Probability-based control of cooking appliance

Abstract

A cooking appliance (100) comprising: a heating element (102) and / or a motor (104) for processing a food product from a first state to a desired state, each state being associated with a physical quantity; a sensor (106) for measuring the physical quantity; a microcontroller (110) configured to control the heating element and / or the motor at one or more setpoints; and a memory (120) connected to the microcontroller (110) for storing information, the memory (120) storing a loss function and a sensor function; wherein the microcontroller (110) receives sensor information relating to the physical quantity from the sensor (106) and is configured to: initiate processing of the food product by activating the heating element (102) and / or the motor (104); determining, at a time tn, a first probability associated with the food product being in the first state or the desired state based on the sensor information and the sensor function; determining a loss value for each setpoint of the heating element (102) and / or the motor (104), the loss value based on the first probability and the loss function; and operating the heating element (102) and / or the motor (104) at the respective setpoint having the lowest loss value.

Description

[0001] Related Applications

[0002] This application claims conventional priority from Australian Provisional Patent Application No. 2022902958, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0003] The present invention relates to a cooking appliance controlled by a microcontroller according to a probability-based control algorithm. Background Art

[0004] Consumer demands for cooking appliances increasingly include greater precision in the results of food processing steps provided by the cooking appliance. For example, a blender should now be able to reliably process ice cubes to a certain chip size, and an oven should be able to reliably cook meat to a precise internal temperature, preferably without the use of a meat temperature probe. In particular, protein-based food processing (such as cooking eggs, making custard, or recipes that are greatly affected by the boiling point of water) should automatically adjust based on environmental factors, such as whether the appliance is used in a high altitude town in South America or below sea level in the Netherlands.

[0005] Existing deterministic control solutions are increasingly unable to meet these consumer demands because the complexity of deterministic systems and their sensor inputs grow exponentially to meet these requirements, while placing demands on control system developers to gain insight into how the consumer environment may differ from the test environment. Summary of the invention

[0006] It is an object of the present invention to at least substantially address one or more of the above-mentioned disadvantages, or at least to provide a useful alternative to control systems for cooking appliances as described above.

[0007] In a first aspect, the present invention provides a cooking appliance, the cooking appliance comprising:

[0008] a heating element and / or a motor for processing the food product from a first state to a desired state, each state being associated with a physical quantity;

[0009] A sensor for measuring the physical quantity;

[0010] a microcontroller configured to control the heating element and / or the motor at one or more set points; and

[0011] a memory connected to the microcontroller for storing information, the memory storing a loss function and a sensor function;

[0012] The microcontroller receives sensor information related to the physical quantity from the sensor and is configured to:

[0013] Initiating processing of the food by activating the heating element and / or the motor;

[0014] At time t n , determining a first probability associated with the food being in the first state or the desired state based on the sensor information and the sensor function;

[0015] determining a loss value for each set point of the heating element and / or the motor, the loss value being based on the first probability and the loss function; and

[0016] The heating element and / or the motor are operated at the respective set point having the lowest loss value.

[0017] Preferably, the sensor function comprises a physics-based model defining a relationship between the physical quantity measured by the sensor and the probability that the food is in the first state or the desired state, and the microcontroller is configured to determine the first probability also based on the physics-based model.

[0018] Preferably, the physical quantity changes value between the first state and the desired state, and the sensor function comprises:

[0019] a variable function defining a relationship between the physical quantity and a probability that the food is in the first state or the desired state; and

[0020] A constant function defines a relationship between a physical constant and a probability that the food is in the first state or the desired state, wherein the physical constant does not change value between the first state and the desired state.

[0021] Preferably, the microcontroller is configured to:

[0022] At time t n+1 , based on the sensor function and the microcontroller at t n With t n+1 determining a second probability associated with the food being in the first state or the desired state by applying Bayesian reasoning to the first probability based on the sensor information received between the first and second states; and

[0023] The loss value for each set point of the heating element and / or the motor is determined based on the second probability and the loss function.

[0024] Preferably, the microcontroller is configured based on the microcontroller at t n With t n+1 The sensor information received during the period is used to update the sensor function using a Kalman filter.

