Intelligent frying and roasting equipment adaptive temperature control method and system based on multi-modal sensor

Abstract

The present application relates to the technical field of intelligent household appliance control, in particular to a self-adaptive temperature control method and system for intelligent frying and roasting equipment based on multi-modal sensors; comprising multi-source perception, feature calculation, confidence assessment, model deduction and fusion control module; the system collects infrared, image and angle data, calculates temperature fluctuation and angle change rate; the core is to generate sensor confidence factor by using confidence decay model, and to calculate and deduce temperature in combination with heat balance equation; based on confidence, the weight of measured value and deduced value is dynamically adjusted to generate fusion temperature for closed-loop control; the present application effectively identifies oil fume shielding and cover opening interference, solves the problem of perception distortion, and realizes accurate temperature control in complex environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent home appliance control technology, specifically to an adaptive temperature control method and system for intelligent grilling equipment based on multimodal sensors. Background Technology

[0002] In practical cooking applications of intelligent grilling equipment, the system typically relies on infrared and contact sensors to collect temperature data in order to maintain a preset constant temperature control state. However, the later stages of cooking are often accompanied by the generation of high concentrations of oil fumes and frequent opening of the lid by the user, creating a complex and dynamically changing thermal environment. Existing temperature control solutions generally adopt a closed-loop feedback mechanism that directly accepts sensor readings, lacking the ability to evaluate and compensate for the reliability of the observed data in real time. When oil fumes obstruct the sensor window or the environment changes abruptly due to opening the lid, the infrared sensor readings will experience non-physical drastic jumps or perception distortion, usually manifesting as false low temperature alarms. Such data distortion causes the control algorithm to misjudge the current thermal state, thereby incorrectly driving the heating element to operate at high power continuously, resulting in food burning, energy waste, and failure of the overall temperature control logic.

[0003] Therefore, how to effectively identify sensor distortion and maintain the accuracy and robustness of temperature control under complex working conditions such as oil fume obstruction and interference from opening the lid has become an urgent technical problem to be solved. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an adaptive temperature control method and system for intelligent grilling equipment based on multimodal sensors. Specifically, the technical solution of this invention is as follows:

[0005] An adaptive temperature control method for intelligent grilling equipment based on multimodal sensors includes:

[0006] Step 1: The instantaneous temperature of the plate is collected by an infrared sensor, the average brightness of the food surface image is collected by a wide-angle camera, the real-time input power of the heating element is collected by a power detection circuit, the opening and closing angle of the cover is collected by an angle sensor, and the ambient temperature is collected by an NTC sensor at the cold end of the machine body.

[0007] Step 2: Calculate the temperature fluctuation variance within a preset sliding window based on the instantaneous temperature of the disk surface;

[0008] Step 3: Calculate the rate of change of the cover plate angle based on the opening and closing angle of the cover plate;

[0009] Step 4: Combining the temperature fluctuation variance and the cover plate angle change rate, the confidence factor of the sensor is generated using the confidence decay model;

[0010] Step 5: Based on the real-time input power, ambient temperature, preset thermodynamic parameters, and the model-induced temperature of the previous moment, calculate the model-induced temperature of the current moment using the discretized lumped parameter heat balance equation.

[0011] Step 6: Based on the sensor confidence factor, determine the first weight of the instantaneous temperature of the disk surface and the second weight of the model-inferred temperature at the current moment;

[0012] Step 7: Combine the first weight, the instantaneous temperature of the trading surface, the second weight, and the model-deduced temperature at the current moment, and generate the fused feedback temperature through weighted summation calculation;

[0013] Step 8: Based on the fused feedback temperature and the preset target temperature, a power control signal for adjusting the heating element is generated through a closed-loop feedback control algorithm.

[0014] Preferably, the confidence factor of the sensor is generated using a confidence decay model, including:

[0015] The temperature fluctuation variance is multiplied by a preset noise sensitivity coefficient to generate a temperature noise term;

[0016] The absolute value of the rate of change of the cover angle is multiplied by the preset cover opening interference penalty coefficient to generate the cover opening interference term;

[0017] The change in the average image brightness is multiplied by a preset oil fume sensitivity coefficient to generate an oil fume occlusion term.

