Cooking robot control method and system based on knowledge graph

Abstract

The invention discloses a cooking robot control method and system based on a knowledge graph, and relates to the technical field of cooking robot control, and the method comprises the steps: collecting multi-source cooking data, constructing the knowledge graph, receiving user demands, recommending dishes based on the graph, determining cooking parameters according to the recommended dishes, and controlling the cooking robot according to the cooking parameters. Comprise time, temperature, seasoning dosage and the like. The parameters are sent to a cooking robot controller, data are collected by a sensor during cooking, the data are compared with standard data, and the parameters are adjusted by a feedback algorithm to ensure that cooking meets the expectation. According to the method, personalized dish recommendation and cooking parameter adaptive adjustment can be realized, dish quality and taste are ensured, user health is considered, nutrition balance is promoted, the intelligent degree of the cooking robot is improved, the market competitiveness of the cooking robot is enhanced, and the requirement of people for high-quality cooking experience is met.

Description

Technical Field

[0001] The present invention relates to the technical field of cooking robot control, and in particular to a cooking robot control method and system based on a knowledge graph. Background Art

[0002] With the acceleration of modern life and the continuous advancement of science and technology, people are increasingly pursuing a convenient and efficient lifestyle. In terms of diet, cooking robots have emerged, which can help people reduce the time spent in the kitchen and improve cooking efficiency. However, the current cooking robots still have many shortcomings in the control method.

[0003] Traditional cooking robots are mostly operated by preset programs. These preset programs are often based on fixed cooking time, temperature and operation steps, and lack comprehensive consideration of factors such as food characteristics, user taste preferences and nutritional needs. For example, the same food from different origins and different seasons may have significant differences in texture, water content and other characteristics, and the preset programs cannot make flexible adjustments to these differences, resulting in uneven taste and quality of the cooked dishes.

[0004] At the same time, existing cooking robots are also relatively weak in terms of dish recommendations. They usually can only provide a limited number of fixed dishes, and cannot recommend suitable dishes based on the user's personalized needs, such as specific taste preferences (such as spicy or sweet flavors), dietary taboos (such as allergies to certain ingredients), nutritional needs (such as the need to control calorie intake or supplement specific nutrients), etc.

[0005] In addition, during the cooking process, due to the lack of real-time perception and feedback adjustment capabilities of the actual cooking environment and changes in ingredients, it is difficult for the cooking robot to ensure that every dish can achieve the ideal cooking effect. For example, when the amount of ingredients in the pot changes or the power of the heating equipment fluctuates, the robot cannot adjust the cooking parameters in time, which may cause the dish to be overcooked or undercooked.

[0006] With people's increasing demands for food quality and personalization, as well as the development of artificial intelligence and knowledge graph technology, the market urgently needs a more intelligent cooking robot control method and system that can comprehensively consider multiple factors. This system can use knowledge graphs to integrate cooking knowledge, achieve accurate dish recommendations and adaptive cooking control, thereby improving the intelligence level of cooking robots and meeting the diverse needs of users. Summary of the invention

[0007] The present invention proposes a cooking robot control method and system based on knowledge graph to solve the problems mentioned in the above-mentioned prior art.

