Cooking robot menu intelligent management method and system

Abstract

The invention discloses an intelligent menu management method and system for a cooking robot, and relates to the technical field of cooking robotics.The management method comprises the steps that a pre-trained operation semantic analysis model is utilized, the heat conduction characteristic of cookware and the heat capacity characteristic of food materials are combined, theoretical heat output power values of all operation stages are determined, and according to a time boundary matching mechanism, the theoretical heat output power values of all the operation stages are calculated; determining a theoretical heat output power time function of the current original menu to execute a heating operation; determining an actual heat output power path function according to the collected actual temperature change data of the bottom of the cooker of the cooking robot; comparing and analyzing the heat output power path function with a theoretical heat output power path function of the current original menu, judging whether the currently executed heating operation has a heating degree deviation risk or not, and correspondingly sending out a heating degree response regulation and control instruction; and performing response regulation and control on the theoretical heat output power path function of the current original menu in time, and updating the theoretical heat output power path function to a menu database in the cooking robot.

Description

Technical Field

[0001] The present invention relates to the technical field of cooking robots, and in particular to a cooking robot recipe intelligent management method and system. Background Art

[0002] With the deep integration of artificial intelligence and Internet of Things technologies, the application of intelligent cooking robots in home and commercial kitchens is becoming increasingly widespread, and has become a key component of the development trend of smart kitchens. In the field of cooking robots, intelligent recipe management methods, as their core logical means, directly affect the heating control, heat judgment and food quality of the entire machine. However, the recipe management methods currently built into cooking robots are mostly static rule control or experience template matching, lacking an in-depth understanding of the semantics of the recipe text and the ability to dynamically correct the heating behavior process, which affects the flavor restoration of the actual dishes.

[0003] Regarding the recipe management mechanism in cooking robots, existing technologies mostly use a static combination of preset temperature and time to generate dish processing instructions, failing to achieve structured analysis and dynamic adaptation of the operational semantics, heat description and behavioral stage information in the original recipe text; at the same time, the existing recipe management mechanism lacks real-time perception of the response changes of the pot temperature during the actual heating process, and lacks an effective comparison mechanism between the theoretical power path and the actual power path, making it difficult to judge the heat response status, and even more difficult to perform dynamic adjustments and optimizations, resulting in a significant deviation between the recipe standard and the actual heating execution. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a cooking robot recipe intelligent management method and system, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A cooking robot recipe intelligent management method, comprising the following steps: S1. Based on the current original recipe text in the recipe database, a pre-trained operational semantic parsing model is used to parse the operation phase sequence parameters of the current recipe. The theoretical heat output power value for each operation phase is determined by combining the heat conductivity of the cookware and the heat capacity of the ingredients. The theoretical heat output power time function of the current original recipe is then determined based on a time boundary matching mechanism to execute the heating operation. S2. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; S3. Compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function to determine whether there is a risk of temperature deviation in the heating operation performed by the cooking robot for the current original recipe, and issue a corresponding temperature response control instruction; S4. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe is time-responsively controlled and updated to the recipe database in the cooking robot.

[0006] Preferably, the specific steps of S1 include: S11. Based on the current original recipe text in the recipe database, and using the pre-trained operational semantic parsing model, parse the operation phase sequence parameters of the current recipe. The specific parsing method is as follows: The cooking robot uses the current original recipe text stored in the recipe database as the initial input. The current original recipe text includes the behavioral terms of each operation stage, the time nodes of the ingredients, and the description of the heat; By constructing a ternary mapping relationship of "action verb-control target-heat parameter" based on various standard recipe texts stored in the recipe database, combined with the contextual semantic modeling task, and using a multi-layer encoding network to perform semantic embedding learning and control parameter alignment training for each operation stage in the recipe text, a pre-trained operational semantic parsing model is obtained. Based on the current original recipe text and using the pre-trained operational semantic parsing model, the stage number, action category, target temperature and temperature maintenance duration of each operation stage in the current original recipe are identified and labeled, and the operation stage sequence parameters of the current original recipe are constructed.

[0007] Preferably, S12, by extracting the characteristics of the operation stage sequence parameters of the current original recipe and combining the heat conductivity characteristics of the cookware and the heat capacity characteristics of the ingredients, the theoretical heat output power value of each operation stage is determined, specifically: Where Pgv i Indicates the theoretical thermal output power value of the corresponding operation stage, St i Indicates the temperature maintenance time of the corresponding operation stage, Tmb i Indicates the target temperature of the corresponding operation stage, Crr indicates the heat capacity of the ingredients in the current original recipe, Thc indicates the current ambient temperature, and Ncd indicates the heat conduction efficiency of the cookware.

[0008] Preferably, S13, based on the theoretical heat output power values ​​of each operation stage determined in S12 and in combination with the time boundary matching mechanism, the theoretical heat output power time function of the current original recipe is determined to perform the heating operation. The specific process is as follows: By performing feature recognition on the operation phase sequence parameters of the current original recipe, the temperature maintenance start time and temperature maintenance end time of each operation phase are extracted to obtain the effective temperature control time interval of each operation phase. Based on this, a time boundary indicator function is set. The time boundary indicator function is used to determine whether each moment falls within the effective temperature control time interval of the corresponding operation phase on the time axis of each operation phase of the current original recipe; According to the set time boundary indicator function and combined with the theoretical thermal output power value of each operation stage, the theoretical thermal output power path function is constructed, specifically: Where Dhc(t) represents the theoretical thermal output power path function, Pgv i represents the theoretical thermal output power value of the i-th operating stage, represents the time boundary indicator function, which is used to determine whether the tth moment falls within the effective temperature control time interval of the ith operation stage. If so, it is 1, otherwise it is 0; After the constructed theoretical thermal output power path function is encoded through the pulse width modulation control mechanism, the thermal power input signal sequence of the current original recipe is generated by the path reconstruction vector and transmitted to the control port of the heating device through the drive interface to perform the heating operation.

[0009] Preferably, the specific steps of S2 include: S21. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the heat conduction principle inverse model. The specific process is as follows: During the heating process, the infrared temperature sensor deployed at the bottom of the cooking robot's pot collects the actual temperature change data sequence in real time and constructs the pot body temperature response function; Based on the heat capacity characteristics of the ingredients in the current original recipe, combined with the constructed pot temperature response curve function, and using the heat conduction principle to infer the model, the actual heat output power path function is determined, specifically: Where Dsj(t) represents the actual heat output power path function, Tsj(t) represents the pot temperature response function, represents the rate of temperature rise of the pot at time t, Crr represents the heat capacity of the ingredients in the current original recipe, and Ncd represents the heat conduction efficiency of the pot.

