AGV digital twin monitoring method based on three-dimensional simulation model

Abstract

The application discloses an AGV digital twin monitoring method based on a three-dimensional simulation model, relates to the technical field of digital twin monitoring, and comprises the following steps: performing multi-dimensional feature extraction on a target AGV to form an AGV behavior mode vector set; generating a nonlinear coupling matrix representing the correlation strength between AGV motion state parameters and environmental constraint parameters based on the vector set and environmental geometric information; dividing a three-dimensional digital twin model into a plurality of simulation units, and independently simulating local AGV disturbance responses of the units according to the matrix; mapping high-frequency motion events and low-frequency task events into the simulation units based on the local AGV disturbance responses to form a nonlinear disturbance coupling field; and dynamically adjusting a virtual AGV state based on the disturbance propagation trend of the disturbance coupling field. The application realizes nonlinear behavior synchronization of a virtual AGV and a physical AGV, thereby solving the problem that virtual and real behaviors of an AGV are difficult to be consistent in the prior art.

Description

Technical Field

[0001] This invention relates to the field of digital twin monitoring technology, and more specifically, to an AGV digital twin monitoring method based on a three-dimensional simulation model. Background Technology

[0002] With the rapid development of intelligent manufacturing and logistics systems, Automated Guided Vehicles (AGVs) are increasingly widely used in production, warehousing, and distribution scenarios. To improve the scheduling efficiency and operational safety of AGVs, digital twin technology has been introduced into AGV management systems. By establishing a mapping between virtual AGV models and the physical AGV states, operational status monitoring, fault prediction, and task optimization can be achieved.

[0003] Existing AGV digital twin systems are typically based on linear models or rigid mapping methods, focusing primarily on single-dimensional state simulations such as AGV pose, velocity, and path planning. In practical applications, AGVs need to simultaneously perform multiple tasks, including handling, obstacle avoidance, loading, and docking, and their behavior is complexly influenced by task commands, motion responses, and load states. Furthermore, when multiple AGVs operate in a shared environment, they also experience nonlinear coupling with each other and the environment, including velocity gradient changes, load interactions, and local spatial disturbances.

[0004] However, existing technologies have the following shortcomings:

[0005] Multi-timescale nonlinear behavior is difficult to synchronize: Traditional twin models have difficulty capturing high-frequency motion disturbances and low-frequency task events at the same time, making it impossible for virtual AGVs to accurately map the dynamic response of physical AGVs at the task layer, motion layer and energy consumption layer.

[0006] Large deviation between virtual and physical models: Due to insufficient consideration of the nonlinear coupling between AGVs and between AGVs and the environment, there is a large deviation between the virtual AGV state and the physical AGV state;

[0007] Insufficient accuracy in task simulation and scheduling optimization: Accumulated biases and lack of nonlinear behavior make it difficult to achieve ideal results in task scheduling, path planning, and energy consumption optimization based on digital twins.

[0008] Therefore, there is an urgent need for a digital twin monitoring method that can enable virtual AGVs to synchronously respond to the nonlinear behavior of physical AGVs across multiple time scales in complex dynamic environments, in order to improve virtual-physical consistency and the accuracy of task scheduling.

[0009] The above-disclosed technical solutions have at least the following technical problems: existing AGV digital twins mostly adopt linear or rigid coupling models, which cannot accurately reflect the nonlinear dynamic behavior of AGVs in complex tasks and environments, resulting in large deviations in the behavior of virtual AGVs and physical AGVs, making it difficult to achieve accurate task strategy simulation and collision risk warning. Summary of the Invention

[0010] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an AGV digital twin monitoring method based on a three-dimensional simulation model. By forming an AGV behavior pattern vector set and generating a nonlinear coupling matrix based on three-dimensional environmental information, the three-dimensional twin model is divided into simulation units to independently simulate disturbances. The virtual AGV state is dynamically adjusted by driving the disturbance coupling field, thereby achieving synchronization of the nonlinear behavior of the virtual AGV and the physical AGV, thus solving the problem of inconsistent virtual and physical AGV behaviors in the prior art.

