Experimental Process Simulation and Resource Scheduling Optimization System Based on Digital Twin

Abstract

This invention discloses a digital twin-based experimental process simulation and resource scheduling optimization system, belonging to the field of laboratory automation technology. The system includes: a digital twin environment module for constructing a full-element digital twin of the laboratory, containing the geometric attributes, kinematic constraints, and logical states of experimental equipment; an AI shadow pre-simulation engine for accelerating pre-simulation within the digital twin before the formal experiment, using a spatiotemporal scanning algorithm to complete conflict detection; a multi-constraint dynamic scheduling module for generating an optimized instruction set containing timestamps and spatial path constraints based on the pre-simulation results to schedule physical equipment to perform experimental operations; and a bidirectional synchronization closed-loop module for achieving real-time synchronization between the virtual model and the physical environment through an adaptive filtering mechanism. This invention effectively avoids the risk of hard collisions in concurrent multi-robotic arm operations, optimizes experimental process timing, and improves laboratory operating efficiency and fault tolerance.

Description

Technical Field

[0001] This invention relates to the field of laboratory automation technology, and more specifically, to an experimental process simulation and resource scheduling optimization system based on digital twins. Background Technology

[0002] With the rapid development of life sciences, drug development, and chemical synthesis, experimental throughput and complexity have increased significantly, making traditional manual operation methods insufficient to meet the demands for high efficiency and precision. Laboratory automation systems, by introducing robotic arms, automated equipment, and control systems, have achieved standardized and batch processing of experimental procedures. However, existing automated laboratories still face several technical bottlenecks during operation: when multiple robotic arms operate concurrently, the lack of spatial and temporal coordination verification in the instruction logic can easily lead to physical collisions between equipment; the difficulty in accurately predicting the time required for each step during the experimental design phase can result in equipment idling or failure of critical samples; and when an anomaly occurs in any link, the entire production line often comes to a standstill, as the system lacks dynamic fault tolerance and adaptive adjustment capabilities, severely restricting the operational efficiency and safety of the laboratory.

[0003] To address the aforementioned issues, the industry has attempted to introduce digital twin technology to construct virtual laboratory models for the simulation and monitoring of physical experiments. For example, existing technologies disclose digital twin-based methods for pre-simulating ingredient preparation and digital twin systems for automated laboratories used in microbial strain screening. However, these solutions still have significant limitations in practical applications: digital twins are mostly limited to static modeling or remote monitoring functions, lacking a real-time synchronization mechanism with the physical environment; collision detection often relies on simple spatial occupancy judgment, making it difficult to accurately predict the risk of spatiotemporal overlap under complex motion trajectories; resource scheduling is still primarily based on static planning, unable to perform dynamic optimization and task remapping based on real-time status. Therefore, existing technologies have not yet effectively solved core problems in laboratory automation systems such as hard collision risks, timing conflicts, poor fault tolerance, and the disconnect between the virtual and real worlds. Summary of the Invention

[0004] To address the aforementioned technical problems in related technologies, this invention proposes an experimental process simulation and resource scheduling optimization system based on digital twins, which can overcome the above-mentioned shortcomings of existing technologies.

[0005] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows: A digital twin-based experimental process simulation and resource scheduling optimization system; This digital twin-based experimental process simulation and resource scheduling optimization system includes: The digital twin environment module is used to construct a full-element digital twin of the laboratory, which includes at least the geometric properties, kinematic constraints, and logical states of the experimental equipment. The AI ​​shadow pre-simulation engine is used to accelerate the pre-simulation of the experimental process in the digital twin before the formal experiment starts, and to collect the trajectory envelope of the robotic arm on the future time axis through the spatiotemporal scanning algorithm in order to complete the collision detection in the virtual space. The multi-constraint dynamic scheduling module is used to generate an optimized instruction set containing timestamps and spatial path constraints based on the pre-simulation results of the AI ​​shadow pre-simulation engine, so as to schedule physical equipment in the physical laboratory to perform experimental operations.