[0010] Preferably, the specific steps of S3 include: S31. Compare and analyze the theoretical thermal output power path function of the current original recipe with the corresponding actual thermal output power path function. After dimensionless processing, determine the degree of deviation of the real-time thermal output power of the heating operation performed by the cooking robot for the current original recipe at the corresponding moment, specifically: Tpc(t)=Dhc(t)-Dsj(t); where Tpc(t) represents the thermal output power deviation value at moment t, Dhc(t) represents the theoretical thermal output power path function, and Dsj(t) represents the actual thermal output power path function.

[0011] Preferably, S32, judging whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe based on the degree of real-time power deviation of the heating operation, the specific analysis process is as follows: If the heat output power at the corresponding moment is equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is normal, and the corresponding moment is marked as a normal heating operation moment. If the heat output power deviation value at the corresponding moment is not equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is abnormal, and the corresponding moment is marked as an abnormal heating operation moment. The number of moments marked as abnormal heating operations is counted to obtain the number of abnormal heating operation moments. When the number of abnormal heating operation moments exceeds 5% of the total number of moments within the heating time, it indicates that there is a risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, a temperature response control instruction is issued. Otherwise, it indicates that there is no risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, the current original recipe stored in the recipe database of the cooking robot is marked as a standard temperature recommended recipe.

[0012] Preferably, the specific steps of S4 include: S41. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe arranged in time series is compared with the corresponding actual heat output power path function on the time axis, and combined with the sliding window difference method, the time difference between the moment when the theoretical heat output power rises and the corresponding moment when the actual heat output power rises is identified to obtain the thermal response lag time difference at each moment.

[0013] Preferably, in S42, according to the thermal response lag time difference at each moment in S41, the theoretical thermal output power path function of the current original recipe is time-controlled to construct the adjusted thermal output power path function of the current original recipe, specifically: Dth(t)=Dhc[t-Ts(t)]; where Dth(t) represents the adjusted thermal output power path function, Ts(t) represents the thermal response lag time difference at the corresponding moment, and [t-Ts(t)] represents the adjusted time at the corresponding moment; S43. After encoding the adjusted heat output power path function of the current original recipe through a pulse width modulation control mechanism, the adjusted heat power input signal sequence of the current original recipe is generated by the path reconstruction vector, and combined with the operational semantics reverse parsing model, a standard heat recommended recipe text is generated, updated to the recipe database in the cooking robot, and the current original recipe corresponding to the adjusted heat output power path function is marked as a standard heat recommended recipe.

[0014] An intelligent cooking robot recipe management system, comprising a recipe parsing module, a path analysis module, a heat analysis module and a recipe management module; The recipe parsing module is used to parse the operation phase sequence parameters of the current recipe based on the current original recipe text in the recipe database using a pre-trained operational semantic parsing model. It then determines the theoretical heat output power value for each operation phase by combining the heat conductivity of the cookware and the heat capacity of the ingredients. Based on the time boundary matching mechanism, it determines the theoretical heat output power time function of the current original recipe to execute the heating operation. The path analysis module is used to determine the actual heat output power path function during the heating process based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; The heat analysis module is used to compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function, determine whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe, and issue a corresponding heat response control instruction; The recipe management module is used to perform time response control on the theoretical heat output power path function of the current original recipe after receiving the heat response control instruction, and update it to the recipe database in the cooking robot.

[0015] The present invention provides a cooking robot recipe intelligent management method and system, which has the following beneficial effects: (1) The present invention proposes a cooking robot recipe intelligent management method and system. In response to the problems of weak recipe semantic understanding ability, delayed heating control response and lack of dynamic heat control ability in the existing technology, a complete control mechanism covering operation semantic analysis, theoretical power modeling, actual power back-calculation, heat deviation judgment and path control update is constructed, which effectively improves the behavioral intelligence level of the cooking robot in the actual cooking process. By converting the recipe text into structured control parameters, the robot can understand and execute semantically clear heating stage operations, avoiding heat deviation caused by improper template matching. At the same time, a thermal power back-calculation model based on temperature response function is introduced to achieve continuous tracking and fine modeling of the actual heating path. Combined with the real-time deviation judgment mechanism, it supports dynamic issuance of control instructions during the heating process and responsively updates the original recipe, thereby improving the recipe adaptability and heating accuracy. In summary, the present invention not only breaks through the static control bottleneck of traditional recipe management, but also realizes closed-loop feedback control between the recipe model and the actual execution status, comprehensively enhancing the cooking robot's dish restoration ability and adaptability in complex food combinations and dynamic environments, and has significant improvements in intelligence and practicality.

[0016] (2) By introducing the operational semantic parsing model, the multi-dimensional semantic information existing in the original recipe text is deeply modeled, and a semantic mapping system of the "action verb-control target-heat parameter" ternary structure is constructed. The semantic embedding and control parameter extraction are carried out in conjunction with the multi-layer encoder to realize the automatic numbering of the operation stages, action classification and thermal target reconstruction, breaking through the problem of insufficient control accuracy caused by the traditional keyword retrieval and template matching method; on this basis, the theoretical thermal output power value of each operation stage is accurately established based on the thermal power theoretical formula, combining the thermal conductivity parameters of the cookware and the heat capacity characteristics of the ingredients, and introducing the time boundary indicator function to construct a theoretical thermal power path function covering the entire process, realizing the staged heating control capability based on time; this mechanism significantly improves the robot's semantic understanding and control conversion capabilities when facing complex recipes, and can automatically generate a heating operation path with strong adaptability and high control accuracy based on the semantic content, providing an accurate target reference for subsequent actual execution.

