Electronic oven with infrared evaluative control

1. A non-transitory computer-readable medium storing instructions to execute a computer-implemented method for heating an item in a chamber of an electronic oven towards a target state comprising:

heating the item with a set of applications of energy to the chamber while the electronic oven is in a respective set of physical configurations;

sensing, using an infrared sensor, sensor data that defines a respective set of responses by the item to the set of applications of energy;

generating a plan to heat the item in the chamber, wherein the generating is conducted by a control system of the electronic oven, and wherein the generating uses the sensor data;

conducting an optimization analysis to produce an optimization output; and

compiling a sequence of commands using the optimization output;

wherein the optimization analysis uses: (i) a convex optimization solver; (ii) data representing the target state; and (iii) data representing the respective set of responses;

wherein the electronic oven executes the sequence of commands to heat the item towards the target state; and

wherein the set of applications of energy and respective set of physical configurations define a respective set of variable distributions of energy in the chamber.

2. The non-transitory computer-readable medium of claim 1 , wherein:

the optimization output includes an error value and a duration vector;

the duration vector includes a set of durations for the electronic oven to be in each of the physical configurations in the respective set of physical configurations;

the error value quantifies a difference between the target state and an extrapolated end state; and

the convex optimization solver sets the duration vector to minimizes the error value.

3. The non-transitory computer-readable medium of claim 2 , wherein the computer-implemented method further comprises:

segmenting the item into a set of segments;

wherein the convex optimization solver is a non-negative least squares solver;

wherein the data representing the target state is a target state vector;

wherein the target state vector includes a set of target temperature values corresponding to the set of segments;

wherein the data representing the respective set of responses is a set of response vectors; and

wherein each response vector in the set of response vectors is a set of temperature derivatives corresponding to the set of segments.

4. The non-transitory computer-readable medium of claim 3 , wherein the computer-implemented method further comprises:

identifying the item using a classifier and the set of response vectors.

5. The non-transitory computer-readable medium of claim 2 , wherein the sequence of commands:

transitions the electronic oven between the physical configurations in the respective set of physical configurations;

applies the applications of energy; and

assures the electronic oven is in each of the physical configurations in the respective set of physical configurations for a duration that is proportional to a respective element in the duration vector.

6. The non-transitory computer-readable medium of claim 2 , wherein the computer-implemented method further comprises:

determining that the error value exceeds an acceptable error value;

heating the item with an application of energy while the electronic oven is in an additional physical configuration, wherein the additional physical configuration is selected upon determining that the error value exceeds the acceptable error value;

sensing, using the infrared sensor, sensor data that defines a respective response by the item to the application of energy;

conducting a second optimization analysis to produce a second optimization output; and

compiling a second sequence of commands using the second optimization output;

where the second optimization analysis uses: (i) the convex optimization solver; (ii) data representing the target state; (iii) data that represents the respective set of responses; and (iv) data that represents the respective response.

7. The non-transitory computer-readable medium of claim 1 , wherein the computer-implemented method further comprises:

executing a command in the sequence of commands, wherein the command returns the electronic oven to a physical configuration in the respective set of physical configurations;

executing a second command in the sequence of commands, wherein the command applies an application of energy to the chamber while the electronic oven is in the physical configuration; and

sensing, using the infrared sensor, additional sensor data that defines a response by the item to the application of energy;

conducting a second optimization analysis to produce a second optimization output; and

compiling a second sequence of commands using the second optimization output;

wherein the second optimization analysis uses: (i) the convex optimization solver; (ii) the data representing the target state; and (iii) the additional sensor data.

8. The non-transitory computer-readable medium of claim 1 , wherein the computer-implemented method further comprises:

periodically conducting additional iterations of the optimization analysis while the item is being heated toward the target state;

wherein a period between the additional iterations is greater than 3 seconds and less than 15 seconds.