[0006] Furthermore, the AI ​​shadow pre-simulation engine calculates the collision probability of the robotic arm's envelope in four-dimensional spatiotemporal coordinates using the quaternion method. When a potential trajectory overlap is detected, the operating parameters of the robotic arm are automatically adjusted to avoid physical collisions.

[0007] Furthermore, the multi-constraint dynamic scheduling module uses a heuristic search algorithm to optimize the execution order of the experimental process in order to minimize the total experimental time; the heuristic search algorithm includes an improved genetic algorithm or an A* algorithm.

[0008] Furthermore, it also includes a bidirectional synchronous closed-loop module, which collects the pose, load and environmental parameters of physical devices in real time through a message subscription architecture based on edge computing, and uses an adaptive filtering mechanism to correct the prediction deviation in the digital twin, so as to realize the real-time synchronization between the virtual model and the physical environment.

[0009] Furthermore, the adaptive filtering mechanism includes: Construct a multidimensional twin deviation vector field containing position deviation, attitude deviation, and velocity deviation; Dynamic online tuning of the noise covariance matrix of Kalman filter based on long short-term memory network; When the system detects that the pose deviation of a specific region continues to exceed a preset threshold, it automatically triggers local twin model recalibration and corrects the kinematic constraint parameters of the region through backpropagation.

[0010] Furthermore, it also includes a fault diagnosis and repair module, which is used to instantly recalculate the optimal remediation path in the digital twin based on the fault status fed back by the physical device in real time, and generate a seamless switching command to be sent to the physical device.

[0011] Furthermore, it also includes an experimental process verification module, which is used to convert the experimental protocol into a directed acyclic graph and configure the virtual coordinate system of each device in the digital twin environment to perform real-time verification of the experimental process.

[0012] Furthermore, the digital twin environment module deploys ultra-wideband positioning tags and inertial measurement unit sensors, and uses a multi-sensor fusion algorithm to fuse data, in order to construct a high-fidelity digital twin.

[0013] Furthermore, it also includes a spatiotemporal resource competition and cooperation graph construction module, which is used to model experimental equipment and operating stations as spatiotemporal resource nodes and experimental steps as resource occupation arcs to form a three-dimensional spatiotemporal resource competition and cooperation directed graph; when resource deadlock is detected, the task remapping engine is triggered to dynamically migrate conflicting steps to functionally equivalent redundant stations or heterogeneous execution units.

[0014] Furthermore, it also includes a visual interactive interface to provide real-time monitoring of experimental status and issuance of operation commands.

[0015] The beneficial effects of this invention are as follows: By constructing a high-fidelity digital twin containing geometric attributes, kinematic constraints, and logical states, and performing accelerated pre-simulation and collision detection in virtual space, the risk of hard collisions in concurrent multi-robotic arm operations is effectively avoided; by generating an optimized instruction set with timestamps and spatial path constraints and using a heuristic search algorithm for dynamic scheduling, the overall efficiency of the experimental process is significantly improved, and the problems of equipment idleness and sample failure are effectively solved; by introducing an adaptive filtering mechanism based on edge computing to achieve real-time synchronization between the virtual model and the physical environment, and combining it with fault diagnosis and task remapping functions, the overall system fault tolerance and operational stability are comprehensively improved. Detailed Implementation

[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0017] The experimental process simulation and resource scheduling optimization system based on digital twin according to an embodiment of the present invention includes: The digital twin environment module is used to construct a full-element digital twin of the laboratory, which includes at least the geometric properties, kinematic constraints, and logical states of the experimental equipment. The AI ​​shadow pre-simulation engine is used to accelerate the pre-simulation of the experimental process in the digital twin before the formal experiment starts, and to collect the trajectory envelope of the robotic arm on the future time axis through the spatiotemporal scanning algorithm in order to complete the collision detection in the virtual space. The multi-constraint dynamic scheduling module is used to generate an optimized instruction set containing timestamps and spatial path constraints based on the pre-simulation results of the AI ​​shadow pre-simulation engine, so as to schedule physical equipment in the physical laboratory to perform experimental operations.