[0025] Preferably, the memory stores a physics-based model defining the sensor at the time t n+1 The microcontroller is configured to determine the loss value also based on the physics-based model.

[0026] Preferably, the microcontroller is configured to determine the processing intensity of the heating element and / or the motor; and

[0027] The sensor function includes a function based on the time t n+1 and the Markov chain of the processing intensity, so that the sensor function including the Markov chain is used at time t n+2 The first probability ratio calculated using the sensor function without the Markov chain at time t n+2 The calculated third probability is closer to n+2 The second probability is calculated.

[0028] Preferably, the treatment intensity is determined by the microcontroller by determining the heat load applied to the food product based on the set point of the heating element and the time the heating element is operated at the set point.

[0029] Preferably, the food product is processed from a first state to one or more second states and subsequently to the desired state, the microcontroller being further configured to determine the first probability for each second state.

[0030] Preferably, the microcontroller is further configured to determine the second probability for each second state.

[0031] Preferably, the microcontroller is configured to determine the penalty value only for the second state where the second probability exceeds a performance threshold.

[0032] Preferably, the microcontroller is configured to determine the loss value only for the second state in which the loss function exceeds the performance threshold.

[0033] Preferably, the microcontroller is configured to determine the loss function based on user input of the desired state of the food product.

[0034] Preferably, the heating element and / or the motor has a power-off set point, and the microcontroller is configured to determine the loss function so that when the first probability that the food is in the desired state has exceeded a completion threshold, the loss value of the power-off set point is lower than the loss value of the other set points.

[0035] Preferably, at time t nThe microcontroller is configured to adjust the sensor function based on the user input of the first state and a user input model defining a relationship between the user input and a probability that the food is in the first state or the desired state.

[0036] Preferably, the physical quantity is continuous, and there is a continuous distribution of the state of the food between the first state and the desired state, and

[0037] Wherein the first probability is a continuous probability function from the first state to the desired state, and

[0038] The user input model is a continuous probability function based on the continuous distribution of the state of the food input by the user.

[0039] Preferably, the first probability is stored by the memory as first evidence, which is defined as the logarithmic ratio of the first probability that the food is in the relevant state and the first probability that the food is in any state other than the relevant state, so that the first probability can be stored as a signed floating point.

[0040] In a second aspect, the present invention provides a computer readable memory comprising executable instructions for the cooking appliance according to the first aspect, the executable instructions being suitable for configuring the microcontroller according to the first aspect.

[0041] In a third aspect, the present invention provides a method for controlling a cooking appliance, the cooking appliance comprising:

[0042] a heating element and / or a motor for processing the food product from a first state to a desired state, each state being associated with a physical quantity;

[0043] A sensor for measuring the physical quantity;

[0044] a microcontroller configured to control the heating element and / or the motor at one or more set points; and

[0045] a memory connected to the microcontroller for storing information, the memory storing a loss function and a sensor function,

[0046] The method comprises the following steps:

[0047] Initiating processing of the food by activating the heating element and / or the motor;

[0048] At time t n , determining a first probability associated with the food being in the first state or the desired state based on the sensor information and the sensor function;

[0049] determining a loss value for each set point of the heating element and / or the motor, the loss value being based on the first probability and the loss function; and

[0050] The heating element and / or the motor are operated at the respective set point having the lowest loss value. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Preferred embodiments of the present invention will now be described by way of example with reference to the accompanying drawings, in which:

[0052] Figure 1 is an isometric view of a cooking appliance according to an embodiment of the present invention.

[0053] Figure 2 is a cross-sectional view of a cooking appliance according to an embodiment of the present invention.

[0054] Figure 3 is a flow chart of a control system of a cooking appliance according to a preferred embodiment of the present invention.

[0055] Figure 4 yes Figure 3 Another flow chart of a control system includes using a physics-based model to determine a first probability.

[0056] Figure 5 yes Figure 3 Another flow chart of a control system includes determining a second probability based on a first probability using Bayesian reasoning.

[0057] Figure 6 yes Figure 3 Another flow chart of a control system for a food product comprising processing the food product to a second state before processing the food product to a desired state.