9. The non-transitory computer-readable medium of claim 1 , wherein the computer-implemented method further comprises:

generating a second plan to heat the item in the chamber, wherein the generating of the second plan is conducted by the control system of the electronic oven, and where the generating of the second plan uses a deterministic planner;

wherein a cost function of the deterministic planner uses the plan as a heuristic to estimate a future plan cost for the second plan.

10. The non-transitory computer-readable medium of claim 9 , wherein the computer-implemented method further comprises:

automatically heating the item in the chamber towards the target state using a reinforcement learning system;

wherein the second plan is used as a policy for the reinforcement learning system.

11. The non-transitory computer-readable medium of claim 1 , wherein the computer-implemented method further comprises:

altering a reflective element in a set of reflective elements to transition between a first and second physical configuration in the respective set of physical configurations;

wherein the chamber is wholly motionless during each application of energy;

wherein the set of reflective elements includes at least three reflective elements; and

wherein the control system generates commands that independently alter the reflective elements in the set of reflective elements.

12. The non-transitory computer-readable medium of claim 11 , wherein:

an application of energy to the chamber from the set of applications of energy introduces a polarized electromagnetic wave to the chamber;

altering the reflective element alters an orientation of the reflective element from a first orientation to a second orientation;

wherein a dominant polarization of the polarized electromagnetic wave is perpendicular to the first orientation; and

wherein the dominant polarization of the polarized electromagnetic wave is parallel to the second orientation.

13. The non-transitory computer-readable medium of claim 1 , wherein the computer-implemented method further comprises:

segmenting the item into a set of segments using data from the infrared sensor;

wherein the convex optimization solver is a non-negative least squares solver;

wherein the data representing the target state is a target state vector;

wherein the target state vector includes a set of target temperature values corresponding to the set of segments;

wherein the data representing the respective set of responses is a set of response vectors; and

wherein each response vector in the set of response vectors corresponds to the set of segments.

14. The non-transitory computer-readable medium of claim 13 , wherein:

the set of segments includes at least 10 elements; and

the respective set of physical configurations includes at least 10 distinct physical configurations.

15. The computer-implemented method of claim 1 , wherein:

during each application of energy in the set of applications of energy, a respective physical configuration in the respective set of physical configurations is held for at least 0.5 seconds; and

the respective set of physical configurations includes at least 5 distinct physical configurations.

16. A computer-implemented method for heating an item in a chamber of an electronic oven towards a target state comprising:

heating the item with a set of applications of energy to the chamber while the electronic oven is in a respective set of configurations;

sensing, using an infrared sensor, sensor data that defines a respective set of responses by the item to the set of applications of energy; and

generating a plan to heat the item in the chamber, wherein the generating: (i) is conducted by a control system of the electronic oven; (ii) uses the sensor data; (iii) includes conducting an optimization analysis to produce an optimization output; and (iv) includes compiling a sequence of commands using the optimization output;

wherein the optimization analysis uses: (i) a convex optimization solver; (ii) data representing the target state; and (iii) data representing the respective set of responses; and

wherein the electronic oven executes the sequence of commands to heat the item towards the target state.

17. The computer-implemented method of claim 16 , wherein:

the respective set of configurations include at least 5 distinct physical configurations;

each application of energy is at least 0.5 seconds in duration; and

the set of applications of energy and respective set of configurations define a respective set of variable distributions of energy in the chamber.

18. The computer-implemented method of claim 16 , further comprising:

the optimization output includes an error value and a duration vector;

the duration vector includes a set of durations for the electronic oven to be in each of the configurations in the respective set of configurations;

the error value quantifies a difference between the target state and an extrapolated end state; and

the convex optimization solver sets the duration vector to minimizes the error value.

19. The computer-implemented method of claim 18 , wherein the sequence of commands:

transitions the electronic oven between the physical configurations in the respective set of physical configurations;

applies the applications of energy; and

assures the electronic oven is in each of the physical configurations in the respective set of physical configurations for a duration that is proportional to a respective element in the duration vector.

