Intelligent kitchen supervision optimization system based on digital twinning and construction method thereof

Abstract

The invention provides an intelligent cooking supervision optimization system based on digital twinning and a construction method, and solves the problems of insufficient whole-process supervision of a traditional kitchen, lagging of real-time early warning and difficulty in operation standardization. The system constructs a virtual model through high-precision modeling (geometry, physics, behaviors and rules); the whole stir-frying link is uniformly coded by using an industrial internet identification analysis technology, so that data tracking and tracing are realized; a real-time data acquisition, transmission and analysis mechanism is established, and seamless virtual-real mapping between physical equipment and a digital platform is realized; and a visual control interface is constructed to support ordering and remote control. A scheduling rehearsal mechanism based on an improved ant colony algorithm is innovatively proposed, multi-task multi-pot dish distribution scheduling is optimized, trial and error cost is reduced, and meal delivery efficiency and system stability are improved. The system provides support for digital and intelligent transformation of the catering industry, and improves food safety supervision and kitchen operation efficiency.

Description

Technical Field

[0001] The present invention relates to fields such as industrial Internet identification resolution, serial communication, algorithm models, digital twin technology, 3D modeling technology, etc., and specifically relates to an intelligent kitchen supervision optimization system based on digital twin and its construction. Background Art

[0002] Currently, China's catering industry is facing two major bottlenecks: food safety issues and inefficient kitchen management. According to the "2023 Catering Industry Safety Risk Report", approximately 68% of foodborne diseases are closely related to operational errors in the food processing link. The traditional kitchen management mode mainly relies on manual inspections and subjective experience, making it difficult to achieve full-process, precise and effective supervision. Against this background, the introduction of digital twin technology has become an important breakthrough to break the deadlock and improve the level of food safety and cooking science.

[0003] The core value of the digital twin system lies in achieving "holographic perception" of the kitchen operation status. By deploying temperature and humidity sensors, photoelectric sensors, cameras, and various Internet of Things devices at each key node in the kitchen, the system can not only collect and monitor the full life cycle data of ingredients from storage to processing and cooking in real time, but also continuously monitor environmental changes and equipment status. For example, when a piece of chicken is stored in the fresh vegetable storage cabinet for too long or the temperature is abnormal, the system can immediately trigger an early warning mechanism and automatically retrieve relevant historical data, tracing back to the supplier information and transportation records, thus forming a complete chain of responsibility, significantly improving the transparency and traceability of food safety management.

[0004] For high-risk utensils such as robotic arms and woks in the kitchen, relying on its high-fidelity modeling ability, the digital twin system can achieve accurate and rapid collision detection, making up for the blind spots in traditional manual supervision. This management mode based on real-time data analysis and early warning mechanism not only reduces the risk of human error, but also provides a scientific basis for the maintenance and update of kitchen equipment, thereby further improving the overall management efficiency and safety.

[0005] In terms of the visualization and standardized management of the cooking process, the system also has unique advantages. With the help of advanced 3D modeling technology, the system can accurately restore the real kitchen scene, and through dynamic data overlay, it can display key parameters such as the grasping state, movement trajectory of the robotic arm, stove firepower, temperature inside the pot, and the remaining quantity of dishes in real time. This interactive method of integrating virtual and real not only provides intuitive professional operation demonstrations for novice cooks to help them quickly master the standard operation process, but also enables chain catering enterprises to achieve centralized monitoring of multiple kitchens. With the help of data visualization tools such as heat maps, the system can also deeply analyze the efficiency bottlenecks in each link and automatically optimize the meal delivery process, thus transforming valuable cooking experience into reproducible and standardized digital assets. It is worth mentioning that the physical scene relied on by this invention has been patented, ensuring the accuracy and exclusivity of the digital mapping of the real kitchen environment.

[0006] Generally speaking, digital twin technology brings new concepts of monitoring and optimization to traditional kitchen management. It not only effectively improves the accuracy of food safety supervision, greatly reduces the safety hazards caused by human operation errors, but also constructs an intelligent, standardized and sustainable management platform for catering enterprises. The comprehensive system integrating monitoring, early warning and process optimization will promote the catering industry towards higher safety standards and management efficiency. Summary of the Invention

[0007] The present invention discloses an intelligent kitchen supervision and optimization system based on digital twin and its construction method. The technical solutions and innovation points involved in this patent mainly focus on aspects such as model establishment, real-time monitoring, virtual-real mapping, algorithm optimization, and system integration, rather than the reproduction of the physical scene itself. Its purpose is to address the problems existing in traditional kitchen management, such as insufficient full-process supervision, lagging real-time warning, and difficulty in operation standardization. By constructing an intelligent kitchen management platform integrating virtual and real and driven by data, it realizes the holographic perception, risk warning, process optimization, and equipment maintenance of the kitchen operation status, thereby effectively improving food safety and management efficiency.