[0058] Figure 7 yes Figure 3 Another flow chart of a control system of the invention includes determining a processing intensity and advancing a physical model using a Markov chain based on the processing intensity.

[0059] Figure 8 yes Figure 3 Another flow chart of a control system including using a Kalman filter to update a sensor function.

[0060] Fig. 9 yes Figure 3 Another flow chart of a control system including using a physics based model to determine loss values.

[0061] Fig.10 yes Figure 3 Another flow chart of a control system comprising adjusting a sensor function using a user input based on a user input function. DETAILED DESCRIPTION

[0062] like Figure 1 and Figure 2 As shown, the cooking appliance 100 according to a preferred embodiment of the present invention may include, for example, a toaster 10 or a blender 20. In another embodiment, the present invention may include a coffee machine. The cooking appliance 100 generally includes at least one of a heating element 102 and / or a motor 104 for processing a food (not shown) from a first state to a desired state, each state being associated with a physical quantity. In the example of the toaster 10, the food is a piece of bread that is being processed from a first state (unbaked or even frozen) to a desired state (a toasted color selected by a user) by the heating element 102. In this specification, the term "first state" generally refers to the current state of the food, and the term "desired state" generally refers to the state of the food to be achieved by the processing of the cooking appliance 100. In this case, the physical quantity is the progress of the Maillard reaction, caramelization reaction and / or burning reaction on the surface of the toast, which is generally indicated by the browning of the surface of the toast. In the example of the blender 20, the food may be a certain amount of ice cubes that are being processed from a first state (original ice cube shape and size) to a desired state (user-selected ice cube size, or possibly ice slush). In this case, the physical quantity is the fragment size of ice cubes in a blender.

[0063] In another example, the cooking appliance may be a sous vide device (not shown), or an air convection oven (not shown), which includes both the heating element 102 and the motor 104 .

[0064] The cooking appliance 100 also includes a sensor 106 for measuring a physical quantity related to the process performed by the cooking appliance 100. In the case of the toaster 20, this may be one of a photochromic sensor, a temperature sensor, an infrared sensor. In the case of the blender 30, this may be a current sensor for determining the power consumption of the motor, an accelerometer for measuring vibrations, a camera for obtaining a visual indication of the size of the ice cube.

[0065] The cooking appliance 100 includes a microcontroller 110 configured to control the heating element 102 and / or the motor 104 at one or more set points. For example, the heating element 102 of the toaster 10 is typically operable at two set points: "on" and "off." Some toasters 10 may include the ability to operate the heating element 102 at a set point between "on" and "off" to provide a lower heat output of the heating element 102. The blender 20 may be operable at multiple set points between "off" and "full speed."

[0066] The cooking appliance 100 also includes a memory 120, which is connected to the microcontroller 110 for storing information, and the memory stores a loss function and a sensor function. The loss function defines the desirability of operating the heating element 102 and / or the motor 104 for each possible state of the food. In the example of the toaster 20, when the bread is not toasted, the loss function value of the set point "off" may be very high, while the loss function value of the set point "on" may be very low. The sensor function relates the output of the sensor 106 to a first probability. The first probability is a data set including probabilities for each possible state of the food. In the example of the toaster 20, when the sensor 106 is a temperature sensor that outputs room temperature, the first probability value for the "untoasted" state may be very high, while the first probability value for the "chroma-2" state may be very low.

[0067] The microcontroller 110 is configured to receive sensor information related to the physical quantity from the sensor 106 .

[0068] like Figure 3 As shown, in a typical basic operation, the microcontroller 110 is configured to start the processing operation by starting the heating element 102 and / or the motor 104 at step S101. After step S101, or in some cases before or at the same time as step S101, at step S103 and at time t n , the controller 110 determines a first probability associated with the food being in a first state or a desired state based on the sensor information and the sensor function. Then, at step S105, the controller 110 determines a loss value for each set point of the heating element 102 and / or the motor 104, the loss value being based on the first probability and the loss function. Finally, at step S107, the controller 110 operates the heating element 102 and / or the motor 104 at the corresponding set point having the lowest loss value.