[0011] To achieve the above objectives, the present invention provides the following technical solution:

[0012] A digital twin monitoring method for AGVs based on a three-dimensional simulation model includes the following steps: extracting multi-dimensional features from the target AGV to form an AGV behavior pattern vector set; generating a nonlinear coupling matrix characterizing the correlation strength between AGV motion state parameters and environmental constraint parameters based on the vector set and environmental geometric information; dividing the three-dimensional digital twin model into several simulation units, each unit independently simulating local AGV disturbance response according to the matrix; based on the local AGV disturbance response, associating the feature parameters of high-frequency motion events and low-frequency task events to the corresponding simulation units in the space and assigning influence weights to form a nonlinear disturbance coupling field of AGV disturbance and event influence within the simulation unit; and dynamically adjusting the virtual AGV state based on the disturbance propagation trend of the disturbance coupling field.

[0013] In a preferred embodiment, the step of extracting multi-dimensional features from the target AGV to form an AGV behavior pattern vector set specifically involves: collecting the AGV's motion state, task execution instructions, and load change information to construct a dynamic segmentation model based on task events; extracting multi-dimensional parameters within the task segments to form time-series feature data; performing dimensionality reduction and analysis on the feature data to map it into behavior pattern vectors representing the dynamic characteristics of the task stage; and organizing each behavior pattern vector according to the task time series to form a behavior pattern vector set.

[0014] In a preferred embodiment, the construction of a dynamic segmentation model based on task events specifically involves: collecting motion state and load signals during the AGV's task execution; detecting task event trigger points and using these as boundaries to dynamically segment continuous task data; and establishing a task semantic-driven dynamic segmentation model based on the time alignment relationship between motion data and task instructions within the segments.

[0015] In a preferred embodiment, generating a nonlinear coupling matrix characterizing the correlation strength between AGV motion state parameters and environmental constraint parameters includes: extracting time-varying motion characteristic parameters and energy consumption response characteristics of each AGV, and constructing a spatial adjacency relationship between AGVs and environmental constraint parameters by combining the three-dimensional geometric information of the environment; calculating the nonlinear dynamic coupling coefficients between AGVs and between AGVs and environmental constraint parameters based on behavior pattern vectors to form a nonlinear coupling matrix; and dynamically adjusting the weights of the coupling matrix according to the task state through an adaptive mechanism.

[0016] In a preferred embodiment, forming the nonlinear coupling matrix includes: setting the acting AGV as matrix rows, the affected AGV or environmental object as matrix columns, and filling in nonlinear dynamic coupling coefficients; determining the disturbance direction based on the behavior pattern vector and establishing a mapping with the task type; dynamically updating the disturbance intensity and direction of the matrix elements according to changes in the AGV motion state, load state, and environmental state; and using the updated nonlinear coupling matrix as input to the three-dimensional twin model simulation unit.

[0017] In a preferred embodiment, dividing the three-dimensional digital twin model into several simulation units, with each unit independently simulating the local AGV disturbance response according to the matrix, includes: dividing the three-dimensional digital twin model into multiple simulation units based on a nonlinear coupling matrix, with each unit corresponding to a set of spatially associated AGVs and environmental objects; each simulation unit independently calculating the virtual AGV state adjustment based on the disturbance intensity and direction corresponding to the matrix; for cross-neighborhood disturbances, the simulation unit coordinates the disturbance response of neighboring units based on the nonlinear dynamic coupling coefficient across units; and updating the disturbance response in the simulation unit in real time based on AGV behavior patterns and environmental state changes, and feeding the update results back to the global twin model.

[0018] In a preferred embodiment, the formation of the nonlinear disturbance coupling field of AGV disturbance and event influence within the simulation unit specifically involves: separating high-frequency motion events and low-frequency task events during AGV operation based on the disturbance response of each simulation unit; synchronizing high-frequency motion events and low-frequency task events within each simulation unit and determining their comprehensive influence weight based on the nonlinear coupling matrix; and constructing a nonlinear disturbance coupling field within the simulation unit according to the comprehensive influence weight.

[0019] In a preferred embodiment, the dynamic adjustment of the virtual AGV state based on the perturbation propagation trend of the perturbation coupling field includes: extracting the perturbation intensity gradient and direction vector within each simulation unit; determining the dynamic correction region of the virtual AGV according to the perturbation propagation trend, and using the perturbation intensity gradient change rate as the state adjustment trigger condition; when the perturbation intensity gradient exceeds a set threshold, synchronously updating the motion trajectory, execution posture, and energy consumption distribution of the virtual AGV according to the task response parameters in the behavior pattern vector; establishing a perturbation evolution feedback chain between each simulation unit, and feeding back the local adjustment results to the coupling coefficient of adjacent units in real time to correct the subsequent perturbation propagation direction.