[0017] (3) By deploying temperature sensors at the bottom of the cooking robot pot, the real-time pot body temperature response sequence is collected and a temperature response function is constructed. On this basis, a model based on heat conduction inverse is introduced to calculate the actual heat output power path function; then the path function is compared with the corresponding theoretical path function moment by moment to construct a dimensionless heat output power deviation value sequence, and the abnormal heating moment and the heat deviation trend are identified; by introducing a deviation risk threshold judgment mechanism, an effective evaluation is made on whether the actual heating execution deviates from the theoretical control trajectory. If there is a significant deviation, the heat response control instruction is triggered, and the hysteresis response identification and path correction process is entered to construct the adjusted heat output power path function; finally, the adjusted recipe text is generated through the linkage of path reconstruction and semantic reverse parsing model, and the recipe database is automatically updated and recommended. The mechanism realizes closed-loop feedback control between recipe execution and actual behavior, so that the cooking robot has the ability to adapt, self-correct and self-update. Under different dishes, environments and execution errors, the recipe heat control logic can be dynamically adjusted to improve the dish restoration and heat consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of an intelligent recipe management method for a cooking robot according to the present invention; Figure 2 This is a block diagram of an intelligent cooking robot recipe management system of the present invention; Figure 3 This is a partial logical thinking diagram of a cooking robot recipe intelligent management method of the present invention; Figure 4 This is a simplified diagram of the overall logical thinking of an intelligent recipe management method for a cooking robot according to the present invention. DETAILED DESCRIPTION

[0019] 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] Example 1

[0021] See also Figure 1 and Figure 4 The present invention provides a cooking robot recipe intelligent management method, comprising the following steps: S1. Based on the current original recipe text in the recipe database, a pre-trained operational semantic parsing model is used to parse the operation phase sequence parameters of the current recipe. The theoretical heat output power value for each operation phase is determined by combining the heat conductivity of the cookware and the heat capacity of the ingredients. The theoretical heat output power time function of the current original recipe is then determined based on a time boundary matching mechanism to execute the heating operation. S2. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; S3. Compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function to determine whether there is a risk of temperature deviation in the heating operation performed by the cooking robot for the current original recipe, and issue a corresponding temperature response control instruction; S4. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe is time-responsively controlled and updated to the recipe database in the cooking robot.

[0022] In this embodiment, based on the deep integration of recipe semantic analysis and thermal output power path modeling, the intelligent control ability and dish restoration accuracy of the cooking robot in the cooking process are significantly improved; through the structured analysis of the operation stages of the original recipe text in S1, the robot can not only accurately identify the target temperature and control time of each heating stage, but also combine the thermal conductivity characteristics of the pot and the heat capacity parameters of the ingredients to establish a corresponding theoretical thermal power function to ensure that the heating process has a scientific basis and temperature rhythm control logic; further introduce the heat conduction inverse model in S2 to construct the actual thermal output power path function, realize the dynamic modeling of the robot's real heating behavior, and effectively overcome the defect of the traditional system's insufficient perception of the pot temperature change; in S3, by The theoretical and actual paths are compared in time sequence to accurately identify the risk of temperature deviation and respond immediately, so that the system has the ability to control temperature in real time; finally, a time response adjustment mechanism for the recipe thermal power path is established in S4, so that the original recipe can be automatically corrected and dynamically updated according to actual feedback, forming a self-optimizing closed-loop control system; generally speaking, the core benefit of the present invention lies in the construction of an integrated semantic-driven thermal control mechanism from "recipe text understanding-theoretical modeling-behavior inference-risk identification-adaptive update", which not only breaks the execution limitations of previous static rule recipes, but also realizes the dynamic optimization of recipe content and intelligent guarantee of temperature accuracy, enhancing the control accuracy and dish consistency of the cooking robot in response to changing ingredients and environmental conditions.

[0023] Example 2

[0024] Please refer to Figure 1 and Figure 3 , specifically: S1 specific steps include: S11. Based on the current original recipe text in the recipe database, and using the pre-trained operational semantic parsing model, parse the operation phase sequence parameters of the current recipe. The specific parsing method is as follows: The cooking robot uses the current original recipe text stored in the recipe database as the initial input. The current original recipe text includes the behavioral terms of each operation stage, the time nodes of the ingredients, and the description of the heat; It should be noted that the current raw recipe text refers to the original recipe content stored in the cooking robot's recipe database and has not yet undergone structured parsing. It usually includes natural language descriptions of each operation stage, including behavioral terms and heat descriptions. This text serves as the initial input for semantic parsing and is the basic data source for subsequently constructing operation stage sequence parameters and extracting control targets and temperature control parameters. Its role is to provide the cooking robot with native human recipe expressions, enabling the system to convert natural language into executable control instructions through semantic modeling, realizing an intelligent bridge from language understanding to behavioral execution. By constructing a ternary mapping relationship of "action verb-control target-heat parameter" based on various standard recipe texts stored in the recipe database, combined with the contextual semantic modeling task, and using a multi-layer encoding network to perform semantic embedding learning and control parameter alignment training for each operation stage in the recipe text, a pre-trained operational semantic parsing model is obtained. It should be noted that the pre-trained operational semantic parsing model refers to a model formed by constructing semantic triplets of "action verbs-control targets-heat parameters" based on a large number of standard recipe texts, and combining them with context modeling tasks, using a multi-layer encoding network to perform semantic embedding and control parameter alignment training on each operation stage in the recipe; this model has the ability to automatically identify key control information such as operation stage number, action category, target temperature and maintenance time from the original recipe text, and is the core tool for realizing the conversion from natural language to control logic; its role is to give the system the ability to understand the semantics of natural language recipes, so that the cooking robot can accurately convert unstructured recipes written by humans into structured heating execution parameters, thereby supporting refined and intelligent cooking control; the training process of this model relies on a large amount of annotated standard recipe data, and through deep learning technology, the mapping relationship between semantics and control parameters is continuously optimized in the embedding layer and alignment mechanism, ultimately forming a semantic recognition engine with general parsing capabilities; Based on the current original recipe text and using the pre-trained operational semantic parsing model, the stage number, action category, target temperature and temperature maintenance duration of each operation stage in the current original recipe are identified and labeled, and the operation stage sequence parameters of the current original recipe are constructed.