[0069] Go to Figure 4 , the operation may be improved by including a physics-based model that defines a relationship between a physical quantity measured by sensor 106 and a first probability that the food product is in a first state or a desired state. At step S109, the physical quantity is measured using sensor 106. At step S111 and at time t n , based on the sensor information and the physics-based model, a first probability associated with the food being in a first state or a desired state is determined. Then, at step S113, the controller 110 determines a loss value for each set point of the heating element 102 and / or the motor 104, the loss value being based on the first probability and the loss function. Finally, at step S115, the controller 110 operates the heating element 102 and / or the motor 104 at the corresponding set point having the lowest loss value.

[0070] As food is processed, the sensor information collected by sensor 106 may change as the processing operation progresses. For example, a photochromic sensor may show browning. Otherwise, the natural divergence of sensor information will cause the readings of sensor 106 to vary over time. To incorporate this new information, the sensor 106 may be configured to read the food as it is processed. Figure 5 The method proposes that at step S117 and at time t n+1 , based on the sensor function and at t n With t n+1 , and determining a second probability associated with the food being in the first state or the desired state by applying Bayesian inference to the first probability. Then, at step S119, the controller 110 determines a loss value for each set point of the heating element 102 and / or the motor 104, the loss value being based on the second probability and the loss function, and operates the heating element 102 and / or the motor 104 at the corresponding set point with the lowest loss value at step S121. For continuous distribution variables, a Kalman filter can be used, such as Figure 8 As shown in step S135.

[0071] In some cases, it may be desirable to process a food product from a first state to a second state, and then process the food product from the second state to a desired state. This allows the loss function and / or sensor function to be defined such that the optimal path to the second state is first indicated by the loss value, and then the optimal path is indicated by the desired state. For example, oven baking may first require a searing step, where the set point of the heating element 102 should be relatively high to achieve the desired rapid searing, followed by a longer baking step, where the set point of the heating element 102 is lower to achieve the desired core temperature of the food product without burning the periphery. To provide this functionality, in Figure 6 In the method, the controller 110 determines the first probability also associated with the second state at step S103. In addition, at step S123, the controller 110 determines whether the food has reached the second state. If the second state has not been reached, the controller 110 returns to step S103. If the food has reached the second state, the controller 110 returns to step S125 and at time t n A first probability associated with the food being in a first state or a desired state is determined based on the sensor information and the sensor function. Then, at step S127, the controller 110 determines a loss value for each set point of the heating element 102 and / or the motor 104, the loss value being based on the first probability and the loss function. Finally, at step S129, the controller 110 operates the heating element 102 and / or the motor 104 at the corresponding set point having the lowest loss value.

[0072] Figure 5The method shown is necessarily a reactive control model that allows the controller 110 to select the set points of the heating element 102 and / or motor 104 that will least damage the desired state of the food product. However, by understanding the physical processes involved in the food processing process, it is possible to predict the future time t n+2 The possible first probability of , and adjust the loss function and / or sensor function accordingly, so that the control model is less likely to overshoot or undershoot. Figure 7 In step S131, the controller 110 determines the processing intensity of the heating element 102 and / or the motor 104. In one example, this can be done by determining the heat load (the amount of power delivered over time) applied to the food based on the set point of the heating element 102 and the time the heating element 102 operates at the set point. At step S133, the controller 110 uses a time-based method including a time-based method based on the current time t n+1 The first probability and sensor function of the Markov chain of processing intensity is determined at time t n+1 Once the controller 110 has determined the loss value using this adjusted first probability and operated the heating element 102 and / or the motor 104 using the corresponding set point, if the controller 110 at time t n+1 If the adjusted first probability has not been used, then at time t n+2 The second probability determined is related to the n+2 The third probability determined will be closer to time t n+2 Therefore, as the first probability at time t n+1 The result of integrating the heat load into the first probability using a Markov chain reduces the n+2 Corrections required to use Bayesian inference.