[0020] In a preferred embodiment, the step of determining the perturbation direction based on the behavior pattern vector and establishing a mapping with the task type specifically means that the perturbation action corresponding to the obstacle avoidance task points in the direction away from the obstacle.

[0021] The path advancement task corresponds to the disturbance effect moving along the target path; the load transportation task corresponds to the disturbance effect being transmitted along the transport direction.

[0022] The technical effects and advantages of the AGV digital twin monitoring method based on a three-dimensional simulation model of the present invention are as follows:

[0023] 1. This invention improves the consistency between virtual and real systems, the accuracy of task simulation, and the reliability of scheduling optimization by modeling the multi-dimensional characteristics of the target AGV's behavior patterns, constructing a nonlinear coupling matrix based on behavior patterns and environmental information, and combining simulation unit partitioning and perturbation coupling field driving. This enables the virtual AGV to synchronously respond to the nonlinear behavior of the physical AGV at multiple time scales in the task layer, motion layer, and energy consumption layer. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating an AGV digital twin monitoring method based on a three-dimensional simulation model according to the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0026] Example 1, Figure 1 This invention presents a digital twin monitoring method for AGVs based on a three-dimensional simulation model, comprising the following steps:

[0027] S1, perform multi-dimensional feature extraction on the target AGV to form an AGV behavior pattern vector set;

[0028] The behavior pattern vector set is used to describe the nonlinear behavior characteristics of AGVs under different environmental and task conditions.

[0029] In this embodiment, the multi-dimensional feature extraction of the target AGV to form an AGV behavior pattern vector set specifically includes:

[0030] Collect motion state data, task execution instructions, and load change information of the target AGV under different task scenarios, and construct a dynamic segmentation model based on the task event trigger point;

[0031] Multidimensional parameters related to task execution strategy, motion response characteristics and energy consumption changes are extracted within each task segment to form a time series feature matrix; the time series feature matrix has task stage as row dimension, behavioral response parameters as column dimension, and energy consumption changes as weight.

[0032] By using feature reduction and covariance analysis, the time series feature matrix is ​​mapped into a behavioral pattern vector that reflects the dynamic characteristics of the task stage;

[0033] Organize all behavior pattern vectors into a behavior pattern vector set according to the task time series.

[0034] The construction of the dynamic segmentation model based on task event trigger points is specifically as follows:

[0035] When the AGV performs various tasks such as loading, handling, obstacle avoidance, and docking, its motion trajectory, speed, acceleration, orientation angle, and load sensing signals are collected in real time.

[0036] Detect key event triggers during task execution, including task switching, path planning updates, payload status changes, and abnormal obstacle avoidance behaviors;

[0037] Using each event trigger point as the segment boundary, the continuous motion and task data stream is dynamically sliced ​​so that each segment corresponds to a task semantic unit.

[0038] Based on the time alignment relationship between high-frequency motion data and low-frequency task instructions within a slice, a task semantic-driven dynamic segmentation model is established.

[0039] The dynamic segmentation model is used to eliminate behavioral ambiguity and task misalignment problems under time-equal sampling, and to achieve a consistent mapping between task logic and physical motion process.

[0040] S2, Based on the vector set and environmental geometric information, generate a nonlinear coupling matrix characterizing the correlation strength between the AGV motion state parameters and environmental constraint parameters;

[0041] The coupling matrix is ​​used to describe the propagation law of physical disturbances in three-dimensional space and multi-AGV systems.

[0042] In this embodiment, the step of generating a nonlinear coupling matrix characterizing the correlation strength between the AGV motion state parameters and environmental constraint parameters based on the vector set and environmental geometric information specifically involves:

[0043] Based on the behavior pattern vector set, time-varying motion characteristic parameters and energy consumption response characteristics of each AGV during task execution are extracted; the time-varying motion characteristics include velocity change rate, acceleration change rate, and load state; the energy consumption response characteristics include energy consumption change.

[0044] By combining the three-dimensional geometric information of the environment, a spatial adjacency diagram is constructed between the AGV and environmental constraint parameters such as obstacles, terrain slope, and path boundaries.