[0018] Add the temperature noise term, the lid opening interference term, and the oil fume blocking term to the constant 1 to generate the denominator term;

[0019] Calculate the quotient of the numerical value 1 divided by the denominator, and determine the quotient as the sensor confidence factor.

[0020] Preferably, the model-estimated temperature at the current moment is calculated using the discretized lumped-parameter heat balance equation, including:

[0021] Calculate the product of real-time input power and preset electrothermal conversion efficiency to generate energy input items;

[0022] Calculate the difference between the model-induced temperature and the ambient temperature at the previous moment;

[0023] The heat loss term is calculated based on the difference, the preset comprehensive convective heat transfer coefficient, and the preset effective heat dissipation area.

[0024] Determine the heat absorption power of water evaporation at the current moment;

[0025] Calculate the net energy increment of the system based on energy input, heat loss, and heat absorption power from water evaporation.

[0026] Multiply the net energy increment of the system by the preset sampling period and divide by the preset equivalent heat capacity of the system to generate the temperature rise increment;

[0027] The model-predicted temperature from the previous moment is added to the temperature rise increment to generate the model-predicted temperature for the current moment.

[0028] Preferably, determining the heat absorption power of water evaporation at the current moment includes:

[0029] Obtain the cumulative running time since the heating operation started;

[0030] Calculate the product of the preset attenuation coefficient and the cumulative running time to generate the attenuation index;

[0031] The heat absorption power of water evaporation is generated by calculating the product of the preset maximum evaporation power and the negative exponent of the natural constant.

[0032] Preferably, determining the first weight of the instantaneous temperature of the disk surface and the second weight of the model-derived temperature at the current moment includes:

[0033] The sensor confidence factor was directly determined as the first weight.

[0034] Calculate the difference between value 1 and the sensor confidence factor, and determine the difference as the second weight.

[0035] Preferably, a power control signal for adjusting the heating element is generated through a closed-loop feedback control algorithm, including:

[0036] Calculate the difference between the preset target temperature and the fusion feedback temperature to generate the current temperature deviation;

[0037] Based on the current temperature deviation, the temperature deviation at the previous moment, and the temperature deviation at the moment before that, combined with the preset proportional coefficient, integral coefficient, and derivative coefficient, the incremental PID algorithm is used to calculate the power adjustment increment.

[0038] The power control signal from the previous moment is added to the power adjustment increment to generate the power control signal for the current moment.

[0039] Preferably, it also includes an initialization step:

[0040] In response to device startup, detect the initial temperature read by the infrared sensor;

[0041] Set the initial value of the model-induced temperature from the previous moment as the initial temperature.

[0042] The intelligent grilling equipment adaptive temperature control system based on multimodal sensors includes:

[0043] The multi-dimensional acquisition module is used to acquire the instantaneous temperature of the plate surface through an infrared sensor, the average brightness of the food surface image through a wide-angle camera, the real-time input power of the heating element through a power detection circuit, the opening and closing angle of the cover through an angle sensor, and the ambient temperature through an NTC sensor at the cold end of the machine body.

[0044] The confidence calculation module is used to combine the temperature fluctuation variance calculated based on the instantaneous temperature of the disc surface and the cover angle change rate calculated based on the cover opening and closing angle, and generate the sensor confidence factor using the confidence decay model.

[0045] The thermodynamic deduction module is used to calculate the model deduction temperature at the current moment based on real-time input power, ambient temperature, preset thermodynamic parameters, and the model deduction temperature at the previous moment, using the discretized lumped parameter heat balance equation.

[0046] The data fusion module is used to generate a fused feedback temperature by weighted summation of the instantaneous temperature of the disk surface and the model-inferred temperature at the current moment, based on the sensor confidence factor.

[0047] The closed-loop control module is used to generate a power control signal for adjusting the heating element based on the fused feedback temperature and the preset target temperature.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. This invention integrates multimodal data from infrared sensors, wide-angle cameras, and angle sensors to construct a multidimensional feature evaluation system that includes temperature fluctuation variance, lid angle change rate, and image brightness change. Utilizing a confidence decay model, the system can quantitatively assess the reliability of current observation data in real time, accurately identifying non-physical data jumps caused by oil fumes obscuring the viewing window or user opening the lid during the later stages of cooking. This effectively avoids misjudgments caused by blindly accepting low sensor readings during sudden environmental changes in traditional temperature control solutions, solving the perception distortion problem mentioned in the background technology.