[0008] In order to achieve the above object, the present invention adopts the following technical solution: a cooking robot control method based on knowledge graph, comprising the following steps: Knowledge graph construction steps: Collect multi-source cooking data, which includes food data, seasoning data, cooking equipment data, cooking process data, user health data and dish nutrition data; the food data covers food type, origin, seasonal information, and food characteristics, which are expressed in vector form as , represents the 1st to mth food characteristics; the seasoning data includes the type of seasoning, taste, saltiness, and spiciness information, recorded as , Indicates the 1st to nth seasoning attributes; cooking equipment data includes equipment type, power, heating method, and capacity information, expressed as , e1-e k represents the 1-kth equipment parameters; cooking process data involves cooking methods, cooking time range, heat requirements, operation steps, and It means that p1-p q represents the 1-qth process attributes; user health data includes the user's allergic ingredients, dietary taboos, and chronic disease information, recorded as , h1-h u represents the 1-uth health attribute; the nutritional data of the dish includes the calorie, protein content, fat content, and vitamin content of the dish, recorded as , n1-n w Represents the 1-w nutritional attributes, uses semantic analysis and entity recognition technology to process the collected data and construct a cooking knowledge graph , where V is a node set, including the ingredient node V F , Seasoning Node V S , cooking equipment node V E , Cooking Process Node V P User Health Node V H And the food nutrition node V N , E is the edge set, which represents the semantic relationship between nodes, including the matching relationship between ingredients and seasonings, the adaptation relationship between ingredients and cooking techniques, the relationship between user health status and ingredient taboos, and the weight of the relationship. Determined through statistical analysis of cooking data and expert experience evaluation, it indicates the closeness of the relationship. The weight calculation method is: ,in represents the number of times entity i and entity j appear together in the data set, and J is the number of all entities related to entity i; Dish recommendation step: receiving the demand information input by the user, the demand information includes the number of diners, dish taste preferences, ingredient preferences, cooking time requirements, nutritional requirements, and budget constraints; searching and matching based on the user demand information in the cooking knowledge graph G, using a recommendation algorithm based on multi-attribute fusion, and calculating the matching degree M between the user demand and the dish information in the graph. The matching degree calculation formula is: , where α, β, γ, δ, and ε are weight coefficients, which respectively represent the importance of ingredients, seasonings, cooking techniques, user health, and dish nutrition in the matching process, and , M F is the food matching degree, M S is the seasoning matching degree, M P is the cooking process matching degree, M H For user health matching, M N is the nutritional matching degree of the dishes, and the calculation methods are , , in and They represent the normalized values ​​of the user demand and the i-th ingredient characteristic in the dish, respectively, and m represents the number of ingredient characteristics. represents the weight of food characteristics, and They represent the user demand and the normalized value of the j-th seasoning attribute in the dish, respectively. represents the weight of the seasoning attribute, n represents the number of seasoning attributes, and Represent the normalized values ​​of user demand and the rth process attribute in the dish, respectively. and represent the normalized value of the user demand and the vth health attribute in the dish, respectively. and They represent the normalized value of the user demand and the xth nutritional attribute in the dish respectively; Cooking process determination step: According to the recommended dish information, query the corresponding cooking process information in the cooking knowledge graph G, including the detailed cooking step sequence, cooking time required for each step, cooking temperature, heat control requirements, seasoning order and amount, and the relationship between cooking time and ingredient type, quantity and cooking equipment power is: ,in is the time coefficient, is the number of food types in step i, is the total amount of ingredients in step i, is the power of the cooking equipment used in step i, As the basic time constant; determine the temperature control parameters through the fuzzy logic reasoning system. The system input is the characteristics of the ingredients and the cooking process, and the output is the temperature control parameters, including the fire size and the heating time interval. After fuzzifying the input variables, a fuzzy rule base is formulated based on expert experience, including: IF the moisture content of the ingredients is high AND the cooking process is stir-fry THEN the fire size is large AND the heating time interval is short. The fuzzy reasoning engine is used for reasoning, and finally the specific temperature control parameter values ​​are obtained through defuzzification. Cooking control steps: Send the determined cooking process information to the controller of the cooking robot. The controller executes each step in sequence according to the cooking step sequence. During the execution process, the on and off time of the cooking equipment is controlled according to the cooking time information. The heating power is controlled according to the cooking temperature information to maintain the target temperature. The flame size or the power output of the heating element is adjusted according to the fire control requirements. The seasoning feeding device is controlled according to the seasoning feeding sequence and feeding amount information. The cooking environment data and cooking equipment operation data are collected in real time through sensors and compared with the preset standard cooking data. The adaptive feedback control algorithm is used to adjust the cooking parameters. The feedback control algorithm formula is: ,in Adjust the amount for cooking parameters, and is the feedback coefficient, For cooking environment data, is the standard cooking environment data, Running data for cooking equipment, Run data for standard cooking equipment to ensure that the cooking process is as expected and the dishes meet user needs.

[0009] Furthermore, in the knowledge graph construction step, the sources of food data collection include food supplier database, agricultural product market data platform, agricultural Internet of Things sensor data and user manual input. The number of collected food types is no less than 500, the origin information is accurate to the county level, and the seasonal information is accurate to the month.

[0010] Furthermore, in the knowledge graph construction step, the sources of seasoning data collection include seasoning manufacturer database, food seasoning research institution data, e-commerce platform sales data and user taste feedback data. The spiciness information is measured using Scoville spiciness unit SHU, and the saltiness is expressed as a percentage of salt content.