[0025] It should be noted that the operation phase sequence parameters refer to a set of structured control information extracted from the original recipe text and arranged in chronological order. Specifically, they include the stage number, action category, target temperature, and temperature maintenance duration of each operation phase. Their function is to convert the cooking steps described in natural language into quantitative control instructions that can be parsed and executed by the cooking robot. They are the basis for the subsequent construction of the thermal output power path function and the execution of heating control. This parameter set is automatically generated by a pre-trained operation semantic parsing model. This model is trained on semantic modeling based on a large amount of standard recipe data. It has the ability to identify key semantic elements in recipes and can accurately label and extract the control targets corresponding to each operation phase from the original text, thereby forming a continuous and executable control process chain. Specifically, S12, by extracting the characteristics of the operation phase sequence parameters of the current original recipe and combining the heat conductivity characteristics of the cookware and the heat capacity characteristics of the ingredients, the theoretical heat output power value of each operation phase is determined, specifically: Where Pgv i Indicates the theoretical thermal output power value of the corresponding operation stage, St iIndicates the temperature maintenance time of the corresponding operation stage, Tmb i Indicates the target temperature of the corresponding operation stage, Crr indicates the heat capacity of the ingredients in the current original recipe, Thc indicates the current ambient temperature, and Ncd indicates the heat conduction efficiency of the cookware.

[0026] It should be noted that the formula in S12 is of key significance in the intelligent control system of the cooking robot. It directly provides responsive solutions to the problems mentioned in the background technology, such as weak recipe comprehension ability, poor thermal control accuracy, and serious deviation between heat execution and recipe standards. This formula is used to calculate the theoretical thermal output power value of each operation stage, that is, the standard thermal energy intensity that should be output according to the recipe control target and the actual environmental conditions. It is the core bridge for achieving the quantitative connection between temperature control execution and recipe analysis. In the formula, Crr×(Tmb i -Thc) represents the total heat required to be absorbed by the unit food in the corresponding operation stage, Ncd×St i It represents the product of the heat transfer capacity of the cookware per unit time and the control time in the corresponding operation stage; by integrating these parameters into a rational structure, after parsing the original recipe text, it automatically generates quantitative heat control targets based on the specific ingredients and environmental characteristics, thus getting rid of the limitations of traditional static instructions; the generation of theoretical thermal output power values ​​enables the subsequent heating path function construction, actual power comparison analysis and fire deviation judgment to have a physical basis and dynamic adaptability, effectively realizing the precision of recipe control and the execution of semantic understanding, providing key support for the robot to achieve stable dish output under changing working conditions, and has a high degree of engineering practical value and semantic conversion logic support.

[0027] Specifically, in S13, based on the theoretical heat output power values ​​of each operation stage determined in S12 and in combination with the time boundary matching mechanism, the theoretical heat output power time function of the current original recipe is determined to perform the heating operation. The specific process is as follows: By performing feature recognition on the operation phase sequence parameters of the current original recipe, the temperature maintenance start time and temperature maintenance end time of each operation phase are extracted to obtain the effective temperature control time interval of each operation phase. Based on this, a time boundary indicator function is set. The time boundary indicator function is used to determine whether each moment falls within the effective temperature control time interval of the corresponding operation phase on the time axis of each operation phase of the current original recipe; It should be noted that the time boundary indicator function refers to a function tool used to identify whether a certain moment is within the effective temperature control range of a specific operating stage. Its core function is to perform boundary judgment on the start and end time of each heating stage to build the basic structure of time and power mapping. During the execution process, the function is usually expressed in the form of binary logic: when a certain moment falls within the temperature maintenance time range of a certain operating stage, the output is 1, indicating that the moment belongs to the heating control time interval of the stage; otherwise, the output is 0, indicating that it is not within the control interval. Its function is to structure the time information of the operating stage so that it can accurately load the corresponding thermal output power value according to the stage to which different moments belong, and then generate a theoretical thermal output power time function to realize timing-driven heating control. This function is established by identifying the start and end time of each stage in the operating stage sequence parameters and combining it with the time axis scanning strategy to ensure that the theoretical power can be switched according to the stage rhythm in different operating segments, providing a strict time logic foundation for subsequent pulse modulation control and path function construction. According to the set time boundary indicator function and combined with the theoretical thermal output power value of each operation stage, the theoretical thermal output power path function is constructed, specifically: Where Dhc(t) represents the theoretical thermal output power path function, Pgv i represents the theoretical thermal output power value of the i-th operating stage, represents the time boundary indicator function, which is used to determine whether the tth moment falls within the effective temperature control time interval of the ith operation stage. If so, it is 1, otherwise it is 0; It should be noted that this formula is the key expression for constructing the theoretical thermal output power path function, which reflects a systematic response to the core problems of the background technology, such as the lack of timing control capability, rough temperature adjustment, and the disconnection between recipes and heating behavior. In the formula, Dhc(t) represents the theoretical thermal power value that the robot should output at time t, which serves as a direct reference for heating control. Among them, Pgv i Indicates the theoretical heat output power value of the i-th operation stage, reflecting the specific demand for heat energy in different cooking stages. The time boundary indicator function is used to determine whether the current time point is within the effective temperature control time range of the i-th operation stage. This structure enables the precise loading of heat output instructions corresponding to each stage throughout the cooking process based on the time window, thereby forming a heat power control sequence with rhythmic, continuous, and instruction-level logic. The theoretical heat output power path function transforms the phased "action-temperature-duration" semantic parameters in the recipe into a theoretical power path with time indexing and function-driven characteristics, providing a continuous foundation for subsequent controller output, PWM modulation encoding, actual power comparison, and heat deviation judgment. Compared with traditional control methods based on static temperature control settings and single threshold triggering, the temporal dynamic characteristics and multi-stage fusion mechanism embodied by the theoretical heat output power path function not only improve the refinement of heat control, but also enhance the semantic consistency and execution coherence between the recipe and actual control. This gives the cooking robot a stronger semantic understanding and closed-loop execution capability, truly realizing the intelligent deconstruction and dynamic adaptive control of the original recipe content. After the constructed theoretical thermal output power path function is encoded through the pulse width modulation control mechanism, the thermal power input signal sequence of the current original recipe is generated by the path reconstruction vector and transmitted to the control port of the heating device through the drive interface to perform the heating operation.