[0050] 2. This invention introduces a discretized lumped parameter thermal balance equation, which uses real-time input power, ambient temperature, and thermodynamic parameters to independently deduce the theoretical thermal state of the equipment without relying on external optical sensing. When the infrared sensor fails completely due to severe oil fume pollution or when the reading fluctuates drastically due to heat dissipation from opening the cover, the system can output a logically sound temperature estimate based on the physical laws of energy input and heat loss. This soft sensing mechanism provides a reliable fallback reference for the system, ensuring that the temperature control logic remains online and reasonable during sensor failure.

[0051] 3. This invention establishes a dual-channel weight allocation mechanism based on sensor confidence. Under ideal operating conditions without interference, the system assigns high weight to the measured values ​​to ensure rapid response and sensitivity of temperature control. Under conditions of oil fumes or open lid interference, the system automatically reduces the weight of the measured values ​​and increases the weight of the model-inferred values, smoothly transitioning to a control strategy based on a physical model. This linear interpolation transition not only avoids system oscillations that may be caused by hard switching of control weights, but also effectively prevents continuous high-power heating induced by false low-temperature alarms from sensors, thereby eliminating food burning and energy waste.

[0052] 4. In the thermodynamic deduction, this invention specifically considers the dynamic convective heat transfer coefficient that changes with the angle of the cover plate, and introduces a water evaporation heat absorption power model that decays over time. By simulating the physical fact that the moisture in the food gradually decreases during cooking, the system can dynamically compensate for the heat loss caused by the latent heat of vaporization, and significantly correct the temperature deduction error in the later stages of cooking. Combined with an incremental closed-loop feedback control algorithm, the system exhibits stronger anti-saturation capability and stability when dealing with data source switching and nonlinear time-varying loads. Attached Figure Description

[0053] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0054] Figure 1 This is a flowchart of the method of the present invention;

[0055] Figure 2 This is a structural diagram of the system of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0057] Example 1:

[0058] Please see Figure 1 An adaptive temperature control method for intelligent grilling equipment based on multimodal sensors includes:

[0059] Step 1: The instantaneous temperature of the plate is collected by an infrared sensor, the average brightness of the food surface image is collected by a wide-angle camera, the real-time input power of the heating element is collected by a power detection circuit, the opening and closing angle of the cover is collected by an angle sensor, and the ambient temperature is collected by an NTC sensor at the cold end of the machine body.

[0060] Step 2: Calculate the temperature fluctuation variance within a preset sliding window based on the instantaneous temperature of the disk surface;

[0061] Step 3: Calculate the rate of change of the cover plate angle based on the opening and closing angle of the cover plate;

[0062] Step 4: Combining the temperature fluctuation variance and the cover plate angle change rate, the confidence factor of the sensor is generated using the confidence decay model;

[0063] Step 5: Based on the real-time input power, ambient temperature, preset thermodynamic parameters, and the model-induced temperature of the previous moment, calculate the model-induced temperature of the current moment using the discretized lumped parameter heat balance equation.

[0064] Step 6: Based on the sensor confidence factor, determine the first weight of the instantaneous temperature of the disk surface and the second weight of the model-inferred temperature at the current moment;

[0065] Step 7: Combine the first weight, the instantaneous temperature of the trading surface, the second weight, and the model-deduced temperature at the current moment, and generate the fused feedback temperature through weighted summation calculation;

[0066] Step 8: Based on the fusion feedback temperature and the preset target temperature, a power control signal for adjusting the heating element is generated through a closed-loop feedback control algorithm.

[0067] This embodiment provides an adaptive temperature control method for intelligent grilling equipment based on multimodal sensors, which aims to solve the problem that existing grilling equipment suffers from temperature perception distortion due to oil fumes obscuring the sensors or users opening the lid during the later stages of cooking, thus leading to control failure.