[0011] Furthermore, in the dish recommendation step, user demand information is collected through voice recognition, text input, and image recognition. The taste preference characteristic dimensions include at least the sourness, sweetness, bitterness, spiciness, saltiness, freshness, greasiness, and crispness. The nutritional needs in the demand information are connected to the nutrition expert system to recommend a suitable range of nutritional intake based on the user's age, gender, and physical condition.

[0012] Furthermore, in the cooking process determination step, in the cooking time calculation, the time coefficient ranges from 0.5 to 5, and the basic time constant ranges from 0 to 5 minutes. In the cooking temperature calculation, the temperature coefficient ranges from 10 to 200, and the basic temperature constant ranges from 30°C to 250°C. In the seasoning amount calculation, the amount coefficient ranges from 0.1 to 10, and the basic amount constant ranges from 0 to 20 grams.

[0013] Furthermore, in the cooking control step, the sensors include temperature sensors, humidity sensors, image sensors, oil fume concentration sensors, and weight sensors. The cooking equipment operation data is acquired through the built-in sensors and controllers of the equipment. The feedback coefficient ranges from 0.1 to 1, and the range of is 0.05 to 0.5, which is dynamically adjusted according to the different stages of the cooking process.

[0014] Furthermore, the knowledge graph construction step also includes a knowledge update mechanism, which regularly obtains new cooking knowledge from Internet cooking forums, food blogs, and professional cooking research literature data sources, such as emerging ingredient combinations and innovative cooking techniques, and integrates new knowledge into the existing cooking knowledge graph through entity alignment and relationship update algorithms. The entity alignment algorithm uses a hybrid algorithm based on semantic similarity and structural similarity to calculate the similarity between new entities and existing entities. If the similarity exceeds the set threshold, a merge or association operation is performed. The relationship update algorithm adjusts the weight and type of the relationship between nodes according to the new knowledge data.

[0015] Furthermore, in the cooking process determination step, for dishes with multiple cooking process options, process comparisons and recommendation reasons are provided. The process comparisons include taste comparisons, nutrient retention comparisons, cooking difficulty comparisons, and cooking time comparisons. The recommendation reasons are based on user demand information and statistical analysis of data in the cooking knowledge graph.

[0016] Furthermore, in the cooking control step, when an abnormal situation occurs during the cooking process, the cooking robot can automatically handle it according to the preset exception handling rules, including reducing the heating power, stopping cooking, and issuing an alarm to notify the user. At the same time, the abnormal situation information is recorded in the log file for subsequent analysis and optimization of the cooking process. The exception handling rules are derived through the analysis and summary of a large amount of cooking experimental data and actual usage data. The decision tree algorithm is used to construct the exception handling model. The input of the model is the abnormal situation characteristics, and the output is the processing measures.

[0017] Furthermore, a system for implementing a cooking robot control method based on a knowledge graph includes: Knowledge graph construction module: used to implement the knowledge graph construction steps and construct the cooking knowledge graph G; Dish recommendation module: used to implement the dish recommendation steps and recommend dish information according to user needs; Cooking process determination module: used to determine the steps of the cooking process and determine the cooking process information; Cooking control module: used to implement cooking control steps and control the cooking process of the cooking robot; Data acquisition module: used to collect cooking data, user demand information, cooking environment data and cooking equipment operation data, providing data support for each module. The collected data is cleaned, denoised and format converted by the data preprocessing unit; Storage module: used to store cooking knowledge graphs, dish information, cooking process information, and data collected by the data acquisition module. The storage capacity is not less than 500GB. It adopts a distributed storage system and has data backup and recovery functions. Human-computer interaction module: provides an interactive interface between the user and the cooking robot, including a display screen, voice interaction equipment, and a touch control panel. Users can input demand information, view recommended dishes, monitor the cooking process, and obtain cooking tips and suggestions; Communication module: used to realize the communication connection between the cooking robot and external devices, supporting Bluetooth, Wi-Fi, 4G / 5G communication protocols, so as to receive remote control, including updating knowledge graphs and cooking programs.

[0018] Compared with the prior art, the present invention has the following beneficial effects: First of all, in terms of dish recommendations, this patent can achieve highly personalized dish recommendations. By constructing a knowledge graph containing information on ingredients, seasonings, cooking techniques, user health, and dish nutrition, the system can fully understand the needs of users, including taste preferences, dietary taboos, nutritional needs, etc., and make accurate matches in a large amount of dish data based on this, and recommend dishes that best meet their needs to users. Whether it is an elderly person who likes light flavors, a young person who pursues heavy flavors, or a person with special dietary needs, they can all get satisfactory dish recommendations.