[0008] In the system design, advanced 3D modeling software is first used to model the physical equipment in the kitchen with high fidelity, accurately restore the real operation scenes of the four stages of food preparation, food transportation, frying and serving, and provide an accurate physical basis for the subsequent digital twin. Then, unified information identification and coding are implemented for the equipment involved in each stage. The food identification records the source, type, storage location, inventory and shelf life of the food in detail, and supports real-time tracking and scheduling; the wok identification covers the real-time status, temperature, heat setting and frying time of the equipment and other key data; the fried finished product identification integrates the food and wok information to form a complete dish production record, ensuring that the release time, operating temperature and other data of each ingredient can be traced. All generated identification codes generate unique identification codes through the identification registration system and are uploaded to the industrial Internet identification resolution platform, thereby realizing seamless connection and data synchronization between physical equipment and digital environment.

[0009] Based on this high-fidelity modeling and precise coding, the present invention constructs a digital twin control interface, and through dynamic data superposition and visualization technology, realizes an intuitive display of the real-time operation status of the kitchen. The control interface not only displays key parameters such as the gripping state of the robotic arm, motion trajectory, stove firepower, pot temperature, and the number of remaining dishes, but also supports remote monitoring and operation, providing managers with a standardized, intuitive and real-time supervision platform. At the same time, on the basis of virtual-real connection, the present invention further introduces an optimization scheduling and preview simulation module, which verifies the rationality and efficiency of tasks at each stage by previewing the scheduling algorithm in a digital twin environment, and reduces unnecessary energy consumption and material loss. The system adopts an innovative multi-task and multi-pot scheduling algorithm under the constraints of side dish resources, combined with a greedy algorithm, a priority algorithm and a shortest time job first (SJF) algorithm, to realize the intelligent allocation of multi-pot parallel operations; in scenarios with high task scheduling complexity, an ant colony algorithm is also introduced to generate a global optimal scheduling solution, which significantly improves the efficiency of serving meals and minimizes the idle time of robotic arms and equipment.

[0010] In summary, the present invention realizes intelligent monitoring, real-time tracking and efficient scheduling of the entire kitchen process through high-precision modeling, unified coding, seamless connection between virtual and real, intuitive twin control interface and advanced scheduling algorithm preview, which significantly improves the transparency of food safety management and meal delivery efficiency, and provides reliable technical support for the transformation of catering enterprises towards standardization, intelligence and digitalization.

[0011] Step 1: Building a high-fidelity physical model in a twin environment

[0012] The intelligent kitchen modeling is divided into five major modules, namely the pre-cut vegetable cabinet, the robotic arm, the slide rail, the cooking pan, and the ingredient modeling. Each module is finely designed using 3DMAX and SOLIDWORK modeling software to ensure the complete and accurate restoration of the actual kitchen scene in the digital twin environment. Specifically, in the modeling process of the present invention, not only the precise construction of the geometric shape is concerned, but also all-round modeling is carried out in terms of physics, behavior, rules, etc.

[0013] In terms of geometric modeling, the focus is on capturing the shape, structure, and dimensional data of each physical device to ensure that the spatial relationship and layout between components are highly consistent with the real scene. Physical modeling focuses on factors such as the material properties, motion parameters, and mechanical characteristics of the device, so as to accurately simulate the physical response of the actual device in the digital environment. In terms of behavior modeling, by collecting the dynamic data and operation modes of the device during operation, an action model of the device under different working conditions is constructed, and then the real-time restoration of the device operation state is realized. Rule modeling involves the setting of device operation processes, interaction rules, and safety constraints, etc., providing a strict behavioral basis for subsequent intelligent scheduling and control.

[0014] Through the comprehensive modeling in the above four aspects of geometry, physics, behavior, and rules, the present invention realizes the construction of a high-fidelity physical model, providing a solid physical foundation and data support for the digital twin system.

[0015] Step 2. Design of identification and resolution coding

[0016] In order to realize the real-time recording of the operation states of each component, the present invention adopts advanced identification and resolution coding technology to carefully divide the entire cooking process into four links: vegetable preparation, transportation, frying, and serving. In each link, unified identification coding is carried out for relevant equipment and operation elements. Taking the vegetable preparation link as an example, the system encodes the information of each type of dish in detail, including the source, type, storage location, inventory, and shelf life, etc., to ensure that each piece of data can be accurately traced; in the transportation link, the focus is on encoding the operation state of the robotic arm, and dynamically records its dynamic information such as grasping, moving, and handover; while in the frying link, the state data of the frying pan is identified, covering temperature, heat setting, frying time, and other key parameters. All encoded device information will be automatically uploaded to the industrial Internet identification and resolution platform, realizing seamless data docking from physical devices to the digital environment, providing a solid data foundation and technical support for the subsequent real-time monitoring and dynamic scheduling of the digital twin system. Through this series of fine identification and resolution and coding measures, the present invention not only ensures the accuracy and integrity of data collection, but also lays a solid foundation for the intelligent management and safety warning of the entire system.