[0068] The system constructs a state space and acquires four key parameters characterizing the cooking environment through multi-source heterogeneous data acquisition; the system uses an infrared sensor to collect the instantaneous temperature of the plate surface in a non-contact manner. The image information of the food surface is captured by a wide-angle camera, and the average brightness of the image is extracted by image processing algorithms. As an auxiliary feature, the real-time input power of the heating element is monitored in real time using a power detection circuit. As the energy input variable of the system; the opening and closing angle of the cover plate is collected using an angle sensor. The ambient temperature is collected using an NTC sensor located at the cold end of the fuselage. The data collection frequency for the above data is set to... ;

[0069] The system is based on the instantaneous temperature of the disk surface. Select a length of Using a sliding window, calculate the variance of temperature fluctuations within the window. This indicator quantifies the dispersion of the signal over a short period of time; when oil fume obstruction occurs, the non-physical, drastic jumps in the infrared readings will significantly increase this variance; the system collects the opening and closing angle of the cover plate. By performing differentiation, the rate of change of the cover plate angle is obtained. , used to capture sudden thermal environment events;

[0070] Combining the temperature fluctuation variance and the cover plate angle change rate, the system uses a confidence decay model to generate a sensor confidence factor. This factor is a dimensionless normalized value ranging from 0 to 1, which quantitatively describes the reliability of the current infrared sensor readings.

[0071] While acquiring sensor data, the system runs a thermodynamic inverse problem solver in parallel; based on the principle of energy conservation, it utilizes real-time input power... and ambient temperature Combining preset thermodynamic parameters and the model-induced temperature from the previous moment... By using the discretized lumped parameter heat balance equation, the model-induced temperature at the current moment can be calculated. This process does not rely on current sensor readings, but only on energy integration based on physical laws.

[0072] The system is based on the sensor confidence factor ,Will The instantaneous temperature of the disk surface is determined as the first weight, representing the degree of confidence in the measured data; the value 1 is compared with... The difference is determined as the second weight of the model-predicted temperature at the current moment, representing the degree of dependence on the physical model's predicted value; a weighted summation algorithm is used to calculate the instantaneous temperature of the disk surface. Temperature in Model Derivation Perform fusion to generate fusion feedback temperature ;

[0073] Fusion feedback temperature As a feedback variable of the system, it is related to the user-preset target temperature. The method involves comparing the results; using a closed-loop feedback control algorithm, the adjustment amount is calculated based on the deviation, and a control signal is generated to adjust the power of the heating element; this method constructs a dual-channel mechanism of sensor observation and model inference, which automatically identifies low-confidence states and smoothly transitions control to a physical model based on energy conservation when oil fumes cause sensor reading distortion, thus avoiding continuous erroneous heating caused by sensor false alarms of low temperature.

[0074] Example 2:

[0075] Using a confidence decay model, a sensor confidence factor is generated, including:

[0076] The temperature fluctuation variance is multiplied by a preset noise sensitivity coefficient to generate a temperature noise term;

[0077] The absolute value of the rate of change of the cover angle is multiplied by the preset cover opening interference penalty coefficient to generate the cover opening interference term;

[0078] The change in the average image brightness is multiplied by a preset oil fume sensitivity coefficient to generate an oil fume occlusion term.

[0079] Add the temperature noise term, the lid opening interference term, and the oil fume blocking term to the constant 1 to generate the denominator term;

[0080] Calculate the quotient of the numerical value 1 divided by the denominator, and determine the quotient as the sensor confidence factor.

[0081] This embodiment is a concretization of the confidence decay model;

[0082] Sensor confidence factor The calculation employs an inverse proportional decay algorithm based on a variant of the logistic function, as shown in the following formula:

[0083] ;

[0084] in, The temperature noise term is derived from the temperature fluctuation variance. With the preset noise sensitivity coefficient Multiplying yields the result; to ensure dimensional consistency, the unit of k1 is set to [value missing]. ; The values ​​were obtained through laboratory calibration. The calibration process is as follows: a calibration dataset containing different concentrations of oil fume shielding conditions was constructed, and the peak variance of the temperature measured by the infrared sensor in this dataset was statistically analyzed and denoted as . ;set up for The reciprocal of, that is ;

[0085] The opening interference term is determined by the absolute value of the rate of change of the cover angle and the preset opening interference penalty coefficient. Multiplying them together yields the result; The unit is set as This mathematical model, by introducing coefficients with consistent physical dimensions, maps interference signals of different dimensions uniformly to a dimensionless attenuation function, ensuring that the confidence factor remains constant as interference increases. It exhibits a non-linear, rapid descent characteristic;