[0028] In this embodiment, S1 proposes an intelligent heating control mechanism that integrates recipe semantic analysis, thermal power modeling and path function construction, which significantly improves the cooking robot's ability to understand the original recipe text and the control accuracy of the heating process; first, by constructing an operational semantic analysis model based on multi-layer encoding network training, it can semantically embed and structure the behavioral terms, ingredient nodes and heat descriptions in the original recipe, and automatically extract the key parameters of the operation stage number, action category, target temperature and maintenance time, fundamentally breaking through the traditional recipe analysis mode that relies on template rules or keyword matching, and realizing the accurate deconstruction of complex semantic content; secondly, combining the thermal conductivity of the cookware with the thermal conductivity of the ingredients corresponding to the current recipe, the cooking robot can realize the cooking process of the original recipe. Based on the thermal power derivation formula, the system can establish accurate theoretical thermal output power values ​​for each operation stage, so that the constructed thermal control model can truly reflect the thermal physical characteristics of the food and the environment; in addition, the effective temperature control time interval of each operation stage is extracted through the time boundary matching mechanism, and the time boundary indication function is set accordingly. Then, by integrating the theoretical power value, the system constructs a complete theoretical thermal output power path function, which can not only accurately cover the stage rhythm of the entire heating process, but also can be used for subsequent fire deviation judgment and control; finally, the path function is converted into a thermal power input signal sequence through the pulse width modulation mechanism, and the heating execution port is controlled by the path reconstruction vector, realizing the deep coupling between semantic analysis and physical execution.

[0029] Example 3

[0030] Please refer to Figure 1 , specifically: S2 specific steps include: S21. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the heat conduction principle inverse model. The specific process is as follows: During the heating process, the infrared temperature sensor deployed at the bottom of the cooking robot's pot collects the actual temperature change data sequence in real time and constructs the pot body temperature response function; It should be noted that the pot temperature response function refers to a functional expression that reflects the thermal response behavior of the pot over time, established after the cooking robot collects the temperature change data sequence of the bottom of the pot in real time through the infrared temperature sensor during the heating process; its core function is to describe the temperature change trajectory of the pot in different heating stages and under different energy input conditions, and it is a key intermediate variable for calculating the actual thermal output power; the pot temperature response function uses time as the independent variable and temperature or temperature rise rate as the dependent variable, and truly reflects the dynamic response process of the pot to thermal energy input; the pot temperature response function is obtained by using a high-frequency sampling infrared temperature sensor to collect the temperature value of the bottom of the pot throughout the cooking operation to form a time series data, and the series is numerically fitted or differentially processed to form a function model with time continuity and physical consistency. This function provides basic support for the subsequent derivation of the actual thermal output power path function based on the heat conduction inverse model, ensuring that the power estimation process has physical realism and time accuracy; Based on the heat capacity characteristics of the ingredients in the current original recipe, combined with the constructed pot temperature response curve function, and using the heat conduction principle to infer the model, the actual heat output power path function is determined, specifically: Where Dsj(t) represents the actual heat output power path function, Tsj(t) represents the pot temperature response function, represents the rate of temperature rise of the pot at time t, Crr represents the heat capacity of the ingredients in the current original recipe, and Ncd represents the heat conduction efficiency of the pot.

[0031] It should be noted that the actual thermal output power path function calculation process represented by this formula is a dynamic thermal power estimation model constructed based on the temperature rise response of the pot during the actual heating process. Its core purpose is to achieve continuous quantitative perception of the actual heating behavior of the cooking robot; in the formula, It is the derivative of the pot body temperature response function with respect to time, indicating the instantaneous heating rate of the pot body during the heating process. This model effectively solves the common problems of imperceptible heating behavior and unevaluable heating effect in background technology by converting the temperature change data collected by the pot's infrared temperature sensor into thermal power output information. It can not only read the temperature, but also understand the energy behavior corresponding to the temperature. Its logical basis is that the instantaneous output of thermal power should be equal to the product of the temperature rise rate and the heat absorption capacity of the food, and the heat transfer efficiency of the pot should be corrected to obtain the actual heat output level. The introduction of the actual thermal output power path function enables the cooking robot to have the ability to quantitatively describe the actual energy input trajectory, which provides key physical quantity support for subsequent theoretical and practical comparisons, fire deviation judgment and dynamic regulation, and is an important basis for realizing intelligent cooking closed-loop feedback control.

[0032] In this embodiment, by constructing a reverse modeling mechanism based on the real-time temperature response of the cookware, the cooking robot's real-time perception of the heating behavior and the accuracy of the actual heat output modeling during the execution process are significantly enhanced; specifically, by deploying an infrared temperature sensor at the bottom of the cookware, the temperature change data of the cookware during the heating process is continuously collected to form a cookware temperature response function, which effectively solves the problem that traditional cooking equipment lacks dynamic perception of the heating state; on this basis, combined with the heat capacity characteristics of the corresponding ingredients in the original recipe and using the principle of heat conduction, a functional relationship between the temperature rise rate of the cookware and the heat conduction efficiency is constructed, and the real-time actual heat output power path function is further inferred, which can be used as the thermal physical quantity. The path function not only has the characteristics of high resolution and high time precision, but also can completely reproduce the actual heating behavior trajectory of the robot in non-ideal environment, providing quantitative support for subsequent fire deviation judgment and regulation; the outstanding advantage of this step is that it realizes the intelligent closed loop from temperature change perception to thermal power inversion modeling, improves the adaptability to the thermal response differences of different ingredients, and ensures that the cooking robot can accurately model each actual execution behavior, thereby avoiding heating deviations caused by the thermal inertia of the pot, ambient temperature difference or equipment difference, and laying a solid foundation for achieving higher-fidelity dish restoration and adaptive cooking control.

[0033] Example 4

[0034] Please refer to Figure 1 , specifically: S3 specific steps include: S31. Compare and analyze the theoretical thermal output power path function of the current original recipe with the corresponding actual thermal output power path function. After dimensionless processing, determine the degree of deviation of the real-time thermal output power of the heating operation performed by the cooking robot for the current original recipe at the corresponding moment, specifically: Tpc(t)=Dhc(t)-Dsj(t); where Tpc(t) represents the thermal output power deviation value at moment t, Dhc(t) represents the theoretical thermal output power path function, and Dsj(t) represents the actual thermal output power path function.