[0018] According to an embodiment of the present invention, the experimental process simulation and resource scheduling optimization system based on digital twins, in a specific implementation, the AI ​​shadow pre-simulation engine calculates the collision probability of the robotic arm envelope in four-dimensional spatiotemporal coordinates using the quaternion method. When potential trajectory overlap is detected, the operating parameters of the robotic arm are automatically adjusted to avoid physical collisions.

[0019] According to an embodiment of the present invention, in a specific implementation of the experimental process simulation and resource scheduling optimization system based on digital twins, the multi-constraint dynamic scheduling module uses a heuristic search algorithm to optimize the execution order of the experimental process in order to minimize the total experimental time; the heuristic search algorithm includes an improved genetic algorithm or an A* algorithm.

[0020] The experimental process simulation and resource scheduling optimization system based on digital twins according to an embodiment of the present invention further includes a bidirectional synchronization closed-loop module in a specific implementation. The bidirectional synchronization closed-loop module collects the pose, load and environmental parameters of physical devices in real time through a message subscription architecture based on edge computing, and uses an adaptive filtering mechanism to correct the prediction deviation in the digital twin, thereby realizing real-time synchronization between the virtual model and the physical environment.

[0021] In a specific embodiment of the experimental process simulation and resource scheduling optimization system based on digital twins according to an embodiment of the present invention, the adaptive filtering mechanism includes: Construct a multidimensional twin deviation vector field containing position deviation, attitude deviation, and velocity deviation; Dynamic online tuning of the noise covariance matrix of Kalman filter based on long short-term memory network; When the system detects that the pose deviation of a specific region continues to exceed a preset threshold, it automatically triggers local twin model recalibration and corrects the kinematic constraint parameters of the region through backpropagation.

[0022] The experimental process simulation and resource scheduling optimization system based on digital twins according to an embodiment of the present invention further includes a fault diagnosis and repair module in a specific implementation, which is used to instantly recalculate the optimal remedial path in the digital twin based on the fault status fed back by the physical device in real time, and generate a seamless switching command to be sent to the physical device.

[0023] The experimental process simulation and resource scheduling optimization system based on digital twin according to an embodiment of the present invention further includes an experimental process verification module in a specific implementation. The experimental process verification module is used to convert the experimental protocol into a directed acyclic graph and configure the virtual coordinate system of each device in the digital twin environment to perform real-time verification of the experimental process.

[0024] According to an embodiment of the present invention, the experimental process simulation and resource scheduling optimization system based on digital twins, in a specific implementation, the digital twin environment module deploys ultra-wideband positioning tags and inertial measurement unit sensors, and uses a multi-sensor fusion algorithm to perform data fusion in order to construct a high-fidelity digital twin.

[0025] The experimental process simulation and resource scheduling optimization system based on digital twins according to an embodiment of the present invention further includes a spatiotemporal resource competition graph construction module in a specific implementation. This module is used to model experimental equipment and operating workstations as spatiotemporal resource nodes and experimental steps as resource occupation arcs, forming a three-dimensional spatiotemporal resource competition directed graph. When a resource deadlock is detected, the task remapping engine is triggered to dynamically migrate conflicting steps to functionally equivalent redundant workstations or heterogeneous execution units.

[0026] The experimental process simulation and resource scheduling optimization system based on digital twins according to an embodiment of the present invention further includes a visual interactive interface in a specific implementation, which provides real-time experimental status monitoring and operation command issuance functions.

[0027] To facilitate understanding of the above technical solutions of the present invention, the following embodiments will be used to describe the above technical solutions of the present invention in detail.

[0028] Example 1 This embodiment provides an experimental process simulation and resource scheduling optimization system based on digital twins. The system includes: a digital twin environment module, an AI shadow pre-simulation engine, a multi-constraint dynamic scheduling module, a bidirectional synchronous closed-loop module, a fault diagnosis and repair module, an experimental process verification module, a spatiotemporal resource competition and cooperation graph construction module, and a visual interactive interface.