[0017] Step 3. Virtual-real mapping between the twin and physical environments

[0018] After completing the construction of high-fidelity physical models and unified identification coding, the present invention further realizes seamless mapping between the physical environment and the digital twin, that is, virtual-real mapping. By establishing a data transmission interface and a real-time synchronization mechanism, all encoded physical equipment status and operating data can be uploaded to the digital twin platform in real time, realizing dynamic mapping between physical equipment and virtual models. Specifically, the system dynamically reproduces the four links of food preparation, transportation, frying and serving in the digital twin environment, including not only the geometric and physical characteristics of the equipment, but also its behavioral status and operating rules. With the help of the unique identification code provided by the industrial Internet identification resolution platform, the system can accurately match the real-time data of each device to the corresponding digital model, ensuring that the equipment status in the virtual environment is highly consistent with the actual situation. In addition, through the data transmission and resolution module, all kinds of data generated in the physical environment, such as temperature and humidity, motion trajectory, equipment working status, etc., can be presented in the twin system with the shortest delay, so that managers can intuitively understand the real-time situation on site through the virtual interface, and achieve rapid early warning and response. The realization of virtual-reality mapping not only provides an accurate data basis for the subsequent scheduling algorithm preview, but also builds a solid bridge for intelligent monitoring, equipment maintenance and safety management, thereby promoting the entire smart kitchen system to move forward in the direction of efficiency, precision and intelligence.

[0019] Step 4: Twin interface construction

[0020] After completing the virtual-real mapping, the present invention further constructs a digital twin control interface to achieve intuitive monitoring and interactive operation of each key link in the kitchen. The twin interface integrates multiple modules, including dish status display, robotic arm status monitoring, ordering interface, dish quantity statistics interface, and a dedicated operation interface for controlling the robotic arm rails and woks. The interface first displays the operating status of each dish in the preparation, transportation, frying and serving links through dynamic heat maps and real-time dynamic charts, so that managers can quickly grasp the overall progress and status of the on-site operation; the robotic arm status monitoring module presents the robot's motion trajectory, grasping status and task execution in real time to ensure the safe and stable operation of the equipment. At the same time, the ordering interface provides users with a convenient order entry, and the dish quantity statistics interface automatically summarizes the dish serving data of each link to achieve accurate management of inventory and production conditions. For the control of the robotic arm rails and woks, this system has designed a special operation interface, through which managers can send real-time control instructions to dynamically adjust and maintain the equipment. The construction of the overall twin interface not only improves the transparency and real-time performance of kitchen management, but also provides an intuitive and interactive platform for intelligent scheduling and remote control, providing strong technical support for catering companies to achieve efficient and standardized management.

[0021] Step 5: Preview of the scheduling algorithm in the twin environment

[0022] After completing the construction of high-fidelity physical models, identification coding, virtual-real mapping and twin interface construction, the present invention further deploys the scheduling algorithm derived from the theory into the digital twin environment for preview simulation. Through this preview process, the system can simulate the scheduling of dishes in the preparation, transportation, frying and serving of dishes in real time on the virtual platform, thereby verifying the feasibility and time matching of the predetermined scheduling scheme. Specifically, during the preview process, the scheduling algorithm drives the virtual models of various equipment modules (such as robotic arms, slide rails, woks, etc.) to work according to the established task sequence and time nodes, and observes whether the dishes can be scheduled in sequence according to the predetermined plan and whether the tasks can be completed within the specified time. At the same time, the system analyzes and visualizes the equipment operation status, task delays and potential collision risks in the scheduling process through real-time data monitoring. If equipment path conflicts, time scheduling mismatches or other abnormal situations are found in the preview, the system will automatically record and feedback relevant information for subsequent adjustment and optimization of algorithm parameters. Through comprehensive rehearsal in a twin environment, the present invention significantly reduces the trial and error costs in real physical scenarios, and provides a scientific basis and reliable guarantee for the optimization of scheduling strategies before actual deployment, thereby ensuring that the final smart kitchen system can operate efficiently, collaboratively and stably.

[0023] The above content elaborates on the overall technical scheme and implementation steps of the present invention in constructing a digital twin smart kitchen supervision and optimization system, from high-fidelity physical model construction, unified identification coding, to virtual-reality mapping, twin interface construction, and scheduling algorithm preview, forming a complete set of data-driven smart kitchen management processes. In this process, the present invention not only achieves an effective breakthrough in the problems of insufficient supervision of the entire process of traditional kitchens and delayed real-time warnings, but also provides a solid technical guarantee for the efficient and coordinated operation of the system. Based on the above main contents, the present invention exhibits a number of innovative points as follows:

[0024] 1. This invention focuses on the application of digital twin technology in the optimization of smart kitchen supervision, with an emphasis on model building, real-time monitoring, virtual-reality mapping, algorithm optimization, and system integration, rather than simply reproducing physical scenes, thereby realizing full-process, data-driven smart kitchen supervision.

[0025] 2. The present invention realizes full-process data tracking and tracing from food ingredients, equipment to fried products by making a detailed division into the four links of food preparation, transportation, frying and serving, and uniformly identifies and codes the equipment and operating elements involved in each link, thereby ensuring the accuracy and completeness of data collection.

[0026] 3. The present invention has designed a twin control interface that integrates dish status display, robotic arm status monitoring, ordering and dish statistics, and equipment control, so that managers can remotely monitor and dynamically operate each key link of the kitchen through an intuitive visualization platform, thereby improving overall supervision efficiency and operational standardization.