[0035] It should be noted that the formula in S31 is used to calculate the thermal output power deviation value of the cooking robot at the t moment, which is the core function for realizing the judgment of the heat deviation. In the formula, Tpc(t) represents the instantaneous difference between the theoretical and actual thermal output powers, which represents the degree of deviation between the current execution state of the robot and the thermal control target of the preset recipe. Among them, Dhc(t) is the theoretical thermal output power path function constructed based on the recipe, and Dsj(t) is the actual thermal output power path function inferred by the temperature rise of the pot. The model compares the expectation and execution moment by moment in a dimensionless processing manner, which significantly solves the problem of background error. In order to solve the problems of difficulty in quantifying heating execution errors and vague standards for heat judgment in Jing technology, a refined judgment logic based on energy behavior differences has been established; its logical basis is that when the actual output power of the robot is lower than the theoretical value, it indicates heating lag and potential heat lag; by calculating the changing trend of the thermal output power deviation value, it can accurately identify abnormal deviations in the heating process, and provide a quantitative basis for subsequent heat regulation, time correction and adaptive update of recipes; it is the key fulcrum for achieving closed-loop control, improving the consistency of dishes and the degree of heat restoration, and gives the heat recognition real-time, accuracy and interpretation capabilities.

[0036] Specifically, S32, based on the degree of real-time power deviation of the heating operation, it is determined whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe. The specific analysis process is as follows: If the heat output power at the corresponding moment is equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is normal, and the corresponding moment is marked as a normal heating operation moment. If the heat output power deviation value at the corresponding moment is not equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is abnormal, and the corresponding moment is marked as an abnormal heating operation moment. The number of moments marked as abnormal heating operations is counted to obtain the number of abnormal heating operation moments. When the number of abnormal heating operation moments exceeds 5% of the total number of moments within the heating time, it indicates that there is a risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, a temperature response control instruction is issued. Otherwise, it indicates that there is no risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, the current original recipe stored in the recipe database of the cooking robot is marked as a standard temperature recommended recipe.

[0037] In this embodiment, by constructing a deviation evaluation mechanism between the theoretical thermal output power path function and the actual thermal output power path function, the cooking robot is able to recognize and judge the temperature deviation state in real time during the cooking process for the first time, effectively improving the temperature control capability of the robot in a dynamic cooking environment. Specifically, the theoretical and actual thermal power path functions are first compared at the same time, and a sequence of thermal output power deviation values ​​is generated through dimensionless processing to accurately measure the degree of deviation of the robot's heating behavior from the recipe target state at each moment. Compared with the traditional method of making simple judgments based on time or temperature setting thresholds, the deviation value analysis model introduced in the present invention has stronger time continuity and energy dimension matching, and can accurately capture small temperature fluctuations caused by food differences, environmental changes or execution errors. Furthermore, by The time at zero is marked as an abnormal operation state, and its proportion in the entire heating cycle is counted. If the proportion exceeds 5%, it is judged that there is a risk of temperature deviation in the current recipe execution, and a temperature response control instruction is issued in real time to form an early warning and control feedback closed loop; on the contrary, if no abnormal deviation occurs, the recipe can be marked as a standard temperature recommendation version and fed back to the recipe database to enhance the repeatability and stability of subsequent use; the outstanding advantage of this step is that by establishing a continuity difference recognition model in the power dimension, it breaks through the technical bottleneck that the traditional recipe execution results only rely on static temperature or time judgment, and gives the cooking robot the ability to actively identify temperature anomalies, which not only improves the adaptation accuracy of the response behavior of different ingredients, but also realizes the closed-loop quality verification and knowledge feedback mechanism of the recipe model and actual execution, fundamentally improving the consistency of dishes and the robustness of the system.

[0038] Example 5

[0039] Please refer to Figure 1 , specifically: S4 specific steps include: S41. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe arranged in time series is compared with the corresponding actual heat output power path function on the time axis, and combined with the sliding window difference method, the time difference between the moment when the theoretical heat output power rises and the corresponding moment when the actual heat output power rises is identified to obtain the thermal response lag time difference at each moment.

[0040] It should be noted that the thermal response lag time difference at each moment refers to the time difference between a certain rising moment in the theoretical thermal output power path and the corresponding rising moment in the actual thermal output power path during the heating process. It is a time offset indicator used to quantify the system response delay. Its function is to reflect the degree of thermal behavior reaction lag caused by the thermal inertia of the pot, sensor lag or control delay factors in the actual execution of the cooking robot. It is an important basis for realizing recipe control and time correction. The thermal response lag time difference is achieved by aligning and comparing the theoretical and actual thermal power path functions on the time axis, and using the sliding window difference method to identify the rising boundary of the power change, that is, finding the moment of rapid power rise in each theoretical heating stage, and performing difference calculation with the corresponding temperature rise response moment in the actual path, so as to obtain the thermal response lag time difference at that moment. This time difference result provides a quantitative basis for the subsequent construction of the adjusted thermal output path function, so that the recipe control logic can accurately compensate and correct the actual response characteristics.

[0041] Specifically, in S42, based on the thermal response lag time difference at each moment in S41, the theoretical thermal output power path function of the current original recipe is time-controlled to construct the adjusted thermal output power path function of the current original recipe, specifically: Dth(t)=Dhc[t-Ts(t)]; where Dth(t) represents the adjusted thermal output power path function, Ts(t) represents the thermal response lag time difference at the corresponding moment, and [t-Ts(t)] represents the adjusted time at the corresponding moment; It should be noted that the formula in S42 is an important calculation process for achieving heat control response compensation; in the formula, Dth(t) is the heat output power path function after adjustment, representing the heating control curve after time adjustment; Dhc[*] is the heat output power path function after adjustment, reflecting the heating intensity requirement preset in the recipe; and Ts(t) represents the thermal response lag time difference detected at the t-th moment, that is, the time offset by which the actual pot body temperature response is later than the theoretical expectation; by shifting Dhc(t) backward along the time axis by Ts(t), that is, constructing a time mapping of t-Ts(t), for each The control signal at each moment is time-compensated, forming a post-adjustment output path that better matches the actual heating response. This mechanism directly addresses the pain points of heating control response lag and recipe model adaptive adjustment in existing technologies, implementing dynamic correction logic based on response deviation. Its logical significance lies in the fact that real heating systems often have control lag caused by the thermal inertia of the pot and environmental fluctuations. This formula allows the measured response time difference to be used to adjust the original recipe heating plan point by point, thereby improving the accuracy of heat restoration and the consistency of thermal control. The post-adjustment heat output power path function serves as the core carrier of closed-loop correction. S43. After encoding the adjusted heat output power path function of the current original recipe through a pulse width modulation control mechanism, the adjusted heat power input signal sequence of the current original recipe is generated by the path reconstruction vector, and combined with the operational semantics reverse parsing model, a standard heat recommended recipe text is generated, updated to the recipe database in the cooking robot, and the current original recipe corresponding to the adjusted heat output power path function is marked as a standard heat recommended recipe.