[0029] The digital twin environment module is used to construct a full-element digital twin of the laboratory, which includes the geometric attributes, kinematic constraints, and logical states of the experimental equipment. Geometric attributes include the equipment's dimensions and installation location; kinematic constraints include parameters such as the robot arm's degrees of freedom, speed, and acceleration; and logical states include the equipment's occupancy / idle status, current temperature, and liquid level. Ultra-wideband positioning tags and inertial measurement unit (IMU) sensors are deployed within the laboratory, and 5G CPEs or industrial Wi-Fi 6 base stations are deployed at key nodes to construct a low-latency, highly reliable industrial communication network. The system employs a multi-sensor fusion algorithm to fuse the absolute position data provided by the ultra-wideband positioning tags with the relative motion data provided by the IMU sensors, correcting the measurement errors of individual sensors through data complementarity.

[0030] The AI ​​shadow simulation engine is used to accelerate the simulation of the experimental process in a digital twin before the formal experiment begins. In this embodiment, the AI ​​performs a 100x accelerated simulation in virtual space. The system uses a spatiotemporal scanning algorithm to calculate the collision probability of the robotic arm envelope in four-dimensional spatiotemporal coordinates using the quaternion method, thus achieving conflict detection. When potential overlap of trajectories between robotic arms is detected, the system automatically adjusts the operating parameters of the robotic arms, delays the operation of the corresponding arm position, and modifies the grasping angle to avoid physical collisions.

[0031] The multi-constraint dynamic scheduling module generates an optimized instruction set containing timestamps and spatial path constraints based on the pre-simulation results of the AI ​​shadow pre-simulation engine, to schedule physical equipment in the physics laboratory to perform experimental operations. In this embodiment, the system introduces an improved genetic algorithm as a heuristic search algorithm to optimize the experimental process while satisfying the logic of the chemical process, thereby minimizing the total experimental time. Through this module, the system can automatically adjust the execution order of experimental steps based on the real-time fault status feedback from the equipment during the experiment.

[0032] The bidirectional synchronous closed-loop module is used to collect the pose, load, and environmental parameters of physical devices in real time through an edge computing-based message subscription architecture. The system employs an adaptive filtering mechanism to correct prediction biases in the digital twin, achieving real-time synchronization between the virtual model and the physical environment. Specifically, the system constructs a multi-dimensional twin bias vector field containing position bias, attitude bias, and velocity bias, and dynamically tunes the noise covariance matrix of the Kalman filter online based on a long short-term memory network. When the system detects that the pose bias in a specific area continuously exceeds a preset threshold (e.g., a pose bias greater than 0.1 mm persists near a centrifuge), it automatically triggers local twin model recalibration, correcting the kinematic constraint parameters of that area through backpropagation.

[0033] The fault diagnosis and repair module is used to instantly recalculate the optimal remediation path in the digital twin based on the fault status fed back by the physical device in real time, and generate a seamless switching command to be sent to the physical device.

[0034] The experimental procedure verification module is used to convert the experimental protocol into a directed acyclic graph and configure the virtual coordinate system of each device in the digital twin environment to perform real-time verification of the experimental procedure.

[0035] The spatiotemporal resource competition and cooperation graph construction module is used to model experimental equipment and operating stations as spatiotemporal resource nodes, and experimental steps as resource-occupying arcs, forming a three-dimensional spatiotemporal resource competition and cooperation directed graph. When a resource deadlock is detected, the task remapping engine is triggered to dynamically migrate conflicting steps to functionally equivalent redundant stations or heterogeneous execution units, and the global spatiotemporal envelope is updated synchronously.

[0036] The visual interactive interface provides real-time monitoring of experimental status and issuance of operation commands, enabling visual management and remote control of the experimental process. The interface clearly displays the current progress of each experimental step, equipment status, material flow trajectory, and the temporal relationship of the entire experimental process.

[0037] Through the above technical solution, this embodiment realizes the dynamic simulation of the experimental process, effectively avoiding the risk of hard collisions caused by instruction logic loopholes, and improving the safety and reliability of the experiment. At the same time, the system can achieve staggered operations of the robotic arms within extremely small gaps, maximizing space utilization and improving the overall efficiency of the laboratory.

[0038] Example 2 This embodiment, based on Embodiment 1, further provides specific implementation steps for the system.