[0027] 4. The present invention innovatively proposes an algorithm preview mechanism, which optimizes the multi-task and multi-pot scheduling scheme by previewing the scheduling algorithm in a digital twin environment. This mechanism combines the greedy algorithm, priority algorithm and ant colony algorithm, and can verify the rationality and time matching of each task in real time on a virtual platform, and identify potential collision risks in advance. In this way, not only the trial and error cost in the real physical environment is significantly reduced, but also the food delivery efficiency and the overall stability of the system are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Smart kitchen physical system

[0029] Figure 2 Smart Kitchen Digital Twin System

[0030] Figure 3 Physical-twin connection diagram

[0031] Figure 4 Digital Twin Control System Interface

[0032] Figure 5 Scheduling algorithm preview flow chart DETAILED DESCRIPTION

[0033] In order to facilitate those skilled in the art to understand and implement the present invention, the present invention is further described in detail and in depth below with reference to examples.

[0034] This paper focuses on five core modules: high-fidelity physical model design, identification resolution and coding design, virtual-reality mapping, twin interface construction, and algorithm scheduling algorithm preview. On the basis of the physical prototype, an intelligent kitchen supervision system based on digital twin technology is constructed. Figure 1 and Figure 2As shown. First, high-end modeling software such as 3DMAX and SOLIDWORK are used to perform geometric, physical, behavioral and rule-based four-dimensional modeling of key modules such as clean vegetable cabinets, robotic arms, slide rails, cooking pots and ingredients to ensure that the virtual model is highly consistent with the real physical equipment. Subsequently, through the unified information identification parsing and coding technology, various types of equipment and operating parameters in the four links of food preparation, transportation, frying and serving are encoded in detail, and the generated unique identification code is uploaded to the industrial Internet identification resolution platform to achieve seamless data docking. Next, a real-time data transmission interface is established to complete the virtual-real mapping of physical equipment and the digital twin environment, so that data such as temperature and humidity, motion trajectory, and equipment status can be fed back to the twin platform in real time. Based on this, a twin control interface integrating dish status display, robotic arm monitoring, ordering and statistics, and equipment control is constructed to provide managers with an intuitive and interactive remote supervision platform. Finally, the present invention innovatively proposes an algorithm preview mechanism, which integrates the greedy algorithm, priority algorithm, SJF algorithm and ant colony algorithm, and verifies the rationality, time matching and potential collision risks of each task through preview simulation of the scheduling plan in a digital twin environment, thereby significantly reducing the cost of trial and error in the real environment and improving the overall food delivery efficiency and system stability.

[0035] Step 1: High-fidelity physical model construction

[0036] In the digital twin system, the construction of a high-fidelity physical model is a key step in achieving virtual-real integration. This process mainly starts from four aspects: geometric model, physical model, behavioral model and rule model.

[0037] Step 101: Geometry model construction

[0038] First, in the geometric model construction stage, high-end modeling software such as 3DMAX and SOLIDWORK are used to accurately restore the appearance, structure, size and spatial layout of key equipment in the kitchen (such as clean vegetable cabinets, robotic arms, slides, cooking pots and food storage areas). By collecting precise measurement data from the equipment, a three-dimensional geometric model that is highly consistent with the actual environment is established to ensure that each component presents a true proportion and structure in the digital twin platform, thereby providing accurate basic data for subsequent modeling at all levels.

[0039] Step 102 Physical model construction

[0040] In terms of physical model construction, emphasis is placed on simulating the material properties, mechanical characteristics, and motion parameters of the equipment. Here, it is necessary not only to define physical parameters such as the mass, friction coefficient, and elastic modulus of each device, but also to simulate the physical responses of the equipment under different working conditions, such as force, vibration, and energy transfer, through numerical simulation techniques. By introducing a physical engine, the virtual model can achieve response behaviors similar to those of real devices, such as the inertial effect of the robotic arm during the handling process and the heat conduction situation of the frying pan during the heating process, so as to ensure that the physical performance of the equipment in the digital twin environment is basically consistent with the actual situation and provide real physical feedback for intelligent scheduling and rehearsal.

[0041] Step 103: Construction of the behavior model

[0042] In the process of constructing the behavior model, the focus is on modeling the dynamic operation process of the equipment. This process collects the dynamic data generated by the equipment during actual operation, analyzes the motion trajectories, operation modes, and response characteristics of the equipment under different working conditions, and then constructs the behavior model of the equipment. For example, during the grasping, moving, and placing processes of the robotic arm, its motion trajectory, acceleration, and pause time are recorded and mapped into the virtual model; similarly, behaviors such as firepower adjustment, stir-frying actions, and temperature changes during the cooking process are also incorporated into the model description. In this way, the digital twin system can not only statically reproduce the appearance of the equipment, but also dynamically restore the behavior characteristics of the equipment during the operation process, realizing real-time simulation and monitoring of the entire operation process.

[0043] Step 104: Construction of the rule model

[0044] In the stage of constructing the rule model, it mainly involves setting the operation process, interaction rules, and safety constraints of the equipment. The rule model constructs a complete set of constraint and guidance systems by defining the operation sequence, data interaction logic, and safety working limits between devices. For example, in the cooking process, the operation connection sequence between the vegetable cleaning cabinet, the robotic arm, and the frying pan is specified, and the time nodes and operation thresholds of each stage are clarified; at the same time, safety rules such as anti-collision, temperature upper limit, and operation emergency shutdown are also embedded in the model to ensure that in the digital twin environment, the system can automatically identify and feedback abnormal situations. By establishing the rule model, the digital twin system realizes the standardized management of the operation behavior of the equipment, making the scheduling, monitoring, and rehearsal in the virtual environment meet the standard requirements of actual operation.

