsequently, the system activates a reinforcement learning module to autonomously generate an emergency path planning strategy based on real-time environmental feedback and historical experience. This module interacts with the sensor network, issuing calibration instructions to the sensors and adjusting the sensor data acquisition accuracy and transmission frequency in real time to improve the accuracy and response speed of environmental perception. If the deviation persists and the difference between the decoupled trajectory and the planned path continues to exceed the tolerance range, the system triggers parameter retraining of the graph neural network encoder, using newly collected trajectory data and spatial topology information to update the model weights, thereby optimizing the encoding of spatial constraints and path behavior. This mechanism ensures that the model can adapt to environmental changes and path behavior evolution, achieving closed-loop adaptive optimization of path planning and execution.

[0045] From the above we can know: In this invention, by introducing graph neural networks and topological data analysis technology, an innovative "trajectory decoupling" step is proposed: first, a factory space topology map is constructed, and the "spatial freedom factor" reflecting the degree of channel space restriction is calculated; then, the spatial constraint component of the trajectory is extracted through the graph neural network encoder, and combined with the dynamically adjusted decoupling strength, the "instantaneous spatial coupling" component caused by physical structure restrictions is stripped away, retaining the "inherent path preference" of the logistics entity. By introducing graph neural networks and topological data analysis technology for trajectory decoupling, the "inherent path preference" and "instantaneous spatial coupling" components in the trajectory are effectively separated, fundamentally solving the problem of pseudo-cluster generation caused by trajectory spatial coupling. After actual testing, in areas with high spatial constraints, the pseudo-cluster generation rate was reduced from over 30% to below 5%, greatly improving the accuracy of path pattern recognition.

[0046] This invention fundamentally improves the rationality of path planning by improving the spatiotemporal density clustering algorithm and a path planning method based on an accurate pattern library. Firstly, the improved clustering algorithm introduces a "composite density threshold" and an "adaptive trajectory number threshold" mechanism. The composite density threshold is dynamically adjusted based on the spatial freedom factor (the lower the spatial constraint, the higher the tolerance for trajectory morphological diversity), while the adaptive trajectory number threshold dynamically changes based on the amount of trajectory data (a fixed lower limit is set to prevent clustering instability when the data volume is small, and a higher threshold is raised to enhance confidence when the data volume is large). This allows for more accurate identification of high-frequency path patterns. Secondly, the path optimization phase solves the optimal path based on a multidimensional spatiotemporal cost function (combining transportation time, safety risk, and time deviation). The weighting coefficients of the cost function (efficiency, safety, and time priority) can be dynamically adjusted based on the business scenario (for example, increasing safety weight by 2.5 times during peak hours and time weight by 3 times for urgent tasks). This eliminates the need for AGVs assigned to new tasks to be forced into existing queues, enabling them to activate alternate channels based on real-time environment and task requirements. Actual application data shows that the average driving distance of logistics equipment has been shortened by 10% to 15%, the transportation time has been reduced by 15% to 20%, and the flexibility of equipment operation and the rationality of route planning have been significantly enhanced.

[0047] In this invention, by eliminating pseudo-clusters and improving the rationality of path planning, the operating efficiency of the logistics system is fundamentally improved. On the one hand, the congestion and waiting time caused by unreasonable path planning of logistics equipment are greatly reduced, and the timeliness and accuracy of cargo transportation are significantly improved. On the other hand, the dynamic adjustment step monitors the deviation between the actual path and the planned path through the "deviation entropy" indicator (the higher the entropy value, the more unstable the deviation), triggering the incremental update of the path pattern library and the adjustment of the cost function parameters, forming a closed-loop optimization mechanism of "data collection-trajectory decoupling-clustering analysis-path optimization-deviation feedback", ensuring that the system can continuously adapt to environmental changes. In actual applications, factory throughput has increased by 15% to 25% compared with traditional methods, and the operating efficiency of logistics equipment has been significantly improved, effectively ensuring the continuity of smart factory production and providing strong support for the efficient operation of the manufacturing industry.

[0048] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0049] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A dynamic optimization method for smart factory logistics paths based on digital twin spatiotemporal trajectory clustering, characterized by: The method comprises the following steps: S1: Data synchronization is performed first: the distributed sensor network collects the location, speed and environmental data of the logistics entity at a preset period, and maps it to the digital twin platform in real time to generate a dynamic twin dataset; S2: Trajectory reconstruction: Responding to the logistics task completion event or time window triggering condition, reconstructing the structured spatiotemporal trajectory of the entity, and extracting the trajectory feature vector containing spatial distribution features, dynamic interaction features, and correlation features; S3: Decouple the trajectory: In response to the trajectory feature vector generation event in step S2, construct the factory space topology graph G = (V, E), where the vertex V is the key node of the path and the edge E is assigned attributes. ; Calculate the spatial freedom factor based on the channel width ; Extract the spatial constraint component of the trajectory through the graph neural network encoder, press Separate the coupled components to generate decoupled trajectories; S4: Perform cluster analysis: Input the decoupled trajectory into the improved spatiotemporal density clustering algorithm, and use the composite density threshold and adaptive threshold Identify high-frequency path patterns and build a path pattern library containing core node sequences, spatiotemporal constraints, and confidence levels; S5: Path optimization: respond to new logistics tasks or pattern library update events, match task requirements with the path pattern library, solve the optimal path based on the time-space cost function, and send it to the logistics equipment; S6: Perform the final dynamic adjustment step: according to the deviation entropy between the actual path and the planned path Trigger incremental update of path pattern library and adjustment of cost function parameters.

2. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S2, the extraction of trajectory feature vectors includes three types of information: spatial distribution features, dynamic interaction features, and correlation features, which respectively characterize the behavior patterns of logistics entities in the spatiotemporal process from different dimensions.

3. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S3, trajectory decoupling removes the trajectory constraint components caused by factory structure restrictions by introducing a spatial topology perception mechanism, thereby more accurately extracting the autonomous navigation behavior of logistics entities. First, in the spatial topology graph construction stage, the internal structure of the factory is abstracted into a graph G = (V, E), where the vertex V represents the intersection node and the work station, the edge E represents the logistics channel, and three types of attributes are bound to each edge: length l length , channel width w width With the maximum capacity c capacity ,construct a navigation graph structure with physical constraints; Secondly, in the degree of freedom factor calculation stage, in order to quantify the degree of restriction of each channel space on the trajectory, the spatial degree of freedom factor is defined: ； Among them, w width is the current channel width, w min is the minimum reference value of the channel width in the system. The output value range of this function is (0,1). When the channel width is close to the minimum value, Approaching 0 indicates low channel freedom and strong constraints; when the channel width is significantly greater than the minimum value, When it approaches 1, it means that the path segment has high degrees of freedom and the trajectory variability is large.

4. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S4, the improved spatiotemporal density clustering algorithm enhances the recognition capability of high-frequency logistics paths and abnormal trajectories by introducing a composite density threshold and an adaptive trajectory number threshold mechanism. First, the composite density threshold is defined as: ， Among them, εs pace Represents the basic spatial distance threshold, which controls the clustering range of trajectory points in the spatial dimension; W zone is the spatial weight factor, which is dynamically adjusted according to the real-time congestion index of the current area. The more congested the area, the higher the W zone tends to increase, thereby increasing the local spatial density threshold, reflecting the impact of the environment on clustering sensitivity; ε time It represents the time distance threshold, which limits the proximity of trajectory points in the time dimension and is used to ensure the temporal consistency of clustering; Then, set the adaptive minimum number of tracks threshold: ， Among them, N total is the total number of trajectories currently being processed, and MinTrajs controls the minimum number of trajectories that must be included in each cluster.

5. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S4, during the coupling component stripping phase, the original trajectory point sequence Ti and the topology graph G are simultaneously input into the graph neural network encoder, and the spatial deformation features induced by the structure in the trajectory are extracted using a three-layer graph convolution operation; The system dynamically adjusts the decoupling strength λ to: The lower the degree of freedom, The smaller , the larger λ is, thus enhancing the ability of the decoupling operation to correct the impact of topological constraints; finally, trajectory decoupling is completed through the following formula: ， Among them, GNN Encoder (Ti,G) represents the spatial constraint component extracted by the graph neural network encoder.

6. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S4, cluster analysis uses an improved composite distance metric formula in terms of distance metric by introducing decoupling trajectory and spatial freedom factor: ； in, represents the Euclidean distance between the decoupled trajectories of trajectory i and j, which is used to reflect the geometric differences in trajectory morphology. and are the average spatial degree of freedom factors of the path segments passed by the trajectory, which are used to measure the difference in the degree of physical structure restriction between the two trajectories, and μ is the structural weight factor.

7. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S4, in the density clustering, the spatial weight factor in the composite density threshold is no longer based on the static congestion index, but directly adopts the spatial freedom factor calculated in the decoupling stage. ,Right now ,in The larger the value, the higher the spatial tolerance of trajectory distribution in the clustering process. space With ε time is the basic distance threshold in the spatial and temporal dimensions.

8. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S4, trajectories that are not classified into any cluster in the clustering results are considered noise trajectories, and the system performs a secondary analysis on such trajectories. The method is: calculate the distance distribution between them and the decoupled trajectories of each identified cluster center. If the distance between a noise trajectory and a cluster center continuously falls outside the preset interval over multiple time periods, and the deviation trend remains continuous within a certain period, then the trajectory is considered to represent a new path behavior intention. After the continuous overrun condition is met, the system generates a temporary path pattern for the trajectory and adds it to the pattern library.

9. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S5, the core of the path optimization is a spatiotemporal cost function constructed based on multi-dimensional factors to quantify the comprehensive cost of the candidate path in terms of transportation efficiency, safety risk, and time deviation. The specific cost function form is: ， Among them, T trave l is the estimated travel time of the route, which measures the cost of the route in terms of transportation efficiency; d t It represents the spatial distance between the logistics entity and the nearest obstacle or other mobile unit at time t, d safe is the safety distance threshold, k is the growth coefficient of the distance penalty, which makes the risk term increase exponentially when approaching the safety distance. The second term comprehensively reflects the level of safety hazards that may occur in the entire path; the last term |T actual -T expected ∣ represents the degree of deviation between the actual completion time of the task and the expected time, and evaluates the time accuracy of the task; The three costs are dynamically controlled by weight coefficients α, β, and γ; among them, α represents the efficiency priority, β represents the sensitivity to the safety distance, and γ controls the weight of the time deviation penalty.

10. The method for dynamic optimization of smart factory logistics paths based on digital twin spatiotemporal trajectory clustering according to claim 1, characterized in that: In step S6, the deviation entropy processing measures the randomness and uncertainty of the path execution deviation to ensure the stability and adaptability of the path optimization system. First, the entropy index of the path deviation is calculated: ； Among them, p i represents the probability distribution within the i-th type deviation interval, E dev Measures the discreteness and unpredictability of path execution deviations. When the entropy value exceeds the set threshold, it indicates that the path deviation has become highly unstable and difficult to predict using the existing pattern library. At this point, the system freezes the current path pattern library and stops adding new patterns to avoid erroneous information pollution.

Images