[0073] In some cases, it may be preferable to integrate a physics-based model into the loss function, rather than Figure 4 In the sensor function shown. Fig. 9 Similarly, the physics-based model defines the relationship between the physical quantity measured by the sensor and the probability that the food is in the first state or the desired state, but this information is applied by adjusting the loss value based on the physics-based model instead of adjusting the first probability.

[0074] Finally, if Fig.10As shown, some information may be obtained through user input. Such as the "frozen", "fruit bread" or "scone" buttons found on some toasters. However, since user error may occur, even if the "fruit bread" button has been selected at step S139, the user input should be adjusted using a user input function that allows for the possibility that the bread type is not "fruit bread". Therefore, the user input is just another function acting on the first probability, and is not deterministic.

[0075] Available Figure 1 A basic concrete example of this operation is explained with reference to a toaster 10 of FIG. 1 . The toaster 10 may have discrete chromaticity settings that the user operates to indicate the desired degree of chromaticity on their toast, i.e., the desired state of the food. The chromaticity s may be an integer between 0 and S. For frozen toast, it may be useful to assign a chromaticity s=-1. In terms of other variables of food state, the bread being toasted may have different properties. For example, it may be sour dough bread, brioche, fruit bread, etc. It is well known that these types of bread react differently to heating. These may be enumerated using index values ​​i from 0 to M.

[0076] Assuming these attributes of food are mutually exclusive and exhaustive, we can assume that: in is a first probability, being the current probability that the toast has chroma s and bread type i, given the initial information I provided to the controller 110. One input to the initial information may be a user input, such as a "frozen", "fruit bread" or "scone" button found on some toasters. As discussed above, a user input model may be used to factor the user input into the first probability. For example, if the user indicates that the bread is a "scone", there is a non-zero probability that the bread is not a "scone". The first probability is adjusted accordingly. Some basic assumptions may also be made, such as based on market research that one-third of bread is frozen at the beginning, nearly two-thirds of bread is untoasted, and a small portion is partially toasted.

[0077] While the food is being processed, the controller 110 will collect further data E. The data D may be obtained from sensors such as temperature sensors, photochromic sensors, light sensors, pressure sensors, humidity sensors, oxygen or other gas sensors. The data E may be associated with the state of the food, in this example, the color s and the bread type i.

[0078] Generally speaking, according to Figure 5 The method uses this data to determine a second probability p(θ|DI). This can be obtained using Bayes' theorem:

[0079]

[0080] p(E|θI), which is the probability of collecting data E given a particular bread state and prior information, can be estimated or determined using the stored model and forms a functional part of the sensor function. p(E|I) is not particularly important because it does not involve active variables and is a normalizing term. p(θ|I) is a first probability.

[0081] According to the method, the loss value is now determined by the controller 110 for each set point of the heating element 102. It can be assumed that the toaster 10 has two set points: "on" (D1) and "off" (D0). The loss value can be expressed as:

[0082]

[0083] <l>0=∑ j p(θ j |EI)L(D0,θ j ).

[0084] Where L() is the value of the assumed set point D given the initial information I and further collected data E. i A set of possible current states θ j An example of a loss function might be:

[0085]

[0086] where θ s is the toast chromaticity for which the first probability of loss value is being calculated. In this example, the target chromaticity is 2, and for the chromaticity θ s When the state exceeds 2, the loss value D1 that determines the continued heating is very high, and for the chromaticity θ s Below 2, the loss value for deciding to stop heating is higher. Therefore, when the first probability indicates that the bread may be below color 2, the controller 110 will continue heating. The shape of the loss function (linear in the above case) can be adjusted to make a faster or slower decision to stop operating the heating element 102.

[0087] In different examples, the variables defining the state of the food can be continuous. For example, when a piece of meat is baked in an oven, the core temperature, surface temperature, or other characteristics of the food are continuous. In these cases, the loss value can be obtained by the following formula:

[0088] <l> j =∫p(θ)L(D j ,θ;α)dθ

[0089] Where p(θ) is a first probability, although of course a second probability p(θ|EI) (determined using a Kalman filter) may also be used, and α is one of the characteristics of the food. If more than one characteristic is monitored, each characteristic is integrated separately to obtain the expected loss value. The loss function relating characteristic α to the desired food state can be defined similarly to the discrete example provided above.