[0039] Step 1: Building a Digital Twin Environment First, the system uses a 3D scanning device to perform a 3D scan of the laboratory environment, acquiring its geometric model and material information. Then, based on the laboratory's functional requirements, it divides the laboratory into zones and defines corresponding kinematic constraints for each zone, including parameters such as the zone's center point coordinates, zone dimensions, and permitted equipment types. Simultaneously, the system establishes kinematic models for the experimental equipment, including the equipment's base coordinate system, the robotic arm's range of motion, and sensor layout. Finally, the system uses an experimental procedure verification module to convert the experimental protocol into a directed acyclic graph (DAG) for subsequent experimental procedure verification.

[0040] Step 2: Experimental Procedure Rehearsal Before the experiment began, the system activated its AI shadow simulation engine, allowing the AI ​​to perform accelerated simulations in a virtual space. During the simulation, the system calculated the state variables of each device in real time, including parameters such as position, attitude, and load, and stored this information in a database. Simultaneously, the system used a spatiotemporal scanning algorithm to calculate the trajectory envelope of the robotic arms on the future timeline, determining whether there was any spatial overlap between the robotic arms. If overlap was found, the system automatically adjusted the motion parameters of the robotic arms to optimize their trajectories.

[0041] Step 3: Experimental Procedure Optimization Based on the pre-simulation results, the multi-constraint dynamic scheduling module generates an optimized instruction set with timestamps and spatial path constraints. The system first initializes the experimental task scheduling table according to the experimental requirements and the current state of the equipment. Then, the system optimizes the execution order of the experimental process using a heuristic search algorithm, minimizing the total experimental time while ensuring all chemical process requirements are met. In this embodiment, the heuristic search algorithm can be the A* algorithm or an improved genetic algorithm. During the optimization process, the system comprehensively considers the equipment's movement time, processing time, waiting time, and various possible waiting scenarios, including equipment failure and untimely material preparation.

[0042] Step 4: Experiment Execution and Monitoring During the experimental execution phase, the system collects real-time status information of the physical equipment through a bidirectional synchronous closed-loop module. The system uses a Kalman filter algorithm to process the collected data and correct prediction biases in the virtual model. When the experimental procedure deviates from the prediction, the system automatically adjusts the execution order of the experimental steps or the motion parameters of the equipment to ensure the experiment proceeds as expected. Simultaneously, the system provides real-time monitoring and remote control functions through a visual interactive interface, allowing operators to promptly understand the experimental progress and make adjustments.

[0043] Step 5: Post-experiment processing After the experiment, the system organizes and analyzes the experimental data. First, the system determines whether the experiment achieved the expected results. If not, it returns to step 2 to re-optimize the experimental process. If the expected results are achieved, the following steps are taken: The fault diagnosis and repair module analyzes the causes of problems encountered during the experiment based on the real-time fault status feedback from the equipment, and optimizes subsequent experimental procedures based on the fault diagnosis results; the system stores the experimental data in the database, providing basic data for subsequent experimental scheme optimization; the system automatically generates an experimental report, including the experimental process, equipment operating parameters, and experimental results.

[0044] Through the above implementation steps, this embodiment achieves dynamic optimization and real-time monitoring of the experimental scheme, effectively solving problems such as hard collision risks and timing conflicts in the experiment, and improving the reliability and efficiency of the experiment. Simultaneously, the system can continuously optimize the experimental process based on the experimental results, improving the success rate of the experiment and product quality.

[0045] In summary, by utilizing the technical solutions described above in this invention, a high-fidelity digital twin containing geometric attributes, kinematic constraints, and logical states is constructed, and accelerated pre-simulation and collision detection are performed in virtual space, thereby effectively avoiding the risk of hard collisions in concurrent operations of multiple robotic arms. By generating an optimized instruction set with timestamps and spatial path constraints and using a heuristic search algorithm for dynamic scheduling, the overall efficiency of the experimental process is significantly improved, and the problems of equipment idleness and sample failure are effectively solved. By introducing an adaptive filtering mechanism based on edge computing to achieve real-time synchronization between the virtual model and the physical environment, and combining fault diagnosis and task remapping functions, the overall system fault tolerance and operational stability are comprehensively improved.