[0045] In summary, through the comprehensive modeling in four aspects of geometry, physics, behavior, and rules, the present invention not only accurately reproduces the actual kitchen environment in terms of the external structure, but also constructs a highly simulated virtual model in terms of physical response, dynamic behavior, and operation specifications. This high-fidelity physical model provides a solid technical foundation for subsequent virtual-real mapping, real-time monitoring, intelligent scheduling, and pre-simulation, ensuring that the digital twin system can achieve efficient monitoring and optimized management of the entire process of the intelligent kitchen with the most authentic and reliable data support.

[0046] Step 2: Design of Identification and Resolution Encoding

[0047] Through the construction of a complete set of identification and resolution systems, the present invention realizes the accurate traceability and efficient scheduling of the entire process from raw materials to finished products, ensuring the high quality and transparency of dish production. The system divides the entire cooking process into four links, namely ingredient preparation, transportation, frying, and serving, and uniformly identifies and encodes the equipment involved in each link, forming a complete traceability chain for real-time monitoring and quick backtracking.

[0048] In the ingredient preparation link, the main equipment includes food storage equipment, vegetable cleaning cabinets, and related sensor networks. In this link, the detailed information of each ingredient is recorded through ingredient identification, such as data on the source, type, storage location, storage time, real-time quantity, and shelf life. Each piece of data is strictly encoded and uploaded to the identification and resolution platform to ensure that the system can dynamically adjust the ingredient usage order according to the inventory and shelf life, realizing accurate scheduling and automatic replenishment. The detailed records of ingredient identification not only provide data support for the efficient management of the ingredient preparation link but also lay a solid foundation for information traceability in subsequent links.

[0049] The equipment mainly involved in the transportation link is the robotic arm and the slide rail system, which are responsible for transferring the prepared ingredients from the vegetable cleaning cabinet to the frying link. In this link, the grasping, movement, positioning of the robotic arm and the transportation state of the slide rail are all identified through a unique coding system. Each robotic arm and slide rail device is attached with real-time status monitoring sensors, and the coding information details the operating status, movement trajectory, grasping actions, and related time nodes of the equipment. Through this coding information, the system can grasp the dynamic changes of the ingredients during transportation in real time, ensuring that the ingredients will not affect the overall operation efficiency due to delays or incorrect operations during transfer.

[0050] The frying process is the core of the entire cooking process, involving key parameters such as the frying pan and its related temperature control, heat intensity, and frying duration. In this process, each frying pan is assigned a unique identification code, which details its real-time temperature, heat setting, frying stage, and specific frying time. Additionally, sensors monitor the temperature distribution and operation process within each frying pan, and bind this data with the ingredient information to form a complete traceability chain. The frying pan identification not only ensures precise control of each frying action but also provides data support for food safety, enabling quick identification of the root cause of problems in case of anomalies.

[0051] The dish serving process mainly focuses on the final presentation and quality statistics of the dishes. In this process, all information about the ingredients and the frying pan is integrated through the identification of the finished fried dishes. Before each dish is served, key information such as the source of ingredients, operation data at each stage, frying pan number, heat control, and serving time is detailedly recorded and encoded. Consumers and management personnel can intuitively understand the production process and raw material information of each dish by querying the identification of the finished fried dishes, thus greatly enhancing the transparency and safety of food quality. In addition, the statistical coding in the dish serving process is also used to monitor the quantity and serving efficiency of dishes in real-time, providing a basis for inventory management and scheduling optimization.

[0052] The entire identification resolution system generates unique identification codes for the ingredient identification, frying pan identification, and identification of the finished fried dishes generated in each process through a unified identification registration system, and uploads them to the industrial Internet identification resolution platform, realizing seamless docking between physical devices and the digital environment. Through this system, the system not only achieves precise tracking of the devices in each process but also can real-time grasp the operating status and operation data of each device in the kitchen, ensuring that in case of food safety problems, it can quickly trace back to the specific process, device, or operation record.

[0053] The industrial Internet identification resolution application service platform has relatively simple and flexible coding rules for more convenient and rapid access to more enterprises. It mainly consists of an identification prefix, a separator, and an identification suffix. The identification prefix is coded by the secondary node platform, including the country code (fixed), industry code (fixed), and enterprise code; the separator " / " is used to separate the identification prefix and the identification suffix; the identification suffix is completely customized by the enterprise, with diverse rules, unique coding, certain meanings, and no duplicates.

[0054] The identification prefix consists of three parts: the national top-level node code, the identification resolution secondary node code, and the enterprise code, separated by the character '.' in the middle and ending with the character " / ". The national top-level node code for identification resolution is "88", the secondary node code for the food industry in identification resolution is "109", and the enterprise node code used in this invention is "80044833".

[0055] The encoding rule of the identification suffix adopts the form of "meal preparation equipment code.meal preparation link code.relevant element code". The specific encoding structure is: "88.109.80044833 / meal preparation equipment code.meal preparation link code.relevant element code", as shown in Table 1. Among them, the meal preparation equipment code increases sequentially starting from "001" according to the number of equipment. The meal preparation link code is a fixed value, including four main links: meal preparation link, dish transportation link, frying link, and meal serving link, corresponding to the codes 001 to 004 respectively. Information such as ingredients, frying pan equipment, and fried finished products is in the relevant element code.