[0042] It should be noted that in the adaptive recipe update process of the present invention, the adjusted thermal output power path function, obtained through time response control, is first used as input and encoded using a pulse width modulation control mechanism. This converts the continuous thermal power variation curve into a set of discrete control pulse sequences with a corresponding relationship between time width and power amplitude. This encoded sequence is then input into the path reconstruction module and converted into an adjusted thermal power input signal sequence through a reconstruction vector mechanism, which is used to actually drive the robot's heating control unit to perform the optimized heating task. After signal generation is completed, the operational semantics reverse parsing model is further invoked to map this set of thermal power signal sequences with clear temperature control logic back into a natural language representation, reconstructing the new recipe text content. This text not only retains the logical structure and semantic style of the original recipe, but also incorporates the optimized temperature control rhythm and operation phase characteristics, thereby generating a standard heat recommendation recipe that has undergone heat verification and execution correction. Ultimately, it is automatically updated to the cooking robot's recipe database for priority call and demonstration execution during subsequent dish processing. This process achieves a complete closed loop from execution behavior correction to recipe language expression optimization, achieving both control feasibility and semantic readability.

[0043] In this embodiment, by introducing the thermal response lag identification and path function time control mechanism, an adaptive correction path of the cooking robot's recipe execution logic is constructed, which substantially realizes the closed-loop self-optimization and knowledge feedback function of the recipe model; first, after receiving the fire response control instruction, the theoretical thermal output power path function of the original recipe and the thermal output power path function inferred in the actual execution are compared moment by moment, and with the help of the sliding window difference method, the time lag difference between the thermal power rising points in each stage is accurately identified, thereby obtaining a thermal response time difference sequence reflecting the lag of the robot's heating reaction; the lag identification mechanism can dynamically capture the control response delay caused by the thermal inertia of the pot body, equipment aging or environmental fluctuations in the actual work of the execution port, filling the gap in the traditional recipe static model's insufficient adaptability to the real response timing; further, based on the above-mentioned response lag data, the theoretical thermal output path function is locally translated and fine-tuned in the time dimension to construct the adjusted thermal output power path function. The output power path function ensures that the new path model is more in line with the actual heating response rhythm; after the adjusted function is encoded by the pulse width modulation mechanism, a new thermal power control signal sequence is generated to drive the robot to execute the corrected heating process; more importantly, on this basis, the operational semantics reverse parsing model is called to convert the new thermal path function into a structured standard heat recommendation recipe text, and automatically update it to the recipe database, marking it as a high-precision recipe template verified by actual heat; this move not only realizes the dynamic evolution and knowledge accumulation of recipe content, but also enables the robot to have a closed-loop intelligent control process of "learning-execution-correction-update"; the core of this step is to break through the non-adaptive barriers between traditional recipe design and execution, so that the cooking robot truly has the self-correction capability for timing errors, and feeds back the micro-control deviation to the recipe model layer, forming a cross-level linkage correction mechanism, improving the recipe self-healing ability and the practicality and intelligence of the recipe management mechanism.

[0044] Example 6

[0045] Please refer to Figure 1 and Figure 2 ,Specifically: A cooking robot recipe intelligent management system, including a recipe parsing module, a path analysis module, a heat analysis module and a recipe management module; The recipe parsing module is used to parse the operation phase sequence parameters of the current recipe based on the current original recipe text in the recipe database using a pre-trained operational semantic parsing model. It then determines the theoretical heat output power value for each operation phase by combining the heat conductivity of the cookware and the heat capacity of the ingredients. Based on the time boundary matching mechanism, it determines the theoretical heat output power time function of the current original recipe to execute the heating operation. The path analysis module is used to determine the actual heat output power path function during the heating process based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; The heat analysis module is used to compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function, determine whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe, and issue a corresponding heat response control instruction; The recipe management module is used to perform time response control on the theoretical heat output power path function of the current original recipe after receiving the heat response control instruction, and update it to the recipe database in the cooking robot.

[0046] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A cooking robot recipe intelligent management method, characterized by: The following steps are involved: S1. Based on the current original recipe text in the recipe database, a pre-trained operational semantic parsing model is used to parse the operation phase sequence parameters of the current recipe. The theoretical heat output power value for each operation phase is determined by combining the heat conductivity of the cookware and the heat capacity of the ingredients. The theoretical heat output power time function of the current original recipe is then determined based on a time boundary matching mechanism to execute the heating operation. S2. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; S3. Compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function to determine whether there is a risk of temperature deviation in the heating operation performed by the cooking robot for the current original recipe, and issue a corresponding temperature response control instruction; S4. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe is time-responsively controlled and updated to the recipe database in the cooking robot.

2. The cooking robot recipe intelligent management method according to claim 1, characterized in that: The specific steps of S1 include: S11. Based on the current original recipe text in the recipe database, and using the pre-trained operational semantic parsing model, parse the operation phase sequence parameters of the current recipe. The specific parsing method is as follows: The cooking robot uses the current original recipe text stored in the recipe database as the initial input. The current original recipe text includes the behavioral terms of each operation stage, the time nodes of the ingredients, and the description of the heat; By constructing a ternary mapping relationship of "action verbs, control targets, and heat parameters" based on various standard recipe texts stored in a recipe database, combined with contextual semantic modeling tasks, and using a multi-layer encoding network to perform semantic embedding learning and control parameter alignment training for each operation stage in the recipe text, a pre-trained operational semantic parsing model is obtained. Based on the current original recipe text and using the pre-trained operational semantic parsing model, the stage number, action category, target temperature and temperature maintenance duration of each operation stage in the current original recipe are identified and labeled, and the operation stage sequence parameters of the current original recipe are constructed.

3. The cooking robot recipe intelligent management method according to claim 2, characterized in that: S12. Feature extraction is performed on the operation phase sequence parameters of the current original recipe, and the theoretical heat output power value of each operation phase is determined by combining the heat conductivity characteristics of the cookware and the heat capacity characteristics of the ingredients. Specifically, Where Pgv i Indicates the theoretical thermal output power value of the corresponding operation stage, St i Indicates the temperature maintenance time of the corresponding operation stage, Tmb i Indicates the target temperature of the corresponding operation stage, Crr indicates the heat capacity of the ingredients in the current original recipe, Thc indicates the current ambient temperature, and Ncd indicates the heat conduction efficiency of the cookware.