[0046] 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-based experimental process simulation and resource scheduling optimization system, characterized in that, include: The digital twin environment module is used to construct a full-element digital twin of the laboratory, which includes at least the geometric properties, kinematic constraints, and logical states of the experimental equipment. The AI ​​shadow pre-simulation engine is used to accelerate the pre-simulation of the experimental process in the digital twin before the formal experiment starts, and to collect the trajectory envelope of the robotic arm on the future time axis through the spatiotemporal scanning algorithm in order to complete the collision detection in the virtual space. The multi-constraint dynamic scheduling module is used to generate an optimized instruction set containing timestamps and spatial path constraints based on the pre-simulation results of the AI ​​shadow pre-simulation engine, so as to schedule physical equipment in the physical laboratory to perform experimental operations.

2. The digital-twin-based experiment flow simulation and resource scheduling optimization system according to claim 1, wherein, The AI ​​shadow pre-simulation engine calculates the collision probability of the robotic arm's envelope in four-dimensional spatiotemporal coordinates using the quaternion method. When a potential trajectory overlap is detected, the operating parameters of the robotic arm are automatically adjusted to avoid physical collisions. 3.The digital-twin-based experiment flow simulation and resource scheduling optimization system of claim 1, wherein, The multi-constraint dynamic scheduling module uses a heuristic search algorithm to optimize the execution order of the experimental process in order to minimize the total experimental time; the heuristic search algorithm includes an improved genetic algorithm or an A* algorithm.

4. The digital-twin-based experiment flow simulation and resource scheduling optimization system according to claim 1, wherein, It also includes a bidirectional synchronization closed-loop module, which collects the pose, load and environmental parameters of physical devices in real time through a message subscription architecture based on edge computing, and uses an adaptive filtering mechanism to correct the prediction deviation in the digital twin, so as to realize the real-time synchronization between the virtual model and the physical environment.

5. The digital-twin-based experiment flow simulation and resource scheduling optimization system according to claim 1, wherein, The adaptive filtering mechanism includes: Construct a multidimensional twin deviation vector field containing position deviation, attitude deviation, and velocity deviation; Dynamic online tuning of the noise covariance matrix of Kalman filter based on long short-term memory network; When the system detects that the pose deviation of a specific region continues to exceed a preset threshold, it automatically triggers local twin model recalibration and corrects the kinematic constraint parameters of the region through backpropagation.

6. The digital-twin-based experiment flow simulation and resource scheduling optimization system of claim 1, wherein, It also includes a fault diagnosis and repair module, which is used to instantly recalculate the optimal remediation path in the digital twin based on the fault status fed back by the physical device in real time, and generate a seamless switching command to be sent to the physical device.

7. The digital-twin-based experiment flow simulation and resource scheduling optimization system according to claim 1, wherein, It also includes an experimental process verification module, which is used to convert the experimental protocol into a directed acyclic graph and configure the virtual coordinate system of each device in the digital twin environment to perform real-time verification of the experimental process. 8.The digital-twin-based experiment flow simulation and resource scheduling optimization system of claim 1, wherein, The digital twin environment module deploys ultra-wideband positioning tags and inertial measurement unit sensors, and uses a multi-sensor fusion algorithm to fuse data in order to construct a high-fidelity digital twin. 9.The digital-twin-based experiment flow simulation and resource scheduling optimization system of claim 1, wherein, It also includes a spatiotemporal resource competition and cooperation graph construction module, which is used to model experimental equipment and operating stations as spatiotemporal resource nodes, and experimental steps as resource occupation arcs to form a three-dimensional spatiotemporal resource competition and cooperation directed graph; When a resource deadlock is detected, the task remapping engine is triggered to dynamically migrate the conflicting steps to a functionally equivalent redundant workstation or heterogeneous execution unit.

10. The digital-twin-based experiment flow simulation and resource scheduling optimization system of claim 1, wherein, It also includes a visual interactive interface for providing real-time monitoring of experimental status and issuing operation commands.