[0019] Secondly, in terms of cooking control, this patent has the advantage of self-adaptation. Using the rich information and precise parameter relationships in the knowledge graph, the system can calculate and adjust cooking parameters such as cooking time, temperature, heat, and seasoning amount in real time according to the actual characteristics of the ingredients (such as texture, water content, etc.), the cooking requirements set by the user, and the status of the cooking equipment. This ensures that each dish can be precisely controlled during the cooking process, avoiding cooking failures caused by differences in ingredients or equipment fluctuations, and significantly improving the quality and taste of the dishes.

[0020] Furthermore, this patent helps to improve the user's dietary health. By combining user health data and dish nutritional data, the system can fully consider nutritional balance when recommending dishes and determining cooking parameters. For example, for users with high blood pressure, the system will recommend low-sodium dishes and control the amount of salt; for users who need protein supplements, the system will recommend protein-rich dishes and ensure the retention of protein during the cooking process. This health-oriented cooking control can provide protection for the user's long-term health.

[0021] Finally, this patent can improve the intelligence level and market competitiveness of cooking robots. As people pursue quality of life and intelligent products, this cooking robot that can provide personalized, healthy, and high-quality cooking experience will be more popular with consumers and has broad market prospects. At the same time, its intelligent control method also provides new ideas and directions for the development of future cooking technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a schematic block diagram of a cooking robot control method based on a knowledge graph proposed in the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise" and "counterclockwise" indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0025] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0026] Reference Figure 1 : A cooking robot control method based on knowledge graph, comprising the following steps: Knowledge graph construction steps: Collect multi-source cooking data, which includes food data, seasoning data, cooking equipment data, cooking process data, user health data and dish nutrition data; the food data covers food type, origin, seasonal information, and food characteristics, such as texture, water content, oil content, and ease of cooking, and is represented in vector form as , represents the 1st to mth food characteristics; the seasoning data includes the type of seasoning, taste, saltiness, and spiciness information, recorded as , Indicates the 1st to nth seasoning attributes; cooking equipment data includes equipment type, power, heating method, and capacity information, expressed as , e1-e k represents the 1-kth equipment parameters; cooking process data involves cooking methods, cooking time range, heat requirements, operation steps, and It means that p1-p q represents the 1-qth process attributes; user health data includes the user's allergic ingredients, dietary taboos, and chronic disease information, recorded as , h1-h u represents the 1-uth health attribute; the nutritional data of the dish includes the calorie, protein content, fat content, and vitamin content of the dish, recorded as , n1-n w Represents the 1-w nutritional attributes, uses semantic analysis and entity recognition technology to process the collected data and construct a cooking knowledge graph , where V is a node set, including the ingredient node V F , Seasoning Node V S , cooking equipment node VE , Cooking Process Node V P User Health Node V H And the food nutrition node V N , E is the edge set, which represents the semantic relationship between nodes, including the matching relationship between ingredients and seasonings, the adaptation relationship between ingredients and cooking techniques, the relationship between user health status and ingredient taboos, and the weight of the relationship. Determined through statistical analysis of cooking data and expert experience evaluation, it indicates the closeness of the relationship. The weight calculation method is: ,in represents the number of times entity i and entity j appear together in the data set, and J is the number of all entities related to entity i; Dish recommendation step: receiving the demand information input by the user, the demand information includes the number of diners, dish taste preferences, ingredient preferences, cooking time requirements, nutritional requirements, and budget constraints; searching and matching based on the user demand information in the cooking knowledge graph G, using a recommendation algorithm based on multi-attribute fusion, and calculating the matching degree M between the user demand and the dish information in the graph. The matching degree calculation formula is: , where α, β, γ, δ, and ε are weight coefficients, which respectively represent the importance of ingredients, seasonings, cooking techniques, user health, and dish nutrition in the matching process, and , M F is the food matching degree, M S is the seasoning matching degree, M P is the cooking process matching degree, M H For user health matching, M N is the nutritional matching degree of the dishes, and the calculation methods are , , in and They represent the normalized values ​​of the user demand and the i-th ingredient characteristic in the dish, respectively, and m represents the number of ingredient characteristics. represents the weight of food characteristics, and They represent the user demand and the normalized value of the j-th seasoning attribute in the dish, respectively. represents the weight of the seasoning attribute, n represents the number of seasoning attributes, and Represent the normalized values ​​of user demand and the rth process attribute in the dish, respectively. and represent the normalized value of the user demand and the vth health attribute in the dish, respectively. and They represent the user's needs and the normalized value of the x-th nutritional attribute in the dish, respectively. The top y dish information C with the highest matching degree is recommended to the user, and y is set according to the user's needs or the system default.