[0045] In the adjacency graph, the nonlinear dynamic coupling coefficients between AGVs and between AGVs and the environment are calculated based on the behavior pattern vectors. The coupling coefficients are used to characterize the time-varying influence of one motion disturbance on the state evolution of the other.

[0046] The nonlinear dynamic coupling coefficients are organized into a matrix according to the adjacency relationship and the task coupling direction to form a nonlinear coupling matrix, where each matrix element corresponds to the nonlinear influence of a specific AGV action on the neighboring AGV and local environmental disturbances under a specific task semantics.

[0047] Furthermore, through an adaptive update mechanism of the matrix, the coupling weights are dynamically adjusted according to changes in the task context and load state to achieve behavior semantic-driven multi-AGV-environment collaborative modeling.

[0048] The nonlinear dynamic coupling coefficients are specifically:

[0049]

[0050]

[0051] The nonlinear coupling matrix is ​​specifically:

[0052]

[0053]

[0054] in, For nonlinear dynamic coupling coefficients, This is a spatial adjacency matrix. This is a preset spatial attenuation coefficient used to adjust the attenuation rate of coupling with distance. , They are respectively The spatial position vector of object j, , , This is the behavioral response weighting coefficient (adjusted according to the actual needs of the project). The velocity difference term reflects the coupling strength of the motion behavior. This is the acceleration difference term, used to characterize the response delay characteristics. This represents the difference in energy consumption per unit time. Under load conditions, To prevent tiny constants with a denominator of zero, It is a nonlinear coupling matrix. This is the dynamically updated nonlinear coupling matrix. The set of task semantic parameters (including task priority, path congestion and risk level). To adaptively adjust the coefficients, ensure that the matrix adjusts the coupling weights in real time as the task and load change.

[0055] The nonlinear dynamic coupling coefficients are organized into a matrix according to adjacency relationships and task coupling directions to form a nonlinear coupling matrix, specifically as follows:

[0056] Each active AGV is represented as a row in the matrix, and each affected AGV or environment object is represented as a column in the matrix.

[0057] The nonlinear dynamic coupling coefficients are filled into the matrix elements according to the row and column correspondence, which are used to characterize the nonlinear disturbance intensity of the object on the neighboring AGV and environmental objects under the current task semantics.

[0058] Based on the nonlinear dynamic coupling coefficient and behavior pattern vector, the perturbation direction of each AGV in the current task is determined, and the task type is mapped to the perturbation direction;

[0059] Based on the changes in AGV motion state, load state and environmental state, the disturbance intensity and direction of each element in the matrix are dynamically updated. The update process directly depends on the nonlinear mapping of the matrix elements from the behavior pattern vector and the adjacency relationship, so as to ensure that the nonlinear coupling matrix reflects the interaction between AGV and environment in real time as the task is executed.

[0060] Using the nonlinear coupling matrix as input for the three-dimensional twin model simulation unit allocation and local disturbance simulation, the state adjustment of the virtual AGV is based on the disturbance intensity and direction of the matrix elements, thereby achieving consistency between the nonlinear behavior of the virtual AGV and the physical AGV.

[0061] The mapping of task type to disturbance direction is specifically as follows:

[0062] The obstacle avoidance task corresponds to a disturbance effect directed away from the obstacle;

[0063] The disturbance effect corresponding to the path advancement task moves along the direction of the target path;

[0064] The direction in which the disturbance effect of the load transportation task is transmitted along the transport direction.

[0065] S3, the three-dimensional digital twin model is divided into several simulation units, and each unit independently simulates the local AGV disturbance response according to the matrix;

[0066] The three-dimensional digital twin model is specifically as follows:

[0067] Constructed based on the actual physical dimensions, motion mechanism, sensor layout, and task operation scenario of the target AGV;

[0068] By combining the AGV behavior pattern vector set, the dynamic behavior characteristics of the physical AGV are mapped onto the three-dimensional model to form a virtual twin;

[0069] Embedding 3D geometric information of the environment into the model, including paths, obstacles, and terrain slopes, enables the twin model to accurately simulate the movement and task execution of AGVs in the real environment.

[0070] In this embodiment, dividing the three-dimensional digital twin model into several simulation units, with each unit independently simulating the local AGV disturbance response according to the matrix, specifically involves:

[0071] Based on the spatial structure and nonlinear coupling matrix of the three-dimensional digital twin model, the model is divided into several independent simulation units. Each simulation unit corresponds to a set of AGVs and environmental objects in the spatial neighborhood. During the division process, the number of AGVs and environmental objects and the spatial range contained in each simulation unit are determined according to the disturbance intensity and direction of action in the nonlinear coupling matrix.