4. The cooking robot recipe intelligent management method according to claim 3, characterized in that: S13. Based on the theoretical heat output power values ​​of each operation stage determined in S12 and in combination with the time boundary matching mechanism, the theoretical heat output power time function of the current original recipe is determined to perform the heating operation. The specific process is as follows: By identifying the characteristics of the operation phase sequence parameters of the current original recipe, the temperature maintenance start time and temperature maintenance end time of each operation phase are extracted to obtain the effective temperature control time interval of each operation phase, and the time boundary indicator function is set accordingly; According to the set time boundary indicator function and combined with the theoretical thermal output power value of each operation stage, the theoretical thermal output power path function is constructed, specifically: Where Dhc(t) represents the theoretical thermal output power path function, Pgv i represents the theoretical thermal output power value of the i-th operating stage, represents the time boundary indicator function; After the constructed theoretical thermal output power path function is encoded through the pulse width modulation control mechanism, the thermal power input signal sequence of the current original recipe is generated by the path reconstruction vector and transmitted to the control port of the heating device through the drive interface to perform the heating operation.

5. The cooking robot recipe intelligent management method according to claim 4, characterized in that: The specific steps of S2 include: S21. During the heating process, the actual heat output power path function is determined based on the actual temperature change data collected from the bottom of the cooking robot pot and the heat conduction principle inverse model. The specific process is as follows: During the heating process, the infrared temperature sensor deployed at the bottom of the cooking robot's pot collects the actual temperature change data sequence in real time and constructs the pot body temperature response function; Based on the heat capacity characteristics of the ingredients in the current original recipe, combined with the constructed pot temperature response curve function, and using the heat conduction principle to infer the model, the actual heat output power path function is determined, specifically: Where Dsj(t) represents the actual heat output power path function, Tsj(t) represents the pot temperature response function, represents the rate of temperature rise of the pot at time t, Crr represents the heat capacity of the ingredients in the current original recipe, and Ncd represents the heat conduction efficiency of the pot.

6. The cooking robot recipe intelligent management method according to claim 5, characterized in that: The S3 specific steps include: S31. Compare and analyze the theoretical thermal output power path function of the current original recipe with the corresponding actual thermal output power path function. After dimensionless processing, determine the degree of deviation of the real-time thermal output power of the heating operation performed by the cooking robot for the current original recipe at the corresponding moment, specifically: Tpc(t)=Dhc(t)-Dsj(t); where Tpc(t) represents the thermal output power deviation value at moment t, Dhc(t) represents the theoretical thermal output power path function, and Dsj(t) represents the actual thermal output power path function.

7. The cooking robot recipe intelligent management method according to claim 6, characterized in that: S32. Determine whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe based on the degree of real-time power deviation of the heating operation. The specific analysis process is as follows: If the heat output power at the corresponding moment is equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is normal, and the corresponding moment is marked as a normal heating operation moment. If the heat output power deviation value at the corresponding moment is not equal to zero, it means that the cooking robot's heating operation for the current original recipe at the corresponding moment is abnormal, and the corresponding moment is marked as an abnormal heating operation moment. The number of moments marked as abnormal heating operations is counted to obtain the number of abnormal heating operation moments. When the number of abnormal heating operation moments exceeds 5% of the total number of moments within the heating time, it indicates that there is a risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, a temperature response control instruction is issued. Otherwise, it indicates that there is no risk of temperature deviation in the heating operation performed by the cooking robot on the current original recipe. At this time, the current original recipe stored in the recipe database of the cooking robot is marked as a standard temperature recommended recipe.

8. The cooking robot recipe intelligent management method according to claim 7, characterized in that: The specific steps of S4 include: S41. After receiving the heat response control instruction, the theoretical heat output power path function of the current original recipe arranged in time series is compared with the corresponding actual heat output power path function on the time axis, and combined with the sliding window difference method, the time difference between the moment when the theoretical heat output power rises and the corresponding moment when the actual heat output power rises is identified to obtain the thermal response lag time difference at each moment.

9. The cooking robot recipe intelligent management method according to claim 8, characterized in that: S42. Based on the thermal response lag time difference at each moment in S41, the theoretical thermal output power path function of the current original recipe is time-controlled to construct the adjusted thermal output power path function of the current original recipe, specifically: Dth(t)=Dhc[t-Ts(t)]; where Dth(t) represents the adjusted thermal output power path function, Ts(t) represents the thermal response lag time difference at the corresponding moment, and [t-Ts(t)] represents the adjusted time at the corresponding moment; S43. After encoding the adjusted heat output power path function of the current original recipe through a pulse width modulation control mechanism, the adjusted heat power input signal sequence of the current original recipe is generated by the path reconstruction vector, and combined with the operational semantics reverse parsing model, a standard heat recommended recipe text is generated, updated to the recipe database in the cooking robot, and the current original recipe corresponding to the adjusted heat output power path function is marked as a standard heat recommended recipe.

10. An intelligent cooking robot recipe management system, used to implement the cooking robot recipe management method according to any one of claims 1 to 9, characterized in that: Including recipe parsing module, path analysis module, heat analysis module and recipe management module; The recipe parsing module is used to parse the operation phase sequence parameters of the current recipe based on the current original recipe text in the recipe database using a pre-trained operational semantic parsing model. It then determines the theoretical heat output power value for each operation phase by combining the heat conductivity of the cookware and the heat capacity of the ingredients. Based on the time boundary matching mechanism, it determines the theoretical heat output power time function of the current original recipe to execute the heating operation. The path analysis module is used to determine the actual heat output power path function during the heating process based on the actual temperature change data collected from the bottom of the cooking robot pot and the inverse model based on the heat conduction principle; The heat analysis module is used to compare and analyze the theoretical heat output power path function of the current original recipe with the corresponding actual heat output power path function, determine whether there is a risk of heat deviation in the heating operation performed by the cooking robot on the current original recipe, and issue a corresponding heat response control instruction; The recipe management module is used to perform time response control on the theoretical heat output power path function of the current original recipe after receiving the heat response control instruction, and update it to the recipe database in the cooking robot.

Images