[0090] Performing these types of operations on a microprocessor, which is typically used as the controller 110 in a desktop device, can be difficult due to limitations on the memory 120. Floating point variables can be used to efficiently store values ​​in a microprocessor. To help store the probabilities and values ​​involved in these calculations, which are often very small or very large, the following conversion tools can be used to store the information as evidence:

[0091]

[0092] This converts probabilities represented in the space 0 to 1 into a probability format. For example, a 10:1 probability would be probability 0.09090909.. and could be represented as evidence 10. Therefore, the evidence can use the entire addressable space of signed floating point numbers, rather than just the space between 0 and 1, improving the efficiency and accuracy of the computation. Many operations required in the control algorithm can be performed directly using evidence. For example, a Bayesian evidence update might take the following form:

[0093]

[0094] If necessary, the probability can be recovered from the evidence using:

[0095] p(A|I)=(1+10 -e (A|I) / 10 ) -1

[0096] In some cases, it may be preferred to divide the sensor function into a variable function and a constant function, wherein the variable function defines the relationship between the physical quantity and the probability that the food is in a first state or a desired state, and the constant function is related to the physical quantity or parts of the measurement signal derived from the physical quantity, which parts do not change value between the first state and the second state.

[0097] To improve computation time, the controller can be configured to determine the loss value only for the first state, the second state, or the desired state, where the second probability for the state exceeds the performance threshold. For states with very low probabilities, the value of the loss function is unlikely to raise the probability above the probabilities of other states.

[0098] The advantages of the disclosed method will now be discussed.

[0099] Because the heating element 102 and / or motor 104 are operated on the basis of a probabilistic heating algorithm, the control algorithm is better able to absorb differences in environmental factors, different initial conditions prior to the process operation, and operational differences between devices. The incorporation of physics-based models allows for meaningful incorporation of data from the sensors 106 to help determine a first probability and / or loss value upon which the controller 110 makes decisions between set points. Thus, calibration curves, models, and / or regressions may be used to help feed sensor information into a probabilistic food process control model.

[0100] Decomposing the sensor function into a variable function and a constant function reduces the computational load on the potentially limited controller 110 , which is typically embodied as an embedded processing device of limited capabilities.

[0101] Using Bayesian inference and / or Kalman filters to update the probabilistic control model based on new evidence allows for continuous updating of the probability distribution underlying the control model. By including a Markov chain in the sensor function, the model is advanced based on processing intensity (such as thermal load), allowing the controller 110 to advance the food product to a desired state more quickly, reducing overshoot or undershoot by reducing the difference between probability updates in each Bayesian inference or Kalman filter update.< / l> < / l>

Claims

1. A cooking utensil, comprising: a heating element and / or a motor for processing the food product from a first state to a desired state, each state being associated with a physical quantity; A sensor for measuring the physical quantity; a microcontroller configured to control the heating element and / or the motor at one or more set points; and a memory connected to the microcontroller for storing information, the memory storing a loss function and a sensor function; The microcontroller receives sensor information related to the physical quantity from the sensor and is configured to: Initiating processing of the food product by activating the heating element and / or the motor; At time t n , determining a first probability associated with the food being in the first state or the desired state based on the sensor information and the sensor function; determining a loss value for each set point of the heating element and / or the motor, the loss value being based on the first probability and the loss function; and The heating element and / or the motor are operated at the respective set point having the lowest loss value.

2. The cooking appliance of claim 1 , wherein the sensor function comprises a physics-based model defining a relationship between the physical quantity measured by the sensor and a probability that the food is in the first state or the desired state, and the microcontroller is configured to determine the first probability also based on the physics-based model.

3. The cooking appliance according to claim 1 or 2, wherein the physical quantity changes value between the first state and the desired state, and the sensor function comprises: a variable function, wherein the variable function defines a relationship between the physical quantity and a probability that the food is in the first state or the desired state; and A constant function that defines a relationship between a physical constant and a probability that the food is in the first state or the desired state, wherein the physical constant does not change value between the first state and the desired state.