[0072] For each simulation unit, based on the perturbation intensity and direction of the corresponding matrix element in the nonlinear coupling matrix, the perturbation information of the AGV acting on the neighboring AGV and environmental objects is distributed to each virtual AGV in the unit. Each simulation unit independently calculates the state adjustment of the virtual AGV after receiving the perturbation. The state adjustment includes the dynamic response of position, speed, acceleration and task execution strategy to maintain consistency with the behavior of the physical AGV.

[0073] For AGV disturbances with cross-neighborhood influence, the simulation unit coordinates the disturbance response of neighboring simulation units according to the matrix elements of the cross-unit in the nonlinear coupling matrix, so that the disturbance can be cascaded in the model according to the task semantics and coupling direction, ensuring that the local response of each simulation unit is both independent and dynamically coupled with neighboring units, thereby achieving nonlinear consistency of the overall AGV group behavior.

[0074] Based on the AGV behavior pattern vector and environmental state changes, the disturbance response of the virtual AGV in each simulation unit is updated in real time. The update process depends on the disturbance intensity and direction information in the nonlinear coupling matrix to ensure that the interaction between the AGV and the environment reflected by the nonlinear coupling matrix remains effective in subsequent task execution and simulation.

[0075] The virtual AGV state output by each simulation unit is used for the overall dynamic adjustment of the global 3D twin model, including the optimization of the virtual AGV's task execution strategy and the correction of path planning, so that the entire twin model can maintain nonlinear behavior in sync with the physical AGV group and provide accurate input for subsequent task scheduling and control.

[0076] For AGV disturbances with cross-neighborhood influence, the simulation unit coordinates the disturbance response of neighboring simulation units based on the matrix elements across units in the nonlinear coupling matrix, specifically as follows:

[0077] Based on the local disturbance calculation results in the aforementioned simulation unit, the corresponding cross-unit action terms in the nonlinear coupling matrix are identified. These action terms are used to characterize the potential influence direction and intensity of the AGV disturbance on adjacent simulation units in the simulation unit.

[0078] According to the aforementioned action, a cross-unit disturbance transmission channel is established, and the energy component and task state change component of the local disturbance are mapped to the target AGV in the adjacent simulation unit, so that the virtual AGV in the adjacent unit can make synchronous adjustments in task execution strategy and motion state.

[0079] During the disturbance propagation process, the simulation unit dynamically corrects the propagation rate and attenuation coefficient of the cross-unit disturbance based on the time delay parameter and path weight in the nonlinear coupling matrix, thereby ensuring the stability and physical consistency of the disturbance propagation.

[0080] After receiving neighborhood disturbance information, each simulation unit updates the state response of the virtual AGV based on the behavior pattern vector generated in the previous step, so that cross-unit disturbances form a continuous dynamic coupling chain in the global scope, realizing state coordination and nonlinear consistent evolution among different simulation units.

[0081] S4, based on the local AGV disturbance response, associate the characteristic parameters of high-frequency motion events and low-frequency task events with the corresponding simulation units in the space and assign them influence weights to form a nonlinear disturbance coupling field between AGV disturbance and event influence within the simulation unit;

[0082] In this embodiment, the formation of the nonlinear disturbance coupling field between AGV disturbance and event influence within the simulation unit specifically refers to:

[0083] Based on the local disturbance response results and cross-unit disturbance coordination information, frequency domain decomposition is performed on the high-frequency motion events and low-frequency task events during AGV operation. High-frequency events are used to characterize the dynamic stability of AGV in the short time scale, while low-frequency events are used to characterize the continuous state changes at the task execution level.

[0084] Within each simulation unit, the disturbance energy component of high-frequency motion events is time-aligned with the semantic change component of low-frequency task events, and the combined influence weight of the two in the spatial neighborhood is calculated based on the disturbance intensity and direction information corresponding in the nonlinear coupling matrix.

[0085] Based on the comprehensive influence weight, a nonlinear disturbance coupling field is constructed in the simulation unit. The coupling field is used to describe the energy transmission path of high-frequency disturbances under low-frequency task constraints and the dynamic interaction intensity between multiple AGVs.