20. The computer-implemented method of claim 18 , further comprising:

determining that the error value exceeds an acceptable error value;

heating the item with an application of energy while the electronic oven is in an additional physical configuration, wherein the additional physical configuration is selected upon determining that the error value exceeds the acceptable error value;

sensing, using the infrared sensor, sensor data that defines a respective response by the item to the application of energy;

conducting a second optimization analysis to produce a second optimization output; and

compiling a second sequence of actions using the second optimization output;

where the second optimization analysis uses: (i) a convex optimization solver; (ii) data representing the target state; (iii) data that represents the respective set of responses; and (iv) data that represents the respective response.

21. The computer-implemented method of claim 16 , further comprising:

segmenting the item into a set of segments;

wherein the convex optimization solver is a non-negative least squares solver;

wherein the data representing the target state is a target state vector;

wherein the target state vector includes a set of target temperature values corresponding to the set of segments;

wherein the data representing the respective set of responses is a set of response vectors; and

wherein each response vector in the set of response vectors is a set of temperature derivatives corresponding to the set of segments.

22. The computer-implemented method of claim 21 , further comprising:

identifying the item using a classifier and the set of response vectors.

23. The computer-implemented method of claim 16 , further comprising:

periodically conducting the optimization analysis while the item is being heated toward the target state;

wherein a period for the periodic conducting of the optimization analysis is greater than 3 seconds and less than 15 seconds.

24. The computer-implemented method of claim 16 , further comprising:

executing a command in the sequence of commands, wherein the command returns the electronic oven to a physical configuration in the respective set of physical configurations;

executing a second command in the sequence of commands, wherein the command applies an application of energy to the chamber while the electronic oven is in the physical configuration; and

sensing, using the infrared sensor, additional sensor data that defines a response by the item to the application of energy;

conducting a second optimization analysis to produce a second optimization output; and

compiling a second sequence of commands using the second optimization output;

wherein the second optimization analysis uses: (i) the convex optimization solver; (ii) the data representing the target state; and (iii) the additional sensor data.

25. The computer-implemented method of claim 16 , further comprising:

generating a second plan to heat the item in the chamber, wherein the generating of the second plan is conducted by the control system of the electronic oven, and where the generating of the second plan uses a deterministic planner;

wherein a cost function of the deterministic planner uses the plan as a heuristic to estimate a future plan cost for the second plan.

26. The computer-implemented method of claim 25 , further comprising:

automatically heating the item in the chamber towards the target state using a reinforcement learning system;

wherein the second plan is used as a policy for the reinforcement learning system.

27. The computer-implemented method of claim 16 , further comprising:

altering a reflective element in a set of reflective elements to transition between a first and second physical configuration in the respective set of physical configurations;

wherein the chamber is wholly motionless during each application of energy;

wherein the set of reflective elements includes at least three reflective elements; and

wherein the control system generates commands that independently alter the reflective elements in the set of reflective elements.

28. The computer-implemented method of claim 27 , wherein:

an application of energy to the chamber from the set of applications of energy introduces a polarized electromagnetic wave to the chamber;

altering the reflective element alters an orientation of the reflective element from a first orientation to a second orientation;

wherein a dominant polarization of the polarized electromagnetic wave is perpendicular to the first orientation; and

wherein the dominant polarization of the polarized electromagnetic wave is parallel to the second orientation.

29. The computer-implemented method of claim 16 , further comprising:

segmenting the item into a set of segments using data from the infrared sensor;

wherein the convex optimization solver is a non-negative least squares solver;

wherein the data representing the target state is a target state vector;

wherein the target state vector includes a set of target temperature values corresponding to the set of segments;

wherein the data representing the respective set of responses is a set of response vectors; and

wherein each response vector in the set of response vectors corresponds to the set of segments.

30. The computer-