4. The cooking appliance according to any one of claims 1 to 3, wherein the microcontroller is configured to: At time t n+1 , based on the sensor function and the microcontroller at t n With t n+1 determining a second probability associated with the food being in the first state or the desired state by applying Bayesian reasoning to the first probability based on the sensor information received between the two states; and The loss value for each set point of the heating element and / or the motor is determined based on the second probability and the loss function.

5. The cooking appliance according to claim 4, wherein the microcontroller is configured to generate a signal based on the signal received by the microcontroller at t n With t n+1 The sensor information received during the period is used to update the sensor function using a Kalman filter.

6. The cooking appliance according to claim 4 or 5, wherein the memory stores a physics-based model defining the temperature at which the sensor n+1 The microcontroller is configured to determine the loss value also based on the physics-based model.

7. The cooking appliance according to any one of claims 4 to 6, wherein the microcontroller is configured to determine the processing intensity of the heating element and / or the motor; and The sensor function includes a function based on the time t n+1 and the Markov chain of the processing intensity, so that the sensor function including the Markov chain is used at time t n+2 The first probability ratio calculated using the sensor function without the Markov chain at time t n+2 The calculated third probability is closer to the time t n+2 Calculate the second probability.

8. The cooking appliance of claim 7, wherein the treatment intensity is determined by the microcontroller by determining a heat load applied to the food product based on the set point of the heating element and the time the heating element is operated at the set point.

9. The cooking appliance of any one of claims 1 to 8, wherein the food is processed from a first state to one or more second states and then to the desired state, the microcontroller being further configured to determine the first probability for each second state.

10. The cooking appliance of claim 9 when dependent on claim 4, wherein the microcontroller is further configured to determine the second probability for each second state. 11 . The cooking appliance of claim 10 , wherein the microcontroller is configured to determine the penalty value only for a second state in which the second probability exceeds a performance threshold.

12. The cooking appliance of claim 10 or 11, wherein the microcontroller is configured to determine the loss value only for a second state in which the loss function exceeds the performance threshold.

13. The cooking appliance of any one of claims 1 to 12, wherein the microcontroller is configured to determine the loss function based on user input of the desired state of the food product.

14. The cooking appliance of claim 13, wherein the heating element and / or the motor has a power-off set point, and the microcontroller is configured to determine the loss function so that when the first probability that the food is in the desired state has exceeded a completion threshold, the loss value of the power-off set point is lower than the loss values ​​of the other set points.

15. The cooking appliance according to any one of claims 1 to 14, wherein at time t n , the microcontroller is configured to adjust the sensor function based on a user input of the first state and a user input model defining a relationship between the user input and a probability that the food is in the first state or the desired state.

16. The cooking appliance according to claim 15, wherein the physical quantity is continuous, and there is a continuous distribution of states of the food between the first state and the desired state, and wherein the first probability is a continuous probability function from the first state to the desired state, and The user input model is a continuous probability function on the continuous distribution of the state of the food based on the user input.

17. The cooking appliance according to any one of claims 1 to 16, wherein the first probability is stored by the memory as a first evidence, the first evidence being defined as a logarithmic ratio of the first probability that the food is in a relevant state to the first probability that the food is in any state other than the relevant state, so that the first probability can be stored as a signed floating point.

18. A computer readable memory comprising executable instructions for a cooking appliance according to any one of claims 1 to 17, the executable instructions being adapted to configure a microcontroller according to any one of claims 1 to 17.

19. A method for controlling a cooking appliance, the cooking appliance comprising: a heating element and / or a motor for processing the food product from a first state to a desired state, each state being associated with a physical quantity; A sensor for measuring the physical quantity; a microcontroller configured to control the heating element and / or the motor at one or more set points; and a memory connected to the microcontroller for storing information, the memory storing a loss function and a sensor function, The method comprises the following steps: Initiating processing of the food product by activating the heating element and / or the motor; At time t n , determining a first probability associated with the food being in the first state or the desired state based on the sensor information and the sensor function; determining a loss value for each set point of the heating element and / or the motor, the loss value being based on the first probability and the loss function; and The heating element and / or the motor are operated at the respective set point having the lowest loss value.

Images