[0056] Table 1 Intelligent Kitchen Equipment Identification Resolution Encoding Table

[0057]

[0058]

[0059]

[0060] Step 3 Twin - Physical Environment Virtual - Reality Mapping

[0061] The implementation process of twin - physical environment virtual - reality mapping mainly includes four key steps: data collection, transmission, parsing, and mapping, as Figure 3 shown, aiming to build a digital twin platform that can reflect the state of the physical environment in real - time. First, in the data collection stage, by deploying temperature and humidity sensors, photoelectric sensors, cameras, and other Internet of Things devices at key nodes in the kitchen (such as vegetable cleaning cabinets, robotic arms, slide rails, frying pans, etc.), the operating status of the equipment, environmental parameters, and operation process data are collected in real - time. These sensors can not only collect static parameters such as geometric dimensions and temperatures, but also capture dynamic data such as the movement trajectory of the robotic arm, the temperature distribution in the frying pan, and the frying time, forming an all - around data collection network.

[0062] Next, in the data transmission stage, various types of collected data are transmitted to the industrial Internet identification resolution platform through standardized data interfaces and wireless / wired communication networks. In this process, to ensure the real - time and accurate transmission of data, the system adopts a low - latency, high - stability transmission protocol, and at the same time pre - processes and converts the data format to meet the requirements of the industrial Internet identification resolution platform. During the data transmission process, timestamps and unique identification codes are also embedded to ensure the precise matching and traceability of subsequent data.

[0063] In the data parsing stage, the central data processing platform performs real-time parsing and processing on the transmitted raw data. The platform uses a specially designed data parsing module to match the data uploaded by each device with the corresponding information in the pre-established digital model. For example, through the unique identification code attached to the device, the system can accurately bind data such as temperature, movement trajectory, and operation status to each virtual device in the corresponding digital twin model. At the same time, the platform also filters, corrects, and integrates the data to eliminate noise and outliers, ensuring that the information in the virtual environment is highly consistent with the physical environment.

[0064] After completing the data parsing, the last step is the virtual-real mapping, which maps the parsed data to the virtual model in the digital twin platform in real time. This mapping process relies on a carefully designed data transmission interface and a real-time synchronization mechanism to ensure that each device and each parameter in the digital twin environment can reflect the operating state of the actual physical device in real time. Through the dynamic data overlay technology, the virtual model not only presents the static structure of the device but also intuitively shows the dynamic changes during the operation of the device, such as the grasping action of the robotic arm, the temperature fluctuation in the cooking pot, and the dish transfer status. After the mapping is completed, the management personnel can monitor the operating state of the entire kitchen in real time through the digital twin control interface and trigger the warning mechanism when necessary to intervene and adjust abnormal situations.

[0065] Generally speaking, the implementation process of the virtual-real mapping between the twin and physical environments ensures the information consistency and real-time nature between the physical device and the digital model through four steps: data acquisition, transmission, parsing, and real-time mapping, providing a solid data foundation and technical support for the efficient monitoring, scheduling optimization, and safety management of the intelligent kitchen system.

[0066] Step 4 Twin Interface Construction

[0067] The digital twin interface, based on the industrial Internet identification and resolution technology, creates a comprehensive intelligent cooking platform integrating order management, real-time monitoring, and simulation optimization. Specifically, as Figure 4 shown. This interface not only realizes the real-time acquisition and accurate transmission of key cooking parameters but also uses digital twin technology to perform fast and faithful simulations of the entire cooking process, including the grasping of ingredients by the robotic arm, the dispensing of ingredients, and the frying process of the ingredients in the wok, ensuring the transparency, intelligence, and standardization of the entire cooking process. At the same time, the operator can also perform real-time control and adjustment on the position of the robotic arm, the grasping action, and the operating state of the wok through the interface.

[0068] In terms of the dynamic management of dishes and ingredients in Step 401

[0069] In terms of the dynamic management of dishes and ingredients, the system designs an intuitive ordering interface where users can view detailed simulation demonstrations of dishes such as stir-fried broccoli and stir-fried pork with peppers. Relying on the industrial Internet identification and resolution technology, the raw materials required for each dish are accurately tracked, the inventory data and status information of each ingredient are collected in real time, and the preparation progress of the dishes is monitored synchronously. Managers can intuitively understand the remaining amount of ingredients, shelf life, and allocation situation through the interface, promptly discover problems such as insufficient inventory or production delays, and thus respond quickly to ensure the smooth progress of the dish production process.

[0070] Step 402 Remote Scheduling and Operation of Each Component

[0071] In terms of the remote scheduling and operation of each component, the interface provides a dedicated operation control module to achieve the remote monitoring and scheduling of the robotic arm and the wok. Specifically, in the case of the robotic arm, the operator can view the current position of the robotic arm, the progress of the task being executed, and its movement trajectory in real time through the control panel; the system supports manual or automatic scheduling modes, and users can issue precise instructions for the actions of the robotic arm such as material picking, stir-frying, ingredient adding, and dish plating to ensure that each operation conforms to the predetermined process. In terms of wok control, operation buttons such as standby, start, serving, and end are set on the interface, enabling the operator to accurately control the start and stop of the wok when needed, thus ensuring that the dish is cooked in the best state.