[0086] For the oil fume masking factor, it is determined by the change in the average image brightness and the preset oil fume sensitivity coefficient. Multiplying these values ​​yields the following: when the concentration of cooking fumes increases, causing an abnormal decrease in image brightness, this value increases, thereby reducing the confidence factor. Specifically, the change in the mean image brightness. Defined as the average initial image brightness when the device starts up. Image brightness average at current time The difference is calculated using the following formula: In this embodiment, the interior of the grilling equipment is set to a dark, light-absorbing background. When oil fumes are generated, the average image brightness decreases due to light scattering. Therefore, the oil fume concentration is characterized by calculating the decrease in brightness. Furthermore... The values ​​were obtained through laboratory calibration: a standard concentration of oil fume was constructed under occupancy conditions, and the maximum decrease in the mean image brightness relative to the initial value was measured under these conditions. ,set up The reciprocal of the decrease, i.e. This is to ensure that the term is normalized under standard interference.

[0087] Example 3:

[0088] Using the discretized lumped-parameter heat balance equation, the model-estimated temperature at the current moment is calculated, including:

[0089] Calculate the product of real-time input power and preset electrothermal conversion efficiency to generate energy input items;

[0090] Calculate the difference between the model-induced temperature and the ambient temperature at the previous moment;

[0091] The heat loss term is calculated based on the difference, the preset comprehensive convective heat transfer coefficient, and the preset effective heat dissipation area.

[0092] Determine the heat absorption power of water evaporation at the current moment;

[0093] Calculate the net energy increment of the system based on energy input, heat loss, and heat absorption power from water evaporation.

[0094] Multiply the net energy increment of the system by the preset sampling period and divide by the preset equivalent heat capacity of the system to generate the temperature rise increment;

[0095] The model-predicted temperature from the previous moment is added to the temperature rise increment to generate the model-predicted temperature for the current moment.

[0096] This embodiment is a concretization of the discretized lumped parameter heat balance equation;

[0097] To extrapolate temperature in real time using a digital microcontroller, a discrete equation is constructed using the backward Euler method; the model extrapolates the temperature at the current moment. The calculation formula is as follows:

[0098] ;

[0099] Among them, the calculation of real-time input power With the preset electrothermal conversion efficiency The product of these terms generates the energy input term; This is a dimensionless constant derived from the heating component specifications; it is used to calculate the model-derived temperature at the previous moment. With the current ambient temperature The difference; based on this difference, multiply by the comprehensive convective heat transfer coefficient corresponding to the current cover angle. and the preset effective heat dissipation area This generates a heat dissipation loss term; It depends on the opening and closing angle of the cover plate. The changing dynamic parameters are calculated using the following formula:

[0100] ;

[0101] in, The heat transfer coefficient is the closed-cover coefficient. The heat transfer coefficient with the lid fully open is an empirical constant obtained by fitting through natural cooling experiments. The unit is These are geometric parameters calculated based on mechanical design drawings; the heat power lost due to moisture evaporation from the food at the current moment is determined. ;

[0102] Subtracting the heat loss term from the energy input term, and then subtracting the heat absorption power from water evaporation, yields the net energy increment of the system; multiplying the net energy increment of the system by the preset sampling period. And divided by the preset system equivalent heat capacity This generates an increase in temperature. The unit is This is a fixed constant obtained through the equivalent heat capacity test of the whole machine; the temperature at the previous moment... Adding the temperature increase, we obtain the theoretical temperature at the current moment. Even when the sensor fails completely, the model can output a temperature estimate that conforms to physical logic based on the input electrical energy and the inherent thermal characteristics of the system.

[0103] Example 4:

[0104] Determining the heat absorption power of water evaporation at the current moment includes:

[0105] Obtain the cumulative running time since the heating operation started;

[0106] Calculate the product of the preset attenuation coefficient and the cumulative running time to generate the attenuation index;

[0107] The heat absorption power of water evaporation is generated by calculating the product of the preset maximum evaporation power and the negative exponent of the natural constant.

[0108] This embodiment is a specific implementation of the steps for determining the heat absorption power of water evaporation.