[0027] Cooking process determination step: According to the recommended dish information C, query the corresponding cooking process information in the cooking knowledge graph G, including the detailed cooking step sequence, cooking time required for each step, cooking temperature, heat control requirements, seasoning order and amount, and the relationship between cooking time and ingredient type, quantity and cooking equipment power is: ,in is the time coefficient, is the number of food types in step i, is the total amount of ingredients in step i, is the power of the cooking equipment used in step i, As the basic time constant; determine the temperature control parameters through the fuzzy logic reasoning system. The system input is the characteristics of the ingredients and the cooking process, and the output is the temperature control parameters, including the fire size and the heating time interval. After fuzzifying the input variables, a fuzzy rule base is formulated based on expert experience, including: IF the moisture content of the ingredients is high AND the cooking process is stir-fry THEN the fire size is large AND the heating time interval is short. The fuzzy reasoning engine is used for reasoning, and finally the specific temperature control parameter values ​​are obtained through defuzzification. Cooking control steps: Send the determined cooking process information to the controller of the cooking robot. The controller executes each step in sequence according to the cooking step sequence. During the execution process, the on and off time of the cooking equipment is controlled according to the cooking time information. The heating power is controlled according to the cooking temperature information to maintain the target temperature. The flame size or the power output of the heating element is adjusted according to the fire control requirements. The seasoning feeding device is controlled according to the seasoning feeding sequence and feeding amount information. The cooking environment data and cooking equipment operation data are collected in real time through sensors and compared with the preset standard cooking data. The adaptive feedback control algorithm is used to adjust the cooking parameters. The feedback control algorithm formula is: ,in Adjust the amount for cooking parameters, and is the feedback coefficient, For cooking environment data, is the standard cooking environment data, Running data for cooking equipment, Run data for standard cooking equipment to ensure that the cooking process is as expected and the dishes meet user needs.

[0028] In the present invention, in the knowledge graph construction step, the sources of food data collection include food supplier database, agricultural product market data platform, agricultural Internet of Things sensor data and user manual input. The number of collected food types is no less than 500, the origin information is accurate to the county level, and the seasonal information is accurate to the month.

[0029] In the present invention, in the knowledge graph construction step, the sources of seasoning data collection include seasoning manufacturer database, food seasoning research institution data, e-commerce platform sales data and user taste feedback data. The spiciness information is measured using Scoville Spiciness Units (SHU), and the saltiness is expressed as a percentage of salt content.

[0030] In the present invention, in the dish recommendation step, user demand information is collected in a variety of ways such as voice recognition, text input, and image recognition, such as recognizing pictures of ingredients or dishes selected by users. The taste preference characteristic dimensions include at least 9 dimensions such as sourness, sweetness, bitterness, spiciness, saltiness, freshness, greasiness, and crispness. The nutritional needs in the demand information are connected with the nutrition expert system to recommend a suitable range of nutritional intake based on the user's age, gender, and physical condition, such as whether the user has high blood pressure, diabetes, or obesity.

[0031] In the present invention, in the cooking process determination step, in the cooking time calculation, the time coefficient has a value range of 0.5 to 5, and the basic time constant has a value range of 0 to 5 minutes. In the cooking temperature calculation, the temperature coefficient has a value range of 10 to 200, and the basic temperature constant has a value range of 30°C to 250°C. In the seasoning amount calculation, the amount coefficient has a value range of 0.1 to 10, and the basic amount constant has a value range of 0 to 20 grams.

[0032] In the present invention, in the cooking control step, the sensors include temperature sensors, humidity sensors, image sensors, oil fume concentration sensors, and weight sensors. The operation data of the cooking equipment is acquired through the built-in sensors and controllers of the equipment. The feedback coefficient has a value range of 0.1 to 1, and a value range of 0.05 to 0.5. Dynamic adjustments are made according to different stages of the cooking process (such as preheating stage, food processing stage, seasoning stage, and out-of-pot stage).