[0072] Step 403 Real-time Monitoring

[0073] The real-time monitoring of the cooking process is another important function of the digital twin interface. This function module not only focuses on the movement trajectory, current task progress, and scheduling status of the robotic arm but also synchronously displays the progress of dish production and quality assessment data. All links in the entire cooking process are presented on the interface in the form of real-time data streams, enabling the operator to comprehensively grasp the operation of the entire process. The system not only monitors traditional cooking parameters such as temperature and heat but also can capture the detailed changes in micro-actions such as the robotic arm's grasping and stir-frying, thus achieving precise management of every detail in the cooking process. When the system detects abnormal situations such as abnormal temperature, robotic arm jamming, grasping errors, or ingredient shortages, the alarm information bar will immediately issue a warning prompt to remind the administrator to intervene and handle it in a timely manner to ensure that the entire cooking process remains efficient, stable, and safe.

[0074] Generally speaking, the construction of the digital twin interface not only greatly improves the transparency and real-time performance of the smart kitchen, but also provides an intuitive and interactive platform for remote scheduling and refined operation. By integrating order management, dynamic data collection, remote device control, and full-process simulation, the entire system achieves precise monitoring and management of key cooking parameters, ensuring the standardization and intelligence of the entire process from raw material warehousing to dish serving. This comprehensive platform provides strong technical support for catering enterprises, contributing to the digital upgrade of kitchen management and the full-process guarantee of food safety.

[0075] Step 5 Pre-demonstration of the Scheduling Algorithm in the Twin Environment

[0076] The implementation process of pre-demonstrating the scheduling algorithm in the digital twin environment is a multi-level and iterative optimization process. Its core purpose is to fully simulate the side dish task scheduling situation of multiple dishes in the actual kitchen on the virtual platform, thereby verifying and optimizing the scheduling plan to ensure efficient and conflict-free operation in the real physical scenario. The specific flowchart is as Figure 5 shown. The entire pre-demonstration process mainly includes the following steps:

[0077] Step 501 Mapping of Side Dish Tasks and Status

[0078] First, after establishing a complete mathematical model and task set, the system maps all the dishes to be stir-fried and their respective side dish tasks (such as the stir-frying time of the side dishes in the pot, the start and completion times of the tasks, the preparation and movement times of the robotic arm, etc.) to the digital twin environment. At the same time, key devices such as the wok and the robotic arm are also given real-time status parameters consistent with the physical devices. The digital twin platform uses real-time data collection and device status synchronization mechanisms to ensure a high degree of consistency between the models in the virtual environment and the dynamic status of the actual devices, providing an accurate physical basis and operation constraints for the pre-demonstration of the scheduling algorithm.

[0079] Step 502 Obtaining the Optimal Solution

[0080] In the initial stage of the pre-demonstration, the scheduling algorithm generates a preliminary task allocation plan by simulating the collaborative search behavior of an ant colony. Specifically, each "ant" represents a task allocation path in the digital twin environment, and its construction process strictly follows the process priority constraints to ensure that the side dish tasks of a certain dish must be completed in the established order. The probability of path selection is jointly determined by the pheromone concentration and the heuristic factor. Here, the heuristic information mainly reflects the time efficiency of the task, and the parameter control balances the weights of the pheromone and the heuristic information, thereby guiding the ants to search for the optimal path in the solution space. The expression for the shortest time to complete stir-frying is:

[0081]

[0082] To further enhance the global search ability, the algorithm introduces an adaptive pheromone update mechanism. During the rehearsal process, when the optimal solution fails to improve after consecutive iterations, the system gradually increases the pheromone evaporation factor, which helps break the local optimum constraint and enhance path diversity. The path selection probability is jointly determined by the pheromone concentration and heuristic information:

[0083]

[0084] where τ ij is the pheromone concentration of task t ij , η ij = 1 / (p ij + r ij ) is the heuristic factor, reflecting the time efficiency of the task. Parameters α and β respectively control the weights of pheromone and heuristic information.

[0085] Meanwhile, the top 10% of individuals in each generation are retained through the elite strategy to prevent the loss of high-quality solution information, thereby providing a higher-quality initial solution for subsequent iterations.

[0086] In the local optimization stage, the rehearsal system embeds a simulated annealing mechanism to perturb the current optimal scheduling plan. New solutions are generated through random swapping or insertion operations, and the sub-optimal solution is accepted with probability p = e -cT , where Δc is the cost increment, and the temperature parameter T decays according to the logarithmic cooling schedule (T k = T0 / ln(1 + k)). This mechanism effectively balances the exploration and exploitation capabilities of the algorithm and avoids falling into local optima.