[0086] Each simulation unit achieves unified modeling of disturbance energy at different frequency scales through this nonlinear disturbance coupling field, enabling the virtual AGV to synchronously respond to the nonlinear behavior of the physical AGV at multiple time scales.

[0087] The comprehensive influence weight is as follows:

[0088]

[0089]

[0090]

[0091] Define low-frequency components and high-frequency components:

[0092]

[0093] Time-frequency representation:

[0094] The nonlinear perturbation coupling field is specifically:

[0095]

[0096] in, To comprehensively influence the weighting, These are elements in the nonlinear coupling matrix that reflect the strength of the interaction between AGVs. , They are respectively , Standardized quantity, , These are the weighting coefficients for high and low frequencies (which can be set according to the importance of the task). The alignment delay between high-frequency and low-frequency signals. For the spatial influence kernel function based on distance decay, for At the spatial location of time t This refers to high-frequency disturbance energy. This represents low-frequency semantic variation (obtained by mapping behavioral pattern vectors). Let x be the set of neighborhood AGVs that have the ability to influence point x. It is a non-linear aggregation function. To decompose the threshold frequency, it can be determined based on AGV task characteristics or an adaptive algorithm, and used to distinguish between short-term motion disturbances and long-term task semantic changes. This is the time-frequency distribution function.

[0097] S5, dynamically adjust the virtual AGV state based on the disturbance propagation trend of the disturbance coupling field.

[0098] In this embodiment, the dynamic adjustment of the virtual AGV state based on the perturbation propagation trend of the perturbation coupling field specifically involves:

[0099] Based on the spatiotemporal distribution results of the aforementioned nonlinear perturbation coupling field, the perturbation intensity gradient and direction vector within each simulation unit are extracted. The gradient is used to characterize the transmission trend of perturbation energy in space, and the direction vector is used to characterize the dominant path of perturbation propagation.

[0100] Based on the disturbance propagation trend, the dynamic correction region of the virtual AGV is determined, and the rate of change of the disturbance intensity gradient within this region is used as the state adjustment trigger condition; when the disturbance intensity gradient exceeds the set threshold, the local adaptive adjustment process of the virtual AGV state is initiated.

[0101] During the state adjustment process, the virtual AGV updates its motion trajectory, execution posture and energy consumption distribution synchronously based on the task response parameters corresponding to its behavior pattern vector, so that its virtual response under the same task stage is consistent with the actual disturbance response of the physical AGV.

[0102] Meanwhile, the twin model introduces a perturbation evolution feedback chain between each simulation unit, and feeds back the local adjustment results to the coupling coefficients of adjacent units in real time to correct the subsequent perturbation propagation direction, thereby forming a global dynamic consistency update mechanism driven by the nonlinear perturbation coupling field.

[0103] Through the above process, the 3D twin model can achieve synchronous evolution and nonlinear consistent adjustment of virtual and real states in complex dynamic environments, ensuring that the virtual AGV's multi-scale response at the task layer, motion layer, and energy consumption layer matches the real behavior of the physical entity.

[0104] The perturbation coupling field can be used for collaborative optimization of multi-AGV systems, including:

[0105] Predict potential collision points, path conflict areas, and task priority conflicts in 3D simulation;

[0106] The system dynamically generates AGV scheduling optimization strategies based on the prediction results, providing a basis for decision-making in the scheduling system.

[0107] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0108] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0109] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0110] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0111] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0112] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A digital twin monitoring method for AGVs based on a three-dimensional simulation model, characterized in that, Includes the following steps: Multidimensional feature extraction is performed on the target AGV to form an AGV behavior pattern vector set; Based on the vector set and environmental geometric information, a nonlinear coupling matrix characterizing the correlation strength between AGV motion state parameters and environmental constraint parameters is generated. This includes: extracting time-varying motion characteristic parameters and energy consumption response characteristics of each AGV, and constructing the spatial adjacency relationship between AGVs and environmental constraint parameters by combining the three-dimensional geometric information of the environment; calculating the nonlinear dynamic coupling coefficients between AGVs and between AGVs and environmental constraint parameters based on behavior pattern vectors to form a nonlinear coupling matrix; and dynamically adjusting the weights of the coupling matrix according to the task state through an adaptive mechanism. The coupling coefficients are used to characterize the time-varying influence strength of one motion disturbance on the state evolution of the other. The three-dimensional digital twin model is divided into several simulation units, and each unit independently simulates the local AGV disturbance response according to the matrix. Based on the local AGV disturbance response, the characteristic parameters of high-frequency motion events and low-frequency task events are associated with the corresponding simulation units in the space and assigned influence weights, forming a nonlinear disturbance coupling field of AGV disturbance and event influence within the simulation unit; high-frequency events are used to characterize the dynamic stability of AGV in the short time scale, and low-frequency events are used to characterize the continuous state changes at the task execution level. The virtual AGV state is dynamically adjusted based on the perturbation propagation trend of the perturbation coupling field.

2. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 1, characterized in that, The process of extracting multi-dimensional features from the target AGV to form an AGV behavior pattern vector set is specifically as follows: Collect the motion status, task execution instructions and load change information of AGV, and construct a dynamic segmentation model based on task events; Extract multidimensional parameters within task segments to form time series feature data; The feature data is dimensionality reduced and analyzed, and mapped into behavioral pattern vectors that represent the dynamic characteristics of the task stage. The behavioral pattern vectors are organized according to the task time sequence to form a behavioral pattern vector set.

3. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 2, characterized in that, The construction of a dynamic segmentation model based on task events specifically involves: Collect motion status and load signals of the AGV during task execution; Detect task event trigger points and use them as boundaries to dynamically segment continuous task data; Based on the time alignment relationship between motion data within segments and task instructions, a task semantic-driven dynamic segmentation model is established.

4. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 3, characterized in that, The formation of the nonlinear coupling matrix includes: Set the active AGV as matrix rows and the affected AGV or environmental object as matrix columns, and fill in the nonlinear dynamic coupling coefficients. The perturbation direction is determined based on the behavior pattern vector and mapped to the task type. Based on changes in the AGV's motion state, load state, and environmental state, the disturbance intensity and direction of the matrix elements are dynamically updated. The updated nonlinear coupling matrix is ​​used as the input to the simulation unit of the three-dimensional twin model.

5. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 4, characterized in that, The process of dividing the three-dimensional digital twin model into several simulation units, with each unit independently simulating the local AGV disturbance response according to the matrix, includes: The three-dimensional digital twin model is divided into multiple simulation units based on the nonlinear coupling matrix, and each unit corresponds to a set of spatially associated AGVs and environmental objects. Each simulation unit independently calculates the virtual AGV state adjustment based on the corresponding disturbance intensity and direction in the matrix; For cross-neighborhood perturbations, the simulation unit coordinates the perturbation response of neighboring units based on the nonlinear dynamic coupling coefficients across units; Based on the AGV's behavior patterns and changes in environmental conditions, the disturbance response in the simulation unit is updated in real time, and the update results are fed back to the global twin model.

6. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 5, characterized in that, The nonlinear disturbance coupling field formed by the AGV disturbance and event influence within the simulation unit is specifically as follows: Based on the disturbance response of each simulation unit, high-frequency motion events and low-frequency task events in AGV operation are separated. High-frequency motion events and low-frequency task events are synchronized within each simulation unit, and their combined influence weights are determined based on a nonlinear coupling matrix. A nonlinear perturbation coupling field is constructed within the simulation unit based on the comprehensive influence weight.

7. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 6, characterized in that, The dynamic adjustment of the virtual AGV state based on the perturbation propagation trend of the perturbation coupling field includes: Extract the perturbation intensity gradient and direction vector within each simulation unit; The dynamic correction region of the virtual AGV is determined based on the disturbance propagation trend, and the disturbance intensity gradient change rate is used as the state adjustment trigger condition. When the disturbance intensity gradient exceeds the set threshold, the motion trajectory, execution posture and energy consumption distribution of the virtual AGV are updated synchronously based on the task response parameters in the behavior pattern vector. A disturbance evolution feedback chain is established between each simulation unit, and the local adjustment results are fed back to the coupling coefficient of adjacent units in real time to correct the subsequent disturbance propagation direction.

8. The AGV digital twin monitoring method based on a three-dimensional simulation model according to claim 7, characterized in that, The process of determining the perturbation direction based on the behavior pattern vector and establishing a mapping with the task type is as follows: The obstacle avoidance task corresponds to a disturbance effect directed away from the obstacle; The disturbance effect corresponding to the path advancement task moves along the direction of the target path; The direction in which the disturbance effect of the load transportation task is transmitted along the transport direction.

Images