[0033] In the present invention, the knowledge graph construction step also includes a knowledge update mechanism, which regularly obtains new cooking knowledge from Internet cooking forums, food blogs, and professional cooking research literature data sources, such as emerging food combinations and innovative cooking techniques, and integrates new knowledge into the existing cooking knowledge graph through entity alignment and relationship update algorithms. The entity alignment algorithm uses a hybrid algorithm based on semantic similarity and structural similarity to calculate the similarity between new entities and existing entities. If the similarity exceeds a set threshold, a merge or association operation is performed. The relationship update algorithm adjusts the weight and type of the relationship between nodes according to the new knowledge data.

[0034] In the present invention, in the cooking process determination step, for dishes with multiple cooking process options, such as steamed fish and braised fish, process comparison and recommendation reasons are provided, the process comparison includes taste comparison, such as steamed fish has a fresh and tender taste, and braised fish has a rich taste, nutrition retention comparison, such as steamed fish retains more complete nutrition, and some nutrients of braised fish will be lost during the cooking process, cooking difficulty comparison, such as steamed fish has a lower cooking difficulty, and braised fish has a higher cooking difficulty, cooking time comparison, such as steamed fish has a shorter cooking time, and braised fish has a longer cooking time. The recommendation reason is based on user demand information, such as if the user pays attention to health, the steaming process is recommended, and if the user likes rich taste, the braised process is recommended, and data statistical analysis in the cooking knowledge map, such as a certain process has a higher satisfaction in user evaluations.

[0035] In the present invention, in the cooking control step, when an abnormal situation occurs during the cooking process, such as excessive temperature, equipment failure, or burnt ingredients, the cooking robot can automatically handle it according to preset exception handling rules, such as reducing heating power, stopping cooking, and issuing an alarm to notify the user. At the same time, the abnormal situation information, including the type of abnormality, time of occurrence, and related cooking parameters, is recorded in a log file for subsequent analysis and optimization of the cooking process. The exception handling rules are derived through the analysis and summary of a large amount of cooking experimental data and actual usage data, and a decision tree algorithm is used to construct an exception handling model. The input of the model is abnormal situation characteristics, such as temperature change rate, equipment operating current, and oil fume concentration change, and the output is the processing measures.

[0036] The present invention also discloses a cooking robot control system based on a knowledge graph, comprising: Knowledge graph construction module: used to implement the knowledge graph construction steps and construct the cooking knowledge graph G.

[0037] Dish recommendation module: used to implement the dish recommendation steps and recommend dish information according to user needs.

[0038] Cooking process determination module: used to implement cooking process determination steps and determine cooking process information.

[0039] Cooking control module: used to implement cooking control steps and control the cooking process of the cooking robot.

[0040] Data acquisition module: used to collect cooking data, user demand information, cooking environment data and cooking equipment operation data, and provide data support for other modules. The collected data is cleaned, denoised and format converted through the data preprocessing unit to ensure the accuracy and availability of the data.

[0041] Storage module: used to store cooking knowledge graphs, dish information, cooking process information and data collected by the data acquisition module. The storage capacity is not less than 500GB. It adopts a distributed storage system and has data backup and recovery functions to prevent data loss.

[0042] Human-computer interaction module: provides an interactive interface between users and the cooking robot, including a display screen, voice interaction equipment, and a touch control panel. Users can use this module to input demand information, view recommended dishes, monitor the cooking process, and obtain cooking tips and suggestions.

[0043] Communication module: used to realize the communication connection between the cooking robot and external devices, such as smart phones, smart home systems, and cloud servers. It supports Bluetooth, Wi-Fi, and 4G / 5G communication protocols to receive remote control instructions, upload cooking data, and download updated data such as updated knowledge graphs and cooking programs.