[0087] Step 503 checks for resource conflicts

[0088] When resource conflicts of the robotic arm or wok are detected in the digital twin environment (e.g., multiple tasks are simultaneously assigned to the same device with overlapping time), the system automatically triggers the dynamic rescheduling process. At this time, the rehearsal platform locates the conflicting task pairs through time window scanning and calculates the conflict intensity index to quantify the severity of resource competition. For high-intensity conflicts, the system uses the tabu search strategy to locally rearrange the conflicting tasks. Specific operations include setting a short-term tabu list to prohibit recent operation rollbacks and generating new solutions using neighborhood operations (such as task postponement, device reassignment). The adjusted solution is fed back into the main algorithm, and then the guiding role of high-quality paths is strengthened by updating the pheromone concentration: if the total completion time is significantly reduced after conflict resolution, the system will deposit pheromone on the relevant paths, so that these high-quality scheduling plans will obtain higher search priorities in subsequent iterations.

[0089] The preview process of the entire scheduling algorithm is carried out in a high-fidelity simulation manner on the digital twin platform. The platform shows the start, execution, and completion of each task in real time, while recording key parameters such as the operating status of each device (such as the robotic arm and the wok), the idle time of the device, and the resource utilization rate. Operators and system engineers can intuitively observe the execution effect of the scheduling scheme through the virtual interface, timely discover potential bottlenecks or conflicts, and adjust the algorithm parameters based on the feedback data.

[0090] In summary, by previewing the improved ant colony algorithm in the digital twin environment, the system not only realizes the real-time simulation and verification of the task allocation scheme, but also continuously optimizes the scheduling strategy through mechanisms such as adaptive pheromone update, simulated annealing, and tabu search. This preview mechanism effectively reduces the cost of trial and error in the real physical environment, ensures that the final scheduling scheme can achieve the goal of minimizing the total completion time under the condition of limited device resources, and improves the meal delivery efficiency and stability of the entire system.

Claims

1. An intelligent kitchen supervision optimization system based on digital twins, characterized in that, The system includes: 1) A high-fidelity physical model building module, which is used to perform four-dimensional modeling of geometry, physics, behavior, and rules for key equipment in the kitchen (including clean food cabinets, robotic arms, slide rails, cooking pots, and food storage areas) to achieve accurate reproduction of real kitchen scenes; 2) Identification resolution and coding module, which is used to uniformly identify and encode the equipment and operating elements involved in the four links of food preparation, transportation, cooking and serving, and upload the generated unique identification code to the industrial Internet identification resolution platform to achieve seamless data connection between physical equipment and digital environment; 3) Virtual-reality mapping module, which maps the status data of physical equipment (including temperature and humidity, motion trajectory, equipment working status, etc.) to the digital twin platform in real time by establishing a data transmission interface and real-time synchronization mechanism; 4) Digital twin control interface module, which is used to intuitively display the real-time operating status of each key link in the kitchen and supports remote monitoring, equipment control and data visualization; 5) The scheduling algorithm preview module performs a virtual preview of the multi-task and multi-pot scheduling scheme by improving the ant colony algorithm combined with simulated annealing, adaptive pheromone update and taboo search mechanism to verify the rationality of the tasks and time matching, and minimize the total completion time.

2. The system according to claim 1, wherein The high-fidelity physical model building module further includes: 1) The geometric model construction submodule uses modeling software such as 3DMAX and SOLIDWORK to collect the size, structure and spatial layout data of each device and generate a three-dimensional geometric model; 2) Physical model construction submodule, which defines physical parameters such as the mass, friction coefficient, elastic modulus, etc. of the equipment, and uses the physical engine to perform numerical simulation on the force, vibration and energy transfer of the equipment; 3) The behavior model construction submodule collects the dynamic operation data of the equipment, establishes the action model of the equipment under different working conditions, and realizes the real-time reproduction of dynamic behavior; 4) The rule model construction sub-module sets the equipment operation process, interaction rules and safety constraints to ensure that the operations in the digital twin environment meet the actual standards.

3. The system according to claim 1, wherein The identification resolution coding module adopts the structure of "national top-level node code. Second-level node code. Enterprise node code / catering equipment code. Catering link code. Related element code" to uniquely identify information such as ingredients, robotic arms, woks, and fried products in the four links of food preparation, transportation, frying, and serving, thereby realizing data tracking and traceability throughout the entire process.

4. The system according to claim 1, wherein The virtual-reality mapping module uploads the real-time data collected by physical devices to the Industrial Internet identity resolution platform through standardized data interfaces and low-latency communication protocols after preprocessing, format conversion and timestamp marking, and uses the resolution module to match and dynamically map the data with the corresponding devices in the digital twin model in real time.

5. The system according to claim 1, wherein The digital twin control interface module integrates multiple sub-interfaces, including the dynamic management interface for dishes and ingredients, which is used to display in real time the production progress, raw material inventory, shelf life, and allocation status of each dish; the robotic arm status monitoring interface, which presents in real time the position, grasping status, motion trajectory, and task execution progress of the robotic arm; the device remote control interface, which supports the remote scheduling and operation of the robotic arm guide rail and the wok status; and the data visualization interface, which displays real-time data streams and anomaly warning information through dynamic heat maps, dynamic charts, etc.

6. The system according to claim 1, wherein The system pre-acts the scheduling algorithm in the digital twin environment, shows through real-time simulation the scheduling situation of each task in the processes of ingredient preparation, transportation, stir-frying, and meal serving, and provides real-time feedback on the device status, task delays, and potential collision risks, providing an intuitive basis for operators to evaluate the scheduling plan and adjust parameters, and ultimately achieving the goal of minimizing the total completion time.

Images