[0109] Considering that the moisture content of ingredients decreases over time during cooking, and the heat absorption effect of evaporation decreases, this embodiment constructs an evaporation model that varies with time:

[0110] ;

[0111] The moment when the detection fusion feedback temperature first reaches the preset boiling point temperature is recorded as . ; Get the current time relative to Effective evaporation duration If the current temperature has not reached the boiling point, the continuous evaporation process is considered interrupted, and the effective evaporation duration is paused. The cumulative and When the temperature reaches the boiling point again, The accumulated values ​​will continue to accumulate based on the current values; the preset attenuation coefficient will be calculated. With effective evaporation duration The product of; in terms of the natural constant Using the base as the base and the negative of the above product as the exponent, calculate the attenuation factor; then, compare the attenuation factor with the preset maximum evaporation power. Multiply to get the current time value. ;

[0112] and The method for determining the value is as follows: Construct a weightlessness test experiment under standard food load and record the values ​​at different times. Food weight loss rate data Using the least squares method to transform the function right Perform fitting, where The latent heat of vaporization of water is used to determine the fitting parameters. and The value of; among which The unit is set as This model is used to characterize the rate at which the water evaporation rate decays over time; it dynamically compensates for the energy loss caused by water evaporation, improving the estimation accuracy of the thermodynamic model throughout the entire cooking cycle.

[0113] Example 5:

[0114] The first weight for determining the instantaneous temperature of the trading surface and the second weight for the model-derived temperature at the current moment are as follows:

[0115] The sensor confidence factor was directly determined as the first weight.

[0116] Calculate the difference between value 1 and the sensor confidence factor, and determine the difference as the second weight.

[0117] This embodiment is a specific implementation of the weight determination method;

[0118] A complementary weighting strategy is adopted: the sensor confidence factor is directly assigned. Determined as the first weight; calculate the value 1 and the sensor confidence factor. The difference is determined as the second weight; this strategy is based on Bayesian estimation theory, when the sensor confidence level... When the confidence level approaches 1, the system mainly relies on measured values ​​to ensure control sensitivity; when the sensor confidence level... When the value approaches 0, the system automatically tilts towards the model value to ensure control safety; this linear interpolation transition avoids control oscillations that may be caused by hard switching.

[0119] Example 6:

[0120] A closed-loop feedback control algorithm is used to generate a power control signal for adjusting the heating element, including:

[0121] Calculate the difference between the preset target temperature and the fusion feedback temperature to generate the current temperature deviation;

[0122] Based on the current temperature deviation, the temperature deviation at the previous moment, and the temperature deviation at the moment before that, combined with the preset proportional coefficient, integral coefficient, and derivative coefficient, the incremental PID algorithm is used to calculate the power adjustment increment.

[0123] The power control signal from the previous moment is added to the power adjustment increment to generate the power control signal for the current moment.

[0124] This embodiment is a concretization of the closed-loop feedback control algorithm, which adopts an incremental PID algorithm based on discrete-time sampling to adapt to the processing characteristics of digital microcontrollers.

[0125] Calculate the preset target temperature With current fusion feedback temperature The difference is used to generate the current temperature deviation. The power change that needs to be adjusted at the current moment is calculated using the following discretized formula that incorporates the sampling time parameter. :

[0126] ;

[0127] in, The sampling period corresponds to the system's data acquisition frequency, in seconds; and These are the temperature deviations at the previous moment and the two moments before that, respectively.

[0128] This is the proportionality coefficient, in units of... , used to indicate the magnitude of the response deviation;

[0129] The integral coefficient is expressed in units of 1000 ppm. Multiply by in the formula This is to transform the integral action in the continuous domain into a cumulative increment in the discrete domain, ensuring that the calculation result retains the power dimension. ;

[0130] These are the differential coefficients, in units of... Dividing by in the formula This is to approximate the second-order difference of temperature change as a rate of change, ensuring that the calculation result retains the dimensions of power. ;

[0131] The power control signal from the previous moment Add the calculated increment Generate the power control signal for the current moment. The algorithm strictly follows the law of conservation of physical dimensions, and can effectively avoid the system impact caused by integral saturation through incremental calculation when the fusion weight changes due to the switching between sensor data and model data.

[0132] Example 7:

[0133] It also includes an initialization step:

[0134] In response to device startup, detect the initial temperature read by the infrared sensor;

[0135] Set the initial value of the model-induced temperature from the previous moment as the initial temperature.