[0044] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can replace or change the technical solution and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A cooking robot control method based on knowledge graph, characterized in that: The following steps are involved: Knowledge graph construction step: collecting multi-source cooking data, wherein the multi-source cooking data includes food material data, seasoning data, cooking equipment data, cooking process data, user health data and dish nutrition data; The food data includes food type, origin, seasonal information, and food characteristics, which are expressed in vector form as follows: , represents the 1st to mth food characteristics; the seasoning data includes the type of seasoning, taste, saltiness, and spiciness information, recorded as , Indicates the 1st to nth seasoning attributes; cooking equipment data includes equipment type, power, heating method, and capacity information, expressed as , e1-e k represents the 1-kth equipment parameters; cooking process data involves cooking methods, cooking time range, heat requirements, operation steps, and It means that p1-p q represents the 1-qth process attributes; user health data includes the user's allergic ingredients, dietary taboos, and chronic disease information, recorded as , h1-h u represents the 1-uth health attribute; the nutritional data of the dish includes the calorie, protein content, fat content, and vitamin content of the dish, recorded as , n1-n w Represents the 1-w nutritional attributes, uses semantic analysis and entity recognition technology to process the collected data and construct a cooking knowledge graph , where V is a node set, including the ingredient node V F , Seasoning Node V S , cooking equipment node V E , Cooking Process Node V P User Health Node V H And the food nutrition node V N , E is the edge set, which represents the semantic relationship between nodes, including the matching relationship between ingredients and seasonings, the adaptation relationship between ingredients and cooking techniques, the relationship between user health status and ingredient taboos, and the weight of the relationship. Determined through statistical analysis of cooking data and expert experience evaluation, it indicates the closeness of the relationship. The weight calculation method is: ,in represents the number of times entity i and entity j appear together in the data set, and J is the number of all entities related to entity i; Dish recommendation step: receiving the demand information input by the user, the demand information includes the number of diners, dish taste preferences, ingredient preferences, cooking time requirements, nutritional requirements, and budget constraints; searching and matching based on the user demand information in the cooking knowledge graph G, using a recommendation algorithm based on multi-attribute fusion, and calculating the matching degree M between the user demand and the dish information in the graph. The matching degree calculation formula is: , where α, β, γ, δ, and ε are weight coefficients, which respectively represent the importance of ingredients, seasonings, cooking techniques, user health, and dish nutrition in the matching process, and , M F is the food matching degree, M S is the seasoning matching degree, M P is the cooking process matching degree, M H For user health matching, M N is the nutritional matching degree of the dishes, and the calculation methods are , , in and They represent the normalized values ​​of the user demand and the i-th ingredient characteristic in the dish, respectively, and m represents the number of ingredient characteristics. represents the weight of food characteristics, and They represent the user demand and the normalized value of the j-th seasoning attribute in the dish, respectively. represents the weight of the seasoning attribute, n represents the number of seasoning attributes, and Represent the normalized values ​​of user demand and the rth process attribute in the dish, respectively. and represent the normalized value of the user demand and the vth health attribute in the dish, respectively. and They represent the normalized value of the user demand and the xth nutritional attribute in the dish respectively; Cooking process determination step: According to the recommended dish information, query the corresponding cooking process information in the cooking knowledge graph G, including the detailed cooking step sequence, cooking time required for each step, cooking temperature, heat control requirements, seasoning order and amount, and the relationship between cooking time and ingredient type, quantity and cooking equipment power is: ,in is the time coefficient, is the number of food types in step i, is the total amount of ingredients in step i, is the power of the cooking equipment used in step i, As the basic time constant; determine the temperature control parameters through the fuzzy logic reasoning system. The system input is the characteristics of the ingredients and the cooking process, and the output is the temperature control parameters, including the fire size and the heating time interval. After fuzzifying the input variables, a fuzzy rule base is formulated based on expert experience, including: IF the moisture content of the ingredients is high AND the cooking process is stir-fry THEN the fire size is large AND the heating time interval is short. The fuzzy reasoning engine is used for reasoning, and finally the specific temperature control parameter values ​​are obtained through defuzzification. Cooking control steps: Send the determined cooking process information to the controller of the cooking robot. The controller executes each step in sequence according to the cooking step sequence. During the execution process, the on and off time of the cooking equipment is controlled according to the cooking time information. The heating power is controlled according to the cooking temperature information to maintain the target temperature. The flame size or the power output of the heating element is adjusted according to the fire control requirements. The seasoning feeding device is controlled according to the seasoning feeding sequence and feeding amount information. The cooking environment data and cooking equipment operation data are collected in real time through sensors and compared with the preset standard cooking data. The adaptive feedback control algorithm is used to adjust the cooking parameters. The feedback control algorithm formula is: ,in Adjust the amount for cooking parameters, and is the feedback coefficient, For cooking environment data, is the standard cooking environment data, Running data for cooking equipment, Run data for standard cooking equipment to ensure that the cooking process is as expected and the dishes meet user needs.

2. According to claim 1, a cooking robot control method based on knowledge graph is characterized in that: In the knowledge graph construction step, the sources of food data collection include food supplier database, agricultural product market data platform, agricultural Internet of Things sensor data and user manual input. The number of collected food types is no less than 500, the origin information is accurate to the county level, and the seasonal information is accurate to the month.

3. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the knowledge graph construction step, the sources of seasoning data collection include seasoning manufacturer database, food seasoning research institution data, e-commerce platform sales data and user taste feedback data. The spiciness information is measured using Scoville spiciness units (SHU), and the saltiness is expressed as a percentage of salt content.

4. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the dish recommendation step, user demand information is collected through voice recognition, text input, and image recognition. The taste preference characteristic dimensions include at least the sourness, sweetness, bitterness, spiciness, saltiness, freshness, greasiness, and crispness. The nutritional needs in the demand information are connected to the nutrition expert system to recommend a suitable range of nutritional intake based on the user's age, gender, and physical condition.

5. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the cooking process determination step, in the cooking time calculation, the time coefficient ranges from 0.5 to 5, and the basic time constant ranges from 0 to 5 minutes. In the cooking temperature calculation, the temperature coefficient ranges from 10 to 200, and the basic temperature constant ranges from 30°C to 250°C. In the seasoning amount calculation, the amount coefficient ranges from 0.1 to 10, and the basic amount constant ranges from 0 to 20 grams.

6. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the cooking control step, the sensors include temperature sensors, humidity sensors, image sensors, oil fume concentration sensors, and weight sensors. The cooking equipment operation data is obtained through the built-in sensors and controllers of the equipment. The feedback coefficient ranges from 0.1 to 1, and the range of is 0.05 to 0.

5. It is dynamically adjusted according to the different stages of the cooking process.

7. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: The knowledge graph construction step also includes a knowledge update mechanism, which regularly obtains new cooking knowledge from Internet cooking forums, food blogs, and professional cooking research literature data sources, such as emerging ingredient combinations and innovative cooking techniques. The new knowledge is integrated into the existing cooking knowledge graph through entity alignment and relationship update algorithms. The entity alignment algorithm uses a hybrid algorithm based on semantic similarity and structural similarity to calculate the similarity between new entities and existing entities. If the similarity exceeds the set threshold, a merge or association operation is performed. The relationship update algorithm adjusts the weight and type of the relationship between nodes according to the new knowledge data.

8. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the cooking process determination step, for dishes with multiple cooking process options, process comparisons and recommendation reasons are provided. The process comparisons include taste comparisons, nutrient retention comparisons, cooking difficulty comparisons, and cooking time comparisons. The recommendation reasons are based on user demand information and statistical analysis of data in the cooking knowledge graph.

9. The cooking robot control method based on knowledge graph according to claim 1 is characterized in that: In the cooking control step, when an abnormal situation occurs during the cooking process, the cooking robot can automatically handle it according to the preset exception handling rules, including reducing the heating power, stopping cooking, and issuing an alarm to notify the user. At the same time, it records the abnormal situation information in the log file for subsequent analysis and optimization of the cooking process. The exception handling rules are derived through the analysis and summary of a large amount of cooking experimental data and actual usage data. The decision tree algorithm is used to construct the exception handling model. The input of the model is the abnormal situation characteristics, and the output is the handling measures.

10. A system for implementing the cooking robot control method based on knowledge graph according to any one of claims 1 to 9, characterized in that: include: Knowledge graph construction module: used to implement the knowledge graph construction step in claim 1 and construct a cooking knowledge graph G; Dish recommendation module: used to implement the dish recommendation step in claim 1 and recommend dish information according to user needs; A cooking process determination module: used to implement the cooking process determination step in claim 1 and determine the cooking process information; Cooking control module: used to implement the cooking control step in claim 1 and control the cooking process of the cooking robot; Data acquisition module: used to collect cooking data, user demand information, cooking environment data and cooking equipment operation data, providing data support for each module. The collected data is cleaned, denoised and format converted by the data preprocessing unit; Storage module: used to store cooking knowledge graphs, dish information, cooking process information, and data collected by the data acquisition module. The storage capacity is not less than 500GB. It adopts a distributed storage system and has data backup and recovery functions. Human-computer interaction module: provides an interactive interface between the user and the cooking robot, including a display screen, voice interaction equipment, and a touch control panel. Users can input demand information, view recommended dishes, monitor the cooking process, and obtain cooking tips and suggestions; Communication module: used to realize the communication connection between the cooking robot and external devices, supporting Bluetooth, Wi-Fi, 4G / 5G communication protocols, so as to receive remote control, including updating knowledge graphs and cooking programs.

Images