Intelligent ship kitchen system based on artificial intelligence

Abstract

The invention discloses a ship intelligent kitchen system based on artificial intelligence, and the system comprises an intelligent cooking control module which monitors the inventory information of food materials in real time, dynamically adjusts the purchasing, storage and consumption strategies of the food materials, optimizes the cooking parameters, and controls intelligent cooking equipment. The customized meal supply module provides a meal ordering terminal supporting multi-mode interaction and customizes recipes for personnel; the health real-time monitoring module monitors the health condition in real time and realizes health risk early warning; the personnel dining scheduling module plans the dining time period of the personnel and guides the personnel to have meals in real time; the restaurant environment adjusting module automatically adjusts and controls light color temperature, temperature and humidity and background media; and the energy management module formulates an energy-saving control strategy, dynamically executes power regulation and drives waste heat and waste water to be recycled. By integrating the AI algorithm, the Internet of Things and the sensor technology, various problems of food material waste, single dish, hysteretic health monitoring and low energy efficiency of a traditional ship kitchen are solved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to an artificial intelligence-based intelligent ship kitchen system. Background Art

[0002] With the development of the shipping industry, the requirements for the comprehensive level of ship kitchens, service quality, and dining experience of personnel are increasing. Traditional ship kitchen systems have many problems. For example, in the food management link, the manual recording method is inefficient and prone to errors, which often leads to inaccurate inventory data and frequent waste of expired ingredients, resulting in unnecessary economic losses. In terms of catering production, due to the lack of accurate grasp of individual taste requirements, the dishes are monotonous and the flavors are fixed, making it difficult to meet the diverse taste preferences of different personnel, greatly reducing dining satisfaction. In addition, with the increasing complexity of ship shipping missions, diverse personnel arrangements, and uncertain sailing cycles, traditional kitchen systems are unable to flexibly adjust meal supply strategies based on these dynamic factors and are unable to meet the special needs of crew members under special working conditions.

[0003] A Chinese patent with publication number CN120010294A discloses "an intelligent kitchen system for ships," which receives ordering information from crew members through an intelligent ordering system and transmits the ordering information to a central control system; an intelligent inventory management system provides the type, quantity, and weight information of inventory ingredients to the central control system, and receives control information from the central control system to prepare ingredients; and an intelligent cooking system receives ingredients from the intelligent inventory management system and receives control information from the central control system to cook.

[0004] The above solution has initially achieved unmanned cooking and basic food management, but it still has the following limitations: 1) Food management relies solely on static weighing and visual recognition, and cannot combine dynamic factors for demand forecasting and inventory optimization; 2) Although it supports historical dish recommendations, it cannot customize special meals; 3) There is a lack of full-process hygiene monitoring, and it cannot dynamically prevent food safety risks; 4) Energy management is weak, and energy consumption optimization relies solely on equipment power regulation. Therefore, it is urgent to design an artificial intelligence-based ship intelligent kitchen system to solve the problems existing in the above-mentioned existing technologies. Summary of the Invention

[0005] In response to the above-mentioned defects or improvement needs of the existing technology, the present invention provides an artificial intelligence-based ship intelligent dining and kitchen system, which aims to provide dynamically optimized dining services according to the actual operating conditions of the ship, and provide an efficient, convenient, healthy and personalized dining environment for the personnel on the ship, while reducing operating costs, improving energy utilization efficiency, and enhancing the overall living facilities level of the ship.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A first aspect of the present invention provides an artificial intelligence-based intelligent ship kitchen system, comprising: The intelligent food management module is used to monitor food inventory information in real time and dynamically adjust food procurement, storage and consumption strategies based on multi-dimensional data using AI algorithms; Intelligent cooking control module: optimizes cooking parameters through supervised learning models and automatically controls intelligent cooking equipment based on cooking parameters; Customized meal supply module: used to provide a meal ordering terminal that supports multimodal interaction and customize recipes based on employee health data and job requirements; Real-time hygiene monitoring module: A triple monitoring network is built using image recognition cameras, microbial sensors, and odor sensors to capture real-time changes in operator behavior, microbial content, and odor, enabling early warning of hygiene risks. Personnel dining scheduling module: It is configured to connect to the ship's personnel management system, plan dining time periods based on the collected crew shift data, and guide personnel to eat in real time; Restaurant environment adjustment module: automatically adjusts lighting color temperature, temperature and humidity, and background media based on ship sailing time, number of diners, and real-time activity needs; Energy management module: used to monitor energy consumption data in real time, formulate energy-saving control strategies, dynamically perform power regulation and drive the recycling of waste heat and wastewater.

[0007] As an embodiment of the present application, the food intelligent management module includes an inventory management unit, a food planning unit, a quality inspection unit and a food traceability unit; The inventory management unit uses weight sensors, temperature sensors, humidity sensors, and RFID readers to monitor the inventory quantity, shelf life, and storage conditions of food in real time. Based on the ship's sailing plan, crew numbers, and historical consumption data, it predicts demand trends for food and automatically generates a procurement plan. The food planning unit adjusts the storage ratio of various food items based on the ship's voyage planning cycle and the real-time updated remaining sailing time, according to the shelf life of the food items, inventory dynamics, and nutritional balance requirements. At the same time, the food consumption schedule is planned based on the remaining sailing time, and a food consumption curve is generated based on the shelf life and storage conditions of different food items. The quality detection unit uses spectral analysis and image recognition technology to detect the quality of food. If the quality of the food is found to be unqualified, an alarm is automatically triggered and the food is marked and isolated. The food traceability unit uses QR codes and RFID readers to record the source, purchase time, transportation process, and storage conditions of the food, thereby establishing a complete food traceability system.

[0008] As an embodiment of the present application, the intelligent cooking control module includes a cooking parameter optimization unit, an intelligent device control unit, and a dish innovation unit; The cooking parameter optimization unit collects multi-dimensional data to form a structured data set, uses machine learning technology to extract key features of the structured data set, uses a supervised learning model for training, and continuously updates the supervised learning model through an online learning algorithm to dynamically optimize cooking parameters; The intelligent cooking device control unit realizes two-way real-time data interaction with the system by installing an Internet of Things controller on the intelligent cooking device, and automatically completes the cooking task according to the preset cooking parameters; The dish innovation unit automatically generates new dish combinations and cooking plans based on the inventory of ingredients, the taste preferences and nutritional needs of the personnel.

[0009] As an embodiment of the present application, the smart cooking device is also equipped with multiple types of sensors for real-time monitoring of the operating parameters of the device. The operating parameters of the device are analyzed in real time through a machine learning model to predict potential failure risks of the device and trigger an early warning.

[0010] As an embodiment of the present application, the customized meal supply module includes a meal ordering unit, a diet management unit, and a special meal customization unit; The order reservation unit provides an order terminal that supports multimodal interaction; The diet management unit provides personalized diet recommendations and nutrition plans based on the individual's physical condition, exercise status, and nutritional needs; The special meal customization unit customizes special meals according to the special needs of various special workers.

[0011] As an embodiment of the present application, the customized meal supply module also includes a robot service unit, which uses a service robot with intelligent navigation, voice interaction, and object recognition functions to be responsible for the delivery of dishes and cleaning of the table. As an embodiment of the present application, the real-time hygiene monitoring module includes a personnel operation behavior monitoring unit and an abnormal food material management unit; The personnel operation behavior monitoring unit uses an image recognition camera combined with image recognition technology and deep learning algorithms to capture the operation behavior of kitchen staff in real time; The food abnormality management unit monitors the microbial content and odor changes in real time through a microbial sensor and an odor sensor respectively.

[0012] As an embodiment of the present application, the personnel dining scheduling module includes a shift information collection unit, a dining time planning unit and a dining guidance unit; The shift information collection unit collects crew shift information in real time through a data interface; The dining time planning unit uses a mixed integer programming algorithm based on the collected crew shift information to minimize waiting time and balance restaurant load, and dynamically adjusts time allocation in combination with real-time passenger flow monitoring; The dining guidance unit is based on real-time restaurant data and user characteristics, and uses dynamic priority queue division and shortest path seat allocation algorithms, combined with the M / M / 1 queuing model to predict waiting time, to achieve multi-objective optimization. It also uses a full-range voice broadcast system and dynamic display screens to guide people on different shifts to dine in an orderly manner in real time.

[0013] As an embodiment of the present application, the restaurant environment adjustment module includes an atmosphere light control unit, a temperature and humidity adjustment unit, and a background media control unit; The atmosphere light control unit automatically adjusts the restaurant atmosphere light according to the dining content, time and real-time activity atmosphere in the restaurant; The temperature and humidity adjustment unit uses a PID control algorithm to dynamically adjust the air conditioning temperature and humidity based on the number of diners and seasonal factors; The background media control unit synchronously controls background music and TV programs based on meal periods, crew physiological data and questionnaire results, synchronizes lyrics and subtitles through the display screen and regularly updates the media library.

[0014] As an embodiment of the present application, the energy management module includes an energy consumption monitoring unit, an energy saving control unit and an energy recovery and utilization unit; The energy consumption monitoring unit monitors the energy consumption data of the kitchen system in real time through the energy monitoring equipment, and transmits the data to the system for analysis and statistics, and locates the peak hours of energy consumption and the main energy-consuming equipment; The energy-saving control unit formulates a corresponding energy-saving control strategy according to the monitoring results of the energy consumption monitoring unit; The energy recovery and utilization unit collects waste heat generated by the cooking equipment through a heat exchanger and heats water, while a purification system is used to treat and purify the kitchen wastewater.

[0015] The beneficial effects of the present invention are: 1. The present invention uses an intelligent food management module to monitor food inventory information in real time, and combines AI algorithms to dynamically adjust food procurement, storage, and consumption strategies based on multivariate data. The intelligent cooking control module optimizes cooking parameters through a supervised learning model and automatically controls intelligent cooking equipment based on the cooking parameters. Intelligent food management and cooking control reduce food waste and cooking time, improve the efficiency of catering production, and ensure the timeliness of catering services. 2. The present invention provides a meal ordering terminal that supports multimodal interaction through a customized meal supply module. Recipes are customized according to employee health data and job requirements to meet the diverse taste preferences of different employees, greatly improving dining satisfaction. 3. The real-time sanitation monitoring module of the present invention establishes a triple monitoring network through an image recognition camera, a microbial sensor, and an odor sensor. This module captures human operating behavior, microbial content, and odor changes in real time, enabling early warning of sanitation risks. This solves the technical problem of existing technologies that rely on regular manual inspections, which have long intervals and cannot detect sanitation hazards in real time, posing a potential threat to food safety. 4. The present invention uses an energy management module to monitor energy consumption data in real time, formulate energy-saving control strategies, dynamically perform power regulation and drive the recycling of waste heat and wastewater, thereby reducing energy consumption and operating costs and improving the economic benefits of the ship. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a general architecture diagram of an artificial intelligence-based ship intelligent kitchen system provided in an embodiment of the present invention; Figure 2 This is a system block diagram of an artificial intelligence-based food material intelligent management module of a ship intelligent kitchen system provided in an embodiment of the present invention; Figure 3 This is a system block diagram of an intelligent cooking control module of an artificial intelligence-based ship intelligent kitchen system provided in an embodiment of the present invention; Figure 4 This is a system block diagram of a customized catering module of an artificial intelligence-based ship intelligent kitchen system provided in an embodiment of the present invention; Figure 5 This is a system block diagram of a real-time hygiene monitoring module for a ship's intelligent kitchen system based on artificial intelligence provided in an embodiment of the present invention; Figure 6 This is a system block diagram of a personnel dining scheduling module of a ship's intelligent dining and kitchen system based on artificial intelligence provided in an embodiment of the present invention; Figure 7 This is a system block diagram of a restaurant environment adjustment module of an artificial intelligence-based ship intelligent kitchen system provided in an embodiment of the present invention; Figure 8 This is a system block diagram of an energy management module for an artificial intelligence-based ship intelligent kitchen system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] Reference Figures 1-8 The first aspect of the present invention provides an artificial intelligence-based ship intelligent kitchen system, comprising: The intelligent food management module is used to monitor food inventory information in real time and dynamically adjust food procurement, storage and consumption strategies based on multi-dimensional data using AI algorithms; Intelligent cooking control module: optimizes cooking parameters through supervised learning models and automatically controls intelligent cooking equipment based on cooking parameters; Customized meal supply module: used to provide a meal ordering terminal that supports multimodal interaction and customize recipes based on employee health data and job requirements; Real-time hygiene monitoring module: A triple monitoring network is built using image recognition cameras, microbial sensors, and odor sensors to capture real-time changes in operator behavior, microbial content, and odor, enabling early warning of hygiene risks. Personnel dining scheduling module: It is configured to connect to the ship's personnel management system, plan dining time periods based on the collected crew shift data, and guide personnel to eat in real time; Restaurant environment adjustment module: automatically adjusts lighting color temperature, temperature and humidity, and background media based on ship sailing time, number of diners, and real-time activity needs; Energy management module: used to monitor energy consumption data in real time, formulate energy-saving control strategies, dynamically perform power regulation and drive the recycling of waste heat and wastewater.

[0019] As an embodiment of the present application, the food intelligent management module includes an inventory management unit, a food planning unit, a quality inspection unit and a food traceability unit; The inventory management unit deploys weight sensors, temperature sensors, humidity sensors, and RFID readers throughout the food storage area of ​​the ship's kitchen to monitor food inventory, shelf life, and storage conditions in real time. It also predicts food demand trends based on the ship's sailing schedule, crew numbers, and historical consumption data, and automatically generates a procurement plan. Specifically, AI algorithms are used to deeply mine and learn from diverse information, including historical consumption data, ship itineraries, seasonal factors, and changes in personnel's eating habits, to accurately predict the usage of different ingredients at different times. Once the food inventory approaches or falls below a preset safety threshold, the system automatically triggers an early warning mechanism and pushes a detailed purchase list to the procurement department's dedicated terminal. The purchase list not only includes the types and quantities of the required ingredients, but also takes into account the convenience and cost of replenishment in combination with the ship's route and port of call information to ensure the timeliness and economy of food supply. The safety threshold is usually set to meet 3-7 days of consumption, depending on the ship's replenishment cycle. For example: for ocean-going ships with a replenishment cycle greater than 15 days, the safety threshold = average daily consumption × 7 days; for near-shore ships with a replenishment cycle ≤ 7 days, the safety threshold = average daily consumption × 3 days. In addition, there are also cases where the type of ingredients differs: for fresh meat or vegetables, due to their short shelf life, the safety threshold is set at average daily consumption × 3 days; for dried goods and canned goods, due to their long shelf life, the safety threshold can be set at average daily consumption × 15 days.

[0020] The system also uses the first-in-first-out (FIFO) principle and ingredient expiration dates to intelligently schedule the order in which ingredients are shipped out of the warehouse for use, minimizing waste caused by overstocking and expired ingredients. For example, for fresh meat, weight sensors accurately monitor remaining quantities. When inventory falls below a preset safety threshold, the system automatically generates a purchase reminder. Temperature and humidity sensors ensure that meat is stored in suitable conditions. Chilled meat is typically stored at 0°C to 4°C, while frozen meat is stored below -18°C. Humidity in the refrigerated area is controlled at 70% to 85%, and in the frozen area at ≤60%. This minimizes condensation and food drying, extending shelf life. The system also analyzes the ship's sailing schedule and historical consumption data. For example, if a ship plans to dock at a specific port, the system will pre-plan a purchase list based on the availability and prices of ingredients near that port. If historical data indicates a high demand for seafood during a particular voyage, the system will increase seafood purchases accordingly.

[0021] Preferably, the AI ​​algorithm in the present application adopts a time series analysis algorithm (such as ARIMA, LSTM) or a machine learning classification / regression model (such as random forest, gradient boosting tree) to process the time series characteristics and multivariate associations of historical consumption data; the AI ​​algorithm performs in-depth mining and learning as follows: first, integrate historical consumption data, such as the daily consumption of each ingredient in the past three months; ship itinerary data, such as port call time and sailing days; seasonal characteristics, such as seasonal fluctuations in ingredient demand; personnel dietary labels, such as regional taste preferences and differences in energy requirements for different types of work; then convert discrete data, such as seasons and ports, into numerical features, such as seasons coded as 1-4, to construct a three-dimensional feature matrix of "ingredients-time-influencing factors"; finally, through supervised learning, fit the mapping relationship between ingredient consumption and influencing factors, for example, use the LSTM model to capture long-term dependencies in time series, such as the increase in hot drink consumption in winter, and use random forests to identify key influencing factors, such as the impact weight of port call frequency on the purchase of fresh ingredients.

[0022] The food planning unit adjusts the storage ratio of various food items based on the ship's voyage planning cycle and the real-time updated remaining sailing time, according to the shelf life of the food items, inventory dynamics, and nutritional balance requirements. At the same time, the food consumption schedule is planned based on the remaining sailing time, and a food consumption curve is generated based on the shelf life and storage conditions of different food items. Specifically, at the beginning of a long voyage, the system fully utilizes the abundant fresh food resources at the departure port and significantly increases the proportion of dishes made with fresh ingredients, allowing the crew and crew to enjoy fresh and delicious food. As the voyage time gradually increases, the system gradually increases the proportion of dishes made with storable ingredients such as canned and dried goods based on the shelf life of the ingredients, inventory dynamics, and nutritional balance needs. For example, in the first half of the voyage, fresh vegetables can account for up to 70% of the total vegetable supply. In the second half, this proportion can be flexibly adjusted to around 30% based on actual conditions, and the use of dehydrated or pickled vegetables can be increased accordingly to ensure a sustainable supply of food during long voyages. At the same time, the food consumption schedule is carefully planned based on the remaining voyage time. In combination with the shelf life and storage conditions of different ingredients, a scientific and reasonable food consumption curve is formulated to avoid the embarrassing situation of shortages of key ingredients in the later stages of the voyage and ensure a stable and reliable food supply throughout the voyage. For example, a ship is carrying out a 30-day ocean voyage mission. It docks at Port A for the first 15 days, with an adequate supply of vegetables, and no port replenishment for the last 15 days. The system plan is as follows: In the first 15 days: the average daily consumption of fresh vegetables was 50 kg, accounting for 70% of the total vegetable supply, accompanied by 21 kg of canned vegetables, accounting for 30%; meat was mainly fresh meat, accounting for 60%, and frozen meat 40%.

[0023] Last 15 Days: Starting on Day 16, fresh vegetable consumption will be reduced to an average of 20kg per day, representing 30% of the total, while dehydrated vegetables will be increased to 47kg, representing 70% of the total. Meat consumption will be adjusted to 70% frozen and 30% canned. Furthermore, depending on the remaining days, for example, if there are 10 days remaining on Day 20, emergency food such as nuts and energy bars should be supplemented at an average of 5kg per day to avoid nutritional gaps later in the journey.

[0024] The quality detection unit uses spectral analysis and image recognition technology to detect the quality of food. If the quality of the food is found to be unqualified, an alarm is automatically triggered and the food is marked and isolated. Specifically, when food ingredients enter the warehouse, staff use spectrometers to test vegetables, fruits, etc., and analyze their nutritional content (such as the correlation between vitamin C content and spectral absorbance) and the presence of pesticide residues (such as the absorption peak of organophosphorus pesticides at specific wavelengths) through reflectance spectral characteristics. For example, when the pesticide residue value of a batch of spinach is detected to exceed the national standard of 0.1mg / kg, the system automatically marks it as "unqualified", triggering an alarm and isolating the batch of food ingredients. For meat and seafood, image recognition technology, such as convolutional neural network (CNN) models, is used to identify indicators such as color, texture, and foreign matter. For example, when abnormal color spots (indicating deterioration) or impurities are detected on the surface of frozen chicken breast, the system immediately issues an alarm and prohibits the food from entering the processing link. At the same time, the system records the supplier information of the batch of food ingredients for subsequent traceability and processing.

[0025] The food traceability unit uses QR codes and RFID readers to record the source, procurement time, transportation process, and storage conditions of food ingredients throughout the entire process, establishing a complete food traceability system. When food safety issues arise, the source of the problematic food ingredients can be quickly traced back and appropriate measures can be taken to ensure food safety for everyone. Specifically, when food is put into storage, staff use handheld RFID readers to scan the labels on the food packaging and enter detailed information about the food, including type, quantity, origin, production date, shelf life, etc., into the system database. Each batch of food has a unique QR code. By scanning the QR code, staff can view detailed information such as the source, procurement time, transportation process, and storage conditions of the food. If a food safety problem occurs on board the ship, such as some crew members showing symptoms of food poisoning, staff can quickly trace the source of the problem food by scanning the QR code on the remaining food and take appropriate measures to deal with it, such as recalling the same batch of food, notifying suppliers, etc. This not only ensures the food safety of the crew, but also improves the ship operator's ability to respond to food safety issues.

[0026] As an embodiment of the present application, the intelligent cooking control module includes a cooking parameter optimization unit, an intelligent device control unit, and a dish innovation unit; The cooking parameter optimization unit collects multi-dimensional data to form a structured data set, uses machine learning technology to extract key features of the structured data set, uses a supervised learning model for training, and continuously updates the supervised learning model through an online learning algorithm to dynamically optimize cooking parameters; Specifically, the cooking parameter optimization unit collects historical cooking records, such as the temperature curve, heating time, and seasoning ratio when frying steak; finished dish evaluations, such as taste scores and color scores; and food characteristic data, such as fiber density and thermal conductivity of different parts of beef, to form a structured data set. Machine learning techniques, such as principal component analysis (PCA), are used to extract key features. For example, "ingredient type + weight + cooking method" is combined as input features, and "optimal temperature range + time + seasoning amount" is used as output labels. Supervised learning models, such as random forests and neural networks, are used for training. Historical cooking parameters and corresponding dish evaluation data are input to output the optimized parameter combination. For example, through training data, it is learned that "when frying sirloin steak (2cm thick), high temperature lock in juice for 2 minutes + frying at 180℃ for 3 minutes" can increase the tenderness score by 20%; new cooking data is collected in real time, such as parameters manually adjusted by the chef and experimental parameters of new dishes, and the model is continuously updated through online learning algorithms to adapt to dynamic factors such as changes in food freshness and equipment performance differences.

[0027] Specifically, the system analyzes and learns from extensive cooking data to precisely adjust cooking parameters for different ingredients and dishes. For example, when cooking steak, the system automatically adjusts the stovetop temperature and cooking time based on the type and thickness of the steak, as well as the crew member's desired doneness. During the cooking process, the stovetop monitors the steak's internal temperature in real time and automatically stops heating when the preset temperature is reached, ensuring a tender and juicy steak. The system also optimizes the amount of seasoning based on different cooking methods, such as pan-frying, grilling, and deep-frying, to ensure a more balanced flavor.

[0028] The intelligent cooking equipment control unit realizes two-way real-time data interaction with the system by installing an Internet of Things controller on the intelligent cooking equipment, and automatically completes cooking tasks according to preset cooking parameters; among them, intelligent controllers with Internet of Things functions are installed on smart devices such as stoves, ovens, steamers, dishwashers, etc. in the ship's kitchen. These controllers are stably connected to the main server of the ship's intelligent kitchen system through wireless or wired networks to realize two-way real-time data interaction. The system uses intelligent algorithms to automatically and accurately adjust key operating parameters of the equipment such as power, temperature, time, steam volume, etc. according to the cooking tasks issued by the chef, the current equipment operating status, task priority and energy consumption optimization strategy.

[0029] Specifically, the system receives cooking tasks from chefs via the kitchen display, extracts key cooking parameters for each dish, and prioritizes tasks based on mealtime and urgency. Intelligent controllers monitor the current state of the equipment and, combined with historical data, assess its real-time performance degradation. Using a mixed integer programming (MIP) algorithm or reinforcement learning (RL) model, a mathematical model is developed with the goals of maximizing cooking efficiency and minimizing energy consumption to assign parameters to the intelligent cooking devices. The device's built-in temperature sensor provides real-time feedback on the cooking chamber temperature, and the system automatically fine-tunes power using a proportional-integral-derivative (PID) controller to ensure parameter accuracy. For example, during peak breakfast hours, when the stove needs to cook multiple breakfast dishes simultaneously, the system allocates power based on the cooking requirements, ensuring that different dishes, such as omelettes, porridge, and bacon, are cooked simultaneously and efficiently at their optimal heat and time. This significantly improves cooking efficiency and reduces energy consumption through refined energy management.

[0030] As an embodiment of the present application, the smart cooking device is also equipped with multiple types of sensors for real-time monitoring of the operating parameters of the device, including temperature sensors, vibration sensors, current sensors, voltage sensors, and pressure sensors. The operating parameters of the device are analyzed in real time through a machine learning model to predict potential failure risks of the device and trigger an early warning.

[0031] Specifically, multiple types of sensors are used to monitor the operating status of the equipment in real time, including parameters such as the temperature, vibration, current, and voltage of key components of the equipment. Machine learning models are then used to analyze these data in real time. First, the temperature and vibration of key components of smart devices, such as the operating frequency and current of motors, such as the operating current and voltage of oven heating elements, are collected in real time at a sampling frequency of 10Hz to form a time series data set. The fault type, such as "heating element aging" or "motor bearing wear," is annotated through historical maintenance records. Normal operating data is marked as "0," and fault data is marked as "1" or a specific fault code. The data is then cleaned to remove outliers, and samples are generated through a sliding window. Each sample contains multi-dimensional sensor data of 1,000 time steps. A convolutional neural network-long short-term memory network (CNN-LSTM) hybrid model is used: the CNN layer extracts spatial features of sensor data, such as the correlation between temperature and current. The LSTM layer captures long-term dependencies in time series data, such as the trend of increasing vibration values ​​after the equipment has been running continuously at high load for 3 hours. The model input layer has a dimension of time step × sensor type, and the CNN layer includes a 1-dimensional convolution kernel of size 5 to extract local features. The LSTM layer includes 128 neurons to capture time series features. The output layer uses a softmax activation function to output the probability of the fault type. The dataset is divided into normal operating data and historical fault data in a ratio of 8:2 for training, minimizing the cross-entropy loss function. The Adam optimizer is used with a learning rate of 0.001 to optimize the target. Finally, the prediction effect of each fault type is evaluated using a confusion matrix, with an accuracy rate of ≥95% and a recall rate of ≥90%. The model is deployed to edge computing nodes, such as the built-in chip of an intelligent controller, to perform online inference on real-time sensor data, generating a prediction result every 5 minutes. If the "fault probability > 0.8", the warning process is triggered.

[0032] Through machine learning models, the operating parameters of the equipment are analyzed in real time to accurately predict the potential failure risks of the equipment in advance. Once abnormal signs are found, detailed maintenance reminder information is immediately pushed to the kitchen supervisor's mobile terminal, covering the type of failure, possible causes, estimated impact and repair suggestions, etc., to gain valuable time for equipment maintenance and ensure the continuous and stable operation of kitchen equipment.

[0033] For example, a smart oven can automatically adjust the temperature, time, and baking mode based on the dish's requirements. When baking a cake, the oven will first quickly set the shape at a higher temperature before lowering the temperature for a slow bake, resulting in a fluffy interior and a golden brown exterior. For another example, if the stove is cooking a dish that requires a quick stir-fry, the intelligent control system will instantly adjust the stove's power to maximum based on the dish's requirements, ensuring the vegetables are cooked quickly and retain their freshness. At the same time, if it detects that a neighboring oven is preheating too long, indicating a possible heating element failure, it will immediately send an alert message containing the failure details and repair instructions to the control terminal.

[0034] In addition, smart cooking equipment also has an automatic cleaning function. After cooking is completed, the built-in timer or temperature sensor detects that the equipment is in an idle state and can automatically start the cleaning program to clean the internal grease and residue, reducing the workload of manual cleaning. Staff can also start cleaning with one click through the equipment operation interface or management backend. Specifically, steam cleaning is used for ovens and steamers. Ovens and steamers have built-in water tanks or are connected to the ship's water supply system. High-temperature steam is generated by heating, usually 120℃-150℃; the steam nozzle sprays evenly onto the inner wall of the cavity to soften grease and food residue; the built-in rotating spray arm cooperates with steam flushing to ensure that there is no blind spot coverage, such as the top and corners of the oven; sewage flows into the ship's wastewater treatment system through the bottom drainage hole. High-pressure water cleaning is suitable for stoves and fryers. These are connected to the ship's high-pressure water pipeline. A nozzle sprays high-pressure water jets at a pressure of 5-10 MPa, impacting the stovetop and fryer interior. A biodegradable cleaning agent, such as an environmentally friendly degreaser, is evenly sprayed via a circulating pump to break down grease. After cleaning, the system automatically dries the equipment, using either electric heating or a fan, to prevent residual water stains. The system also automatically adjusts cleaning parameters based on the type of equipment and the severity of the stain. For example, after baking greasy foods in the oven, the steam cleaning time is automatically extended to 20 minutes. After frying eggs on the stove, the system activates the quick-clean mode with a 5-minute high-pressure water rinse.

[0035] In addition, smart devices also have fault diagnosis functions, which can monitor the operating status of the equipment in real time. When a fault is detected in the equipment, an alarm will be issued in time and a fault solution will be provided, thereby improving the reliability and service life of the equipment.

[0036] Among them, the fault diagnosis function realizes real-time monitoring and analysis through the sensor network combined with the machine learning model. First, multi-dimensional data collection is carried out. The temperature of the heating element is monitored by the temperature sensor, such as the temperature of the oven heating tube and the temperature of the motor winding. The vibration sensor detects the vibration frequency of the equipment during operation, such as the vibration amplitude of the stove fan and the oven fan. The current sensor and voltage sensor collect the working current and voltage fluctuations in real time. For example, the normal working current of the stove is 10-15A, which jumps to 20A in the event of an abnormality. The pressure sensor monitors the steam pipe pressure. Then, anomaly detection and fault location are carried out, and the safe operating parameter range is preset. For example, if the oven heating element temperature is >300℃ and the normal operating temperature is ≤250℃, an "overheating warning" is triggered; if the stove fan vibration amplitude is >5mm / s and the normal operating temperature is ≤2mm / s, it indicates "bearing wear risk". Combined with machine learning model analysis, the isolation forest (Isolation Forest) is used. Forest algorithms or support vector machines (SVMs) compare real-time data with historical normal data to identify abnormal patterns. For example, when preheating an oven, the normal current rise curve is a smooth increase, from 0 to 10A within 20 seconds. If the current suddenly rises to 15A and remains there, the model identifies a "heating element failure" with over 95% accuracy. The system then responds to and handles the failure, setting emergency faults as level one alarms. For example, if a stove gas leak is detected through abnormal fluctuations in the gas flow meter, the gas supply is immediately cut off, an audible and visual alarm is triggered, and a notification is pushed to all kitchen terminals. Potential faults are set as level two alarms. For example, if the oven preheat time is 30% longer than the historical average, it may be due to aging of the heating element. A reminder is sent to the main terminal: "Oven preheat efficiency has decreased. Check the heating element within 3 days." Finally, the system automatically generates a maintenance guide based on the equipment maintenance knowledge base and provides troubleshooting solutions. For example, if a steamer pressure sensor failure is detected, the system will send the following: "Step 1: Check sensor wiring; Step 2: Calibrate sensor zero point; Step 3: Replace sensor (if the first two steps fail)."

[0037] The dish innovation unit automatically generates new dish combinations and cooking plans based on the inventory of ingredients, crew members' taste preferences, and nutritional needs. At the same time, the system can also recommend personalized dishes based on crew members' dining history and evaluation data to meet the taste needs of different crew members. For example, if there are fresh seafood and vegetables on board, the system will recommend some creative dishes with seafood and vegetables as the main ingredients, such as seafood and vegetable salad, steamed seafood with vegetables, etc. At the same time, the system will also recommend personalized dishes based on the crew members' dining history and evaluation data. If a crew member orders steak multiple times, the system will recommend steak dishes of different types and cooking methods, such as filet mignon, sirloin steak, etc., giving the crew more choices.

[0038] As an embodiment of the present application, the customized meal supply module includes a meal ordering unit, a diet management unit, and a special meal customization unit; The ordering and reservation unit provides a multimodal ordering terminal, allowing passengers to order and reserve food through personal handheld devices or the ship's self-service ordering terminals. The system provides personalized dish recommendations based on the passenger's order history and food preferences. The system also displays real-time dish inventory and estimated wait times, making it easier for passengers to make choices.

[0039] Among them, the ordering terminal adopts a client-server (C / S) architecture. The client develops native applications or cross-platform applications (such as using the React Native framework) based on the mobile operating system (such as Android / iOS). The server builds an API interface through Java / Node.js, supports real-time data synchronization with the main database of the ship's intelligent kitchen system, and the user interface is designed as a multilingual touch interactive interface. The dishes are displayed in categories, support keyword search, integrated slider controls and real-time inventory status display, and real-time push of order data is achieved through the WebSocket protocol. For example, after the user submits the order, the kitchen display immediately displays the tasks to be cooked. At the same time, the system calculates the estimated pick-up time based on the current number of people in the queue and the chef's workload, uses the MySQL database to store user order history, taste preferences and other data, and caches real-time inventory information through Redis to ensure response speed in high-concurrency scenarios. At the same time, the system captures data in real time through interactive behaviors at the ordering terminal, including dish selection records, taste adjustment parameters, reservation time and pick-up location; extracts user preference tags, basic tags and dynamic tags from the original data, uses the K-means algorithm to divide users into different groups, and generates differentiated recommendation strategies for different groups; and screens out available dishes based on the current inventory status of ingredients, matches dishes that meet nutritional standards from the recipe library based on user BMI and energy requirements, and sorts candidate dishes based on user historical scores and current taste adjustments, giving priority to recommending dishes with a similarity of >80%. If the reservation quantity of a dish exceeds the number of servings available in inventory, it will automatically be replaced with a similar dish and a replacement notification will be pushed.

[0040] The diet management unit provides personalized diet recommendations and nutrition plans based on the personnel's physical condition, exercise status and nutritional needs. The system forms an initial health record by entering the personnel's physical examination report into the system. The personnel wear smart watches or bracelets distributed by the integrated ship to synchronize basic data, exercise data, and health monitoring data in real time. Combined with the crew's schedule and work type, it automatically matches the exercise consumption standard to obtain the personnel's physical condition, exercise status and nutritional needs, filter out dishes that do not meet health standards, optimize the combination of dishes to meet nutritional needs, monitor and analyze the crew's actual intake in real time through ordering records and inventory data, and the system will send reminder information to them through personal handheld terminals, suggesting that they adjust their diet structure.

[0041] The special meal customization unit tailors meals to the specific needs of personnel in various special work categories. When a crew member or crew member becomes unwell and seeks medical attention, medical staff enter detailed medical information into the system. The system then quickly initiates the patient meal customization process. Based on the diagnosis, combined with medical nutrition therapy guidelines and individual physical conditions, a personalized patient meal recipe is precisely formulated. The system integrates with the ship's work scheduling system in real time to accurately track overtime personnel information. For overtime personnel, energy-rich and nutritious meals are provided based on overtime hours, work intensity, and nighttime physiological metabolic characteristics. Meal size is dynamically adjusted based on the overtime worker's real-time hunger feedback and an energy expenditure estimation model, ensuring timely and adequate energy replenishment to maintain optimal working conditions. For personnel engaged in specialized operations, such as underwater operations and those in high-temperature cabins, which involve high labor intensity, high risk, and rapid energy consumption, the system customizes specialized high-energy, high-protein recipes based on the type of work, duration, environmental conditions, and research on labor physiology. Provide foods such as steaks rich in high-quality protein and iron, chicken legs that are easy to eat and high in energy density, and nuts rich in unsaturated fatty acids and trace elements. The meal amount is scientifically and rationally planned according to the intensity and time of work. Through flexible methods such as separate meals and additional meals, it is ensured that special operations personnel have sufficient physical strength and energy to complete the tasks before and after high-intensity work, and maintain the physical health and work safety of the operators.

[0042] Among them, the energy consumption prediction model first collects the basic characteristics, work characteristics, physiological characteristics and current environmental characteristics of the personnel, cleans the collected features of outliers, standardizes the features and extracts time series features, uses the LSTM-Transformer hybrid deep learning model to process time series data and capture the nonlinear relationship between multiple features, and the output layer performs regression to predict the energy consumption value. It uses the Adam optimizer and Huber Loss loss function for training, collects data in real time based on the personnel's wearable devices, obtains the cabin temperature and humidity through environmental sensors, corrects the energy consumption calculation, and assesses the degree of fatigue through voice interaction combined with camera recognition of yawning frequency. The model prediction value is subtracted from the intake recorded by the system to calculate the energy gap, and the meal amount is dynamically adjusted according to the energy gap. The energy consumption prediction model also collects user feedback, regularly iterates the model, uses new data to fine-tune the model parameters every week, retrains the model every quarter, incorporates seasonal factors, and continuously optimizes the model.

[0043] As an embodiment of the present application, the customized meal supply module also includes a robot service unit, which includes multiple service robots. Each of the service robots has intelligent navigation, voice interaction, and object recognition functions, and is responsible for the delivery of dishes and cleaning of tables. The service robots communicate with the system through the MQTT protocol, and synchronize in real time the delivery task status, robot power, storage box capacity and other status data, and new task instructions triggered by user needs. When multiple service robots are working at the same time, the system assigns tasks through a scheduling algorithm, such as a greedy algorithm, and gives priority to assigning robots closest to the task starting point to balance the load of each robot and avoid a single robot working continuously for a long time.

[0044] Specifically, the service robot scans the restaurant environment through LiDAR or visual sensors, constructs a two-dimensional grid map or a three-dimensional point cloud map, and locates its own position in real time. The robot calculates the shortest path using a dedicated algorithm or Dijkstra algorithm. It detects dynamic obstacles using real-time infrared obstacle avoidance sensor data, triggers the Dynamic Windowing Algorithm (DWA) to replan the local path, receives delivery orders, analyzes the target location coordinates, and moves along the planned path, updating its location every 0.5 seconds. It pauses and issues a voice prompt when encountering an obstacle, and then alerts the user via voice or light notifications upon reaching the table. The service robot's built-in microphone array collects voice signals, employs beamforming technology to enhance the target speech and suppress background noise, and converts speech to text using a pre-trained Google Speech-to-Text speech recognition model with an accuracy rate of ≥95%. It identifies user needs through keyword matching or the semantic understanding model Rasa, and converts these into natural speech using a speech synthesis engine such as Alibaba Cloud TTS. The service robot is also equipped with an RGB camera or an RGBD camera to take multi-angle photos of the dining table, and use the YOLOv5 or Faster R-CNN target detection model to identify tableware and garbage with an accuracy rate of ≥90%; when the service robot detects the edge outline of the plate and the interior is empty, the confidence level is >0.8 and it is judged as "empty tableware"; if it identifies white irregular objects such as paper towels, it is judged as "garbage". The service robot moves to the target dining table, confirms the items that need to be cleaned through visual recognition, is equipped with a robotic arm, uses a gripper or suction cup to grab the tableware, puts it into the built-in storage box, collects the debris through the bottom roller brush and vacuum device, and deposits it into the trash can.

[0045] As an embodiment of the present application, the real-time hygiene monitoring module includes a personnel operation behavior monitoring unit and an abnormal food material management unit; The personnel operation behavior monitoring unit uses an image recognition camera combined with image recognition technology and deep learning algorithms to monitor 24 hours a day, capturing the operation behavior of kitchen staff in real time, accurately identifying operations that do not comply with hygiene regulations, such as failure to wear masks, hats, and gloves as required, cross-contamination during food processing, and incomplete kitchen cleaning, and issues timely warnings; Specifically, the personnel operation behavior monitoring unit adopts the YOLOv5s target detection model and trains multiple target categories for kitchen scenes. The multiple target categories include detecting whether the mask covers the mouth and nose, with a confidence threshold of ≥0.9; detecting whether the hands are wearing white or blue food-grade gloves, and the image recognition camera captures the picture in real time. The algorithm scans one frame every 0.5 seconds; if it is detected that a staff member is not wearing a mask, the confidence level is 0.95, and an audible and visual alarm is immediately triggered, and a violation record is marked in the management background; using I3D (Inflated 3D ConvNet) model analyzes motion patterns in continuous video frames and identifies various cross-contamination risk behaviors, including mixing raw and cooked ingredients, touching cooked food with bare hands, and mixing containers. By establishing a "cutting board usage log" database, RFID tags are associated with the use of cutting boards. Detection of the same cutting board handling raw meat and ready-to-eat food with an interval of less than 5 minutes and without cleaning the cutting board indicates a risk of mixing raw and cooked ingredients. The image recognition camera identifies ungloved hands directly grasping cooked food as a risk of bare hands touching cooked food. The model combines container color coding (red for raw meat containers and blue for cooked food containers) to identify cross-regional usage. It detects unwashed containers used to hold raw meat and directly used to hold cooked food as a risk of mixing containers. When a violation occurs, the system issues a real-time alert, and the manager's mobile app simultaneously pushes a notification containing a screenshot of the violation. The violation record is automatically associated with the video clip and stored on the cloud server. It supports retrieval by "date + type of work + violation type," and counts the number of violations in each area weekly to assist in management decision-making.

[0046] The personnel operation behavior monitoring unit also uses the U-Net semantic segmentation model to perform pixel-level analysis on the kitchen countertop, floor, and equipment surface, and quantify the cleanliness score. The cleanliness score calculation formula is as follows:

[0047] in, Rate the cleanliness. is the total area to be cleaned, For the stain area, if the cleanliness score If the score is below 85, the system will automatically generate a cleaning work order and assign the nearest cleaner to clean.

[0048] The food abnormality management unit monitors the microbial content and odor changes in real time through a microbial sensor and an odor sensor respectively.

[0049] Among them, the microbial sensor has a built-in miniature sampling probe, which regularly wipes and samples the air, food surface or tableware surface every 30 minutes, or collects water samples directly from the kitchen drainage system. By detecting the changes in the electrical signals of microbial metabolites, the microbial concentration is quantitatively analyzed, or the fluorescent staining technology SYBR Green dye is used to label DNA combined with microscopic imaging. The microbial fluorescence signal is captured by the camera, and the number of microorganisms is calculated by the image analysis algorithm. The microbial sensor has a built-in standard curve database, which compares the detection data with the threshold in real time. If the threshold is exceeded, an alarm signal is immediately sent to the system via the RS-485 bus or Wi-Fi, and the detection time, location and specific data are recorded at the same time.

[0050] The odor sensor uses a metal oxide semiconductor (MOS) sensor or a gas chromatography-mass spectrometry (GC-MS) micromodule to detect volatile organic compounds (VOCs) in the kitchen, such as hydrogen sulfide and methyl mercaptan produced by food spoilage, aldehydes and ketones produced by the decomposition of oil stains that are not cleaned in time, and burnt smells caused by electrical overloads, such as benzene. Once an abnormal odor is detected, the system immediately sends an alarm to the kitchen manager's mobile terminal and the restaurant display screen through various methods such as sound and light alarms and push notifications. At the same time, it records in detail the time, location, specific detection data and other information of the problem, so that managers can quickly locate the root cause of the problem and take targeted corrective measures in a timely manner, effectively ensuring kitchen hygiene and safety and protecting the health of diners.

[0051] As an embodiment of the present application, the personnel dining scheduling module includes a shift information collection unit, a dining time planning unit and a dining guidance unit; The shift information collection unit establishes a deep data connection with the ship's personnel management system, and collects crew shift information in real time through a data interface, including detailed data such as the precise working hours, rest time, handover time of different shifts, and special working hours arrangements for each position.

[0052] The dining time planning unit uses a mixed integer programming algorithm based on the collected crew shift information to minimize waiting time and balance restaurant load, and dynamically adjusts time period allocation in combination with real-time traffic monitoring. Specifically, the dining time planning unit first parses the crew shift information to obtain the shift ID, working period, crew list, and special working hours. It then expands the time window, reserving one hour before and after the working period as a potential dining window. It then cleans historical dining data, calculates restaurant traffic flow for each period over the past 30 days, identifies peak hours, calculates average dining time per person, and evaluates restaurant capacity. A mixed integer programming (MIP) algorithm is used to minimize waiting time and balance restaurant load, and an objective function is constructed:

[0053] in, Indicates the shift Assign to time period Number of people dining, Indicates the total waiting time for a shift In the period The average waiting time, Indicates the flow of people and restaurant capacity at each time period deviation to ensure load balancing.

[0054] Among them, the constraints include avoiding conflicts during working hours, shift Meal time Must be completely included in non-working hours, and the number of people must be constant , For flights Total number of people; time period capacity limit ; Special needs are given priority, such as patients and overtime workers can choose any time period.

[0055] The mixed integer programming algorithm is used to minimize waiting time and balance restaurant load to perform initial allocation strategy. It automatically triggers dynamic adjustment mechanism based on real-time feedback and special circumstances. It uses the camera at the restaurant entrance to count the real-time flow of people. If the number of people in a certain period exceeds the restaurant capacity, It automatically diverts some employees to adjacent time periods and pushes notifications to affected employees. For special circumstances of temporary overtime, it automatically delays the meal time of overtime employees and notifies the kitchen to extend meal supply.

[0056] The dining guidance unit is based on real-time restaurant data and user characteristics, and uses dynamic priority queue division and shortest path matching algorithm, combined with the M / M / 1 queuing model to predict waiting time, to achieve multi-objective optimization, and through a full-range voice broadcast system and dynamic display screen, to guide people in different shifts to dine in an orderly manner in real time.

[0057] Specifically, the core goals are to minimize the average waiting time, balance the load of each pick-up window or table, reduce the gathering of people, and maintain social distance. Real-time data and user characteristics are used as input parameters. Real-time data includes the current number of people in the restaurant, the number of available seats, the queue length of each pick-up window, and the remaining dining time obtained through camera visual counting or seat sensors; user characteristics include personnel shift information and special needs; dynamic priority queue division and shortest path matching algorithm perform dynamic queue management in stages, classifying patients, pregnant women, and special operations personnel as first priority, personnel who arrive early and have limited remaining dining time as second priority, and normal shift personnel as third priority; dynamic queue management includes arranging ordinary users in order of arrival time, assigning each person a unique queue number in the main queue, and inserting high-priority users in real time to ensure that their waiting time is ≤2 minutes in the emergency queue. For example: if mid-shift employee A arrives with 15 minutes remaining dining time, the system will determine him as second priority, insert him to the front of the main queue, and arrange seats first. The available seat coordinates are updated in real time through the seat pool and marked as allocable resources. The seat matching strategy includes the nearest neighbor priority strategy to assign queuing users to the available seats closest to the pick-up window; the regional balancing strategy to evenly distribute the number of people according to the restaurant zone to avoid overcrowding in certain areas; and the special needs adaptation strategy to assign patients seats close to both the pick-up window and the restroom. The M / M / 1 queue model is used to calculate the waiting time of ordinary users: ,in, is the arrival rate, For service rate.

[0058] At the same time, the large display screen at the restaurant entrance scrolls in real time to display information such as the dining schedule for each shift, the current real-time seating situation in the restaurant, and the estimated waiting time, allowing personnel to view information at a glance and plan their dining itinerary independently. Specifically, this application achieves real-time interaction and dynamic adjustment through mobile notifications and on-site guidance. When users arrive at the restaurant entrance, they scan the code to get a number. The APP displays the current queue number, estimated waiting time, and recommended pickup window in real time. The queue status is updated every 5 minutes through voice broadcast. The display screen guides dynamically marking the assigned seats and displays the shortest path to the seats.

[0059] As an embodiment of the present application, the restaurant environment adjustment module includes an atmosphere light control unit, a temperature and humidity adjustment unit, and a background media control unit; The atmosphere light control unit automatically adjusts the restaurant atmosphere light according to the dining content, time and real-time activity atmosphere in the restaurant; specifically, the atmosphere light is connected to the system via the Zigbee or KNX protocol, and the system triggers the preset scene according to the time: the color temperature of the breakfast atmosphere light is controlled between 3000K-4000K, and the brightness is 80%, simulating the natural light in the early morning; the color temperature of the lunch atmosphere light is controlled between 2700K-3000K to help soothe the dining mood; the color temperature of the dinner atmosphere light is controlled between 2500K-2700K, and the brightness is 50%, combined with the local spotlight on the dining table to create a warm feeling.

[0060] The temperature and humidity control unit uses a PID control algorithm to dynamically adjust the air conditioning temperature and humidity, taking into account the number of diners and seasonal factors. Specifically, the system sets the comfortable temperature threshold to 24°C–26°C in summer and 20°C–22°C in winter, and the comfortable humidity threshold to 40%-60%. The PID control algorithm is used to dynamically adjust the air conditioning power based on the number of diners: when the number of diners is greater than 100, the air conditioning cooling power is increased to 120%, with a target temperature of 25°C and a humidity of 55%. During off-peak hours, the power is reduced to 60%, with a target temperature of 26°C and a humidity of 60%, entering energy-saving mode. The temperature and humidity control unit strengthens cooling in summer and heating in winter according to seasonal changes, ensuring a consistently comfortable and pleasant environment within the restaurant.

[0061] The background media control unit synchronizes background music and television programs based on meal times, crew physiological data, and questionnaire results. It also synchronizes lyrics and subtitles via the display screen and regularly updates the media library. Specifically, the system pushes electronic questionnaires to crew members' mobile devices via the ship's internal network. The system compiles real-time statistics and generates visual reports. Program playback strategies are updated based on meal times and crew physiological data collected by integrated wearable devices. The music library and television programs are regularly updated to enhance the dining experience.

[0062] As an embodiment of the present application, the energy management module includes an energy consumption monitoring unit, an energy saving control unit and an energy recovery and utilization unit; The energy consumption monitoring unit uses energy monitoring equipment to monitor the kitchen system's energy consumption data in real time, transmitting the data to the system for analysis and statistics, locating peak energy consumption periods and major energy-consuming devices. Specifically, smart meters are installed at the kitchen's power distribution box, regional distribution boxes, and the incoming lines of large equipment to monitor power consumption by area and device. Gas flow meters are installed at the main gas pipeline valve and the air inlets of gas appliances such as stoves and steamers to monitor real-time gas flow and cumulative usage. Water meters are installed at water points such as the kitchen water main, dishwashers, and sinks to record water consumption at each stage. The monitoring equipment connects to the smart kitchen system's main server via Modbus RTU / ASCII or MQTT protocols, collecting real-time data once per second and generating statistical values ​​every minute to ensure real-time and accurate data. Energy consumption data statistics include calculating the energy consumption contribution of each dining period; ranking equipment energy consumption by day, week, or month; and comparing data from adjacent periods to identify abnormal fluctuations. The analysis method includes generating energy consumption trend charts and pie charts through Tableau or the system's built-in chart components; using the moving average method (MA) or exponential smoothing method (ES) to establish a normal energy consumption baseline, and determining that real-time data exceeds the baseline by ±10% as abnormal.

[0063] The energy-saving control unit formulates a corresponding energy-saving control strategy based on the monitoring results of the energy consumption monitoring unit. During non-meal times, the system automatically reduces the power of the cooking equipment and sets the temperature of the stove and oven to a lower insulation state to reduce energy waste. At the same time, the operating status of air-conditioning, lighting and other equipment is reasonably adjusted according to the number and time of the crew's meals. If the number of people dining is small, the system will automatically turn off the air-conditioning and lighting equipment in some areas to reduce energy consumption. In addition, energy-saving cooking equipment and appliances, such as high-efficiency energy-saving stoves, ovens and refrigerators, are used. These devices have higher energy utilization efficiency and can reduce energy consumption and improve energy utilization efficiency while meeting cooking needs.

[0064] The energy recovery and utilization unit collects the waste heat generated by the cooking equipment through a heat exchanger and heats water, while a purification system is used to treat and purify the kitchen wastewater. Specifically, during the cooking process, the stove and oven will generate a large amount of waste heat, which the system collects through a heat exchanger and uses to heat water. The heated water can be used for washing dishes and washing people, reducing the energy demand for the hot water supply system. At the same time, a wastewater treatment and purification system is used to treat and purify the kitchen wastewater. The treated wastewater can be used to irrigate green plants on the ship or flush toilets, etc., realizing the recycling of water resources. Through energy recovery and utilization, the energy consumption and operating costs of the ship are reduced, and the impact on the environment is also reduced.

[0065] The above description is merely an illustration of some preferred embodiments of the present disclosure and the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also includes other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned inventive concept.

Claims

1. An artificial intelligence-based intelligent kitchen system for ships, characterized in that: include: The intelligent food management module is used to monitor food inventory information in real time and dynamically adjust food procurement, storage and consumption strategies based on multi-dimensional data using AI algorithms; Intelligent cooking control module: optimizes cooking parameters through supervised learning models and automatically controls intelligent cooking equipment based on cooking parameters; Customized meal supply module: used to provide a meal ordering terminal that supports multimodal interaction and customize recipes based on employee health data and job requirements; Real-time hygiene monitoring module: A triple monitoring network is built using image recognition cameras, microbial sensors, and odor sensors to capture real-time changes in operator behavior, microbial content, and odor, enabling early warning of hygiene risks. Personnel dining scheduling module: It is configured to connect to the ship's personnel management system, plan dining time periods based on the collected crew shift data, and guide personnel to eat in real time; Restaurant environment adjustment module: automatically adjusts lighting color temperature, temperature and humidity, and background media based on ship sailing time, number of diners, and real-time activity needs; Energy management module: used to monitor energy consumption data in real time, formulate energy-saving control strategies, dynamically perform power regulation and drive the recycling of waste heat and wastewater.

2. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The food intelligent management module includes an inventory management unit, a food planning unit, a quality inspection unit and a food traceability unit; The inventory management unit uses weight sensors, temperature sensors, humidity sensors, and RFID readers to monitor the inventory quantity, shelf life, and storage conditions of food in real time. Based on the ship's sailing plan, crew numbers, and historical consumption data, it predicts demand trends for food and automatically generates a procurement plan. The food planning unit adjusts the storage ratio of various food items based on the ship's voyage planning cycle and the real-time updated remaining sailing time, according to the shelf life of the food items, inventory dynamics, and nutritional balance requirements. At the same time, the food consumption schedule is planned based on the remaining sailing time, and a food consumption curve is generated based on the shelf life and storage conditions of different food items. The quality detection unit uses spectral analysis and image recognition technology to detect the quality of food. If the quality of the food is found to be unqualified, an alarm is automatically triggered and the food is marked and isolated. The food traceability unit uses QR codes and RFID readers to record the source, purchase time, transportation process, and storage conditions of the food, thereby establishing a complete food traceability system.

3. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The intelligent cooking control module includes a cooking parameter optimization unit, an intelligent device control unit, and a dish innovation unit; The cooking parameter optimization unit collects multi-dimensional data to form a structured data set, uses machine learning technology to extract key features of the structured data set, uses a supervised learning model for training, and continuously updates the supervised learning model through an online learning algorithm to dynamically optimize cooking parameters; The intelligent cooking device control unit realizes two-way real-time data interaction with the system by installing an Internet of Things controller on the intelligent cooking device, and automatically completes the cooking task according to the preset cooking parameters; The dish innovation unit automatically generates new dish combinations and cooking plans based on the inventory of ingredients, the taste preferences and nutritional needs of the personnel.

4. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The smart cooking device is also equipped with multiple types of sensors for real-time monitoring of the device's operating parameters. The device's operating parameters are analyzed in real time through a machine learning model to predict potential equipment failure risks and trigger early warnings.

5. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The customized meal supply module includes a meal ordering unit, a diet management unit, and a special meal customization unit; The order reservation unit provides an order terminal that supports multimodal interaction; The diet management unit provides personalized diet recommendations and nutrition plans based on the individual's physical condition, exercise status, and nutritional needs; The special meal customization unit customizes special meals according to the special needs of various special workers.

6. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The customized meal supply module also includes a robot service unit, which uses a service robot with intelligent navigation, voice interaction, and object recognition functions, and is responsible for the delivery of dishes and cleaning of the table.

7. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The real-time hygiene monitoring module includes a personnel operation behavior monitoring unit and an ingredient abnormality management unit; The personnel operation behavior monitoring unit uses an image recognition camera combined with image recognition technology and deep learning algorithms to capture the operation behavior of kitchen staff in real time; The food abnormality management unit monitors the microbial content and odor changes in real time through a microbial sensor and an odor sensor respectively.

8. The artificial intelligence-based ship intelligent kitchen system according to claim 1 is characterized in that: The personnel dining scheduling module includes a shift information collection unit, a dining time planning unit and a dining guidance unit; The shift information collection unit collects crew shift information in real time through a data interface; The dining time planning unit uses a mixed integer programming algorithm based on the collected crew shift information to minimize waiting time and balance restaurant load, and dynamically adjusts time allocation in combination with real-time passenger flow monitoring; The dining guidance unit is based on real-time restaurant data and user characteristics, and uses dynamic priority queue division and shortest path seat allocation algorithms, combined with the M / M / 1 queuing model to predict waiting time, to achieve multi-objective optimization. It also uses a full-range voice broadcast system and dynamic display screens to guide people on different shifts to dine in an orderly manner in real time.

9. The artificial intelligence-based ship intelligent kitchen system according to claim 1, characterized in that: The restaurant environment adjustment module includes an atmosphere light control unit, a temperature and humidity adjustment unit, and a background media control unit; The atmosphere light control unit automatically adjusts the restaurant atmosphere light according to the dining content, time and real-time activity atmosphere in the restaurant; The temperature and humidity adjustment unit uses a PID control algorithm to dynamically adjust the air conditioning temperature and humidity based on the number of diners and seasonal factors; The background media control unit synchronously controls background music and TV programs based on meal periods, crew physiological data and questionnaire results, synchronizes lyrics and subtitles through the display screen and regularly updates the media library.

10. The artificial intelligence-based ship intelligent kitchen system according to claim 1, characterized in that: The energy management module includes an energy consumption monitoring unit, an energy saving control unit and an energy recovery and utilization unit; The energy consumption monitoring unit monitors the energy consumption data of the kitchen system in real time through the energy monitoring equipment, and transmits the data to the system for analysis and statistics, and locates the peak hours of energy consumption and the main energy-consuming equipment; The energy-saving control unit formulates a corresponding energy-saving control strategy according to the monitoring results of the energy consumption monitoring unit; The energy recovery and utilization unit collects waste heat generated by the cooking equipment through a heat exchanger and heats water, while a purification system is used to treat and purify the kitchen wastewater.

Images