[0136] This embodiment is a supplement to the adaptive temperature control method, adding an initialization step during system startup;

[0137] In response to device startup, the system detects the initial temperature read by the infrared sensor. The system assumes that at the moment of startup... The equipment is not yet severely contaminated by oil and is in a state of relative thermal equilibrium. The initial temperature value extrapolated from the model at the previous moment will be used. Set to equal to the initial temperature This strategy anchors the starting point of the physical model to real physical measurements, eliminating the initial bias of the model's accumulated error.

[0138] Example 8:

[0139] Please see Figure 2 An adaptive temperature control system for intelligent grilling equipment based on multimodal sensors includes:

[0140] The multi-dimensional acquisition module is used to acquire the instantaneous temperature of the plate surface through an infrared sensor, the average brightness of the food surface image through a wide-angle camera, the real-time input power of the heating element through a power detection circuit, the opening and closing angle of the cover through an angle sensor, and the ambient temperature through an NTC sensor at the cold end of the machine body.

[0141] The confidence calculation module is used to combine the temperature fluctuation variance calculated based on the instantaneous temperature of the disc surface and the cover angle change rate calculated based on the cover opening and closing angle, and generate the sensor confidence factor using the confidence decay model.

[0142] The thermodynamic deduction module is used to calculate the model deduction temperature at the current moment based on real-time input power, ambient temperature, preset thermodynamic parameters, and the model deduction temperature at the previous moment, using the discretized lumped parameter heat balance equation.

[0143] The data fusion module is used to generate a fused feedback temperature by weighted summation of the instantaneous temperature of the disk surface and the model-inferred temperature at the current moment, based on the sensor confidence factor.

[0144] The closed-loop control module is used to generate a power control signal for adjusting the heating element based on the fused feedback temperature and the preset target temperature.

[0145] This embodiment provides an adaptive temperature control system for an intelligent grilling device based on a multimodal sensor;

[0146] The system includes a multi-dimensional acquisition module, which integrates an infrared temperature sensor, a wide-angle camera, a power metering chip, an angle sensor, and an NTC thermistor for real-time acquisition of physical quantities. The confidence calculation module, built into the microcontroller, executes the confidence decay model algorithm, monitors the volatility and rate of change of the data, and outputs a confidence factor. The thermodynamic deduction module executes the discretized lumped parameter heat balance equation, independently calculating the theoretical thermal state of the equipment without relying on external sensing; the data fusion module executes weighted logic, combining the observed values ​​and the deduced values ​​into unified decision data; the closed-loop control module executes the PID algorithm, converting the final decision into an electrical signal to drive the heating element to adjust its power; the modules work together to achieve intelligent temperature control with self-diagnosis and self-adaptation capabilities.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An adaptive temperature control method for intelligent grilling equipment based on multimodal sensors, characterized in that, include: Step 1: The instantaneous temperature of the plate is collected by an infrared sensor, the average brightness of the food surface image is collected by a wide-angle camera, the real-time input power of the heating element is collected by a power detection circuit, the opening and closing angle of the cover is collected by an angle sensor, and the ambient temperature is collected by an NTC sensor at the cold end of the machine body. Step 2: Calculate the temperature fluctuation variance within a preset sliding window based on the instantaneous temperature of the disk surface; Step 3: Calculate the rate of change of the cover plate angle based on the opening and closing angle of the cover plate; Step 4: Combining the temperature fluctuation variance and the cover plate angle change rate, the confidence factor of the sensor is generated using the confidence decay model; Step 5: Based on the real-time input power, ambient temperature, preset thermodynamic parameters, and the model-induced temperature of the previous moment, calculate the model-induced temperature of the current moment using the discretized lumped parameter heat balance equation. Step 6: Based on the sensor confidence factor, determine the first weight of the instantaneous temperature of the disk surface and the second weight of the model-inferred temperature at the current moment; Step 7: Combine the first weight, the instantaneous temperature of the trading surface, the second weight, and the model-deduced temperature at the current moment, and generate the fused feedback temperature through weighted summation calculation; Step 8: Based on the fused feedback temperature and the preset target temperature, a power control signal for adjusting the heating element is generated through a closed-loop feedback control algorithm. Using a confidence decay model, a sensor confidence factor is generated, including: The temperature fluctuation variance is multiplied by a preset noise sensitivity coefficient to generate a temperature noise term; The absolute value of the rate of change of the cover angle is multiplied by the preset cover opening interference penalty coefficient to generate the cover opening interference term; The change in the average image brightness is multiplied by a preset oil fume sensitivity coefficient to generate an oil fume occlusion term. Add the temperature noise term, the lid opening interference term, and the oil fume blocking term to the constant 1 to generate the denominator term; Calculate the quotient of the numerical value 1 divided by the denominator, and determine the quotient as the sensor confidence factor.

2. The adaptive temperature control method for intelligent grilling equipment based on multimodal sensors according to claim 1, characterized in that, Using the discretized lumped-parameter heat balance equation, the model-estimated temperature at the current moment is calculated, including: Calculate the product of real-time input power and preset electrothermal conversion efficiency to generate energy input items; Calculate the difference between the model-induced temperature and the ambient temperature at the previous moment; The heat loss term is calculated based on the difference, the preset comprehensive convective heat transfer coefficient, and the preset effective heat dissipation area. Determine the heat absorption power of water evaporation at the current moment; Calculate the net energy increment of the system based on energy input, heat loss, and heat absorption power from water evaporation. Multiply the net energy increment of the system by the preset sampling period and divide by the preset equivalent heat capacity of the system to generate the temperature rise increment; The model-predicted temperature from the previous moment is added to the temperature rise increment to generate the model-predicted temperature for the current moment.

3. The adaptive temperature control method for intelligent grilling equipment based on multimodal sensors according to claim 2, characterized in that, Determining the heat absorption power of water evaporation at the current moment includes: Obtain the cumulative running time since the heating operation started; Calculate the product of the preset attenuation coefficient and the cumulative running time to generate the attenuation index; The heat absorption power of water evaporation is generated by calculating the product of the preset maximum evaporation power and the negative exponent of the natural constant.

4. The adaptive temperature control method for intelligent grilling equipment based on multimodal sensors according to claim 1, characterized in that, The first weight for determining the instantaneous temperature of the trading surface and the second weight for the model-derived temperature at the current moment are as follows: The sensor confidence factor was directly determined as the first weight. Calculate the difference between value 1 and the sensor confidence factor, and determine the difference as the second weight.

5. The adaptive temperature control method for intelligent grilling equipment based on multimodal sensors according to claim 1, characterized in that, A closed-loop feedback control algorithm is used to generate a power control signal for adjusting the heating element, including: Calculate the difference between the preset target temperature and the fusion feedback temperature to generate the current temperature deviation; Based on the current temperature deviation, the temperature deviation at the previous moment, and the temperature deviation at the moment before that, combined with the preset proportional coefficient, integral coefficient, and derivative coefficient, the incremental PID algorithm is used to calculate the power adjustment increment. The power control signal from the previous moment is added to the power adjustment increment to generate the power control signal for the current moment.

6. The adaptive temperature control method for intelligent grilling equipment based on multimodal sensors according to claim 1, characterized in that, It also includes an initialization step: In response to device startup, detect the initial temperature read by the infrared sensor; Set the initial value of the model-induced temperature from the previous moment as the initial temperature.

7. An adaptive temperature control system for intelligent grilling equipment based on multimodal sensors, applied to the adaptive temperature control method for intelligent grilling equipment based on multimodal sensors as described in any one of claims 1-6, characterized in that, include: The multi-dimensional acquisition module is used to acquire the instantaneous temperature of the plate surface through an infrared sensor, the average brightness of the food surface image through a wide-angle camera, the real-time input power of the heating element through a power detection circuit, the opening and closing angle of the cover through an angle sensor, and the ambient temperature through an NTC sensor at the cold end of the machine body. The confidence calculation module is used to combine the temperature fluctuation variance calculated based on the instantaneous temperature of the disc surface and the cover angle change rate calculated based on the cover opening and closing angle, and generate the sensor confidence factor using the confidence decay model. The thermodynamic deduction module is used to calculate the model deduction temperature at the current moment based on real-time input power, ambient temperature, preset thermodynamic parameters, and the model deduction temperature at the previous moment, using the discretized lumped parameter heat balance equation. The data fusion module is used to generate a fused feedback temperature by weighted summation of the instantaneous temperature of the disk surface and the model-inferred temperature at the current moment, based on the sensor confidence factor. The closed-loop control module is used to generate a power control signal for adjusting the heating element based on the fused feedback temperature and the preset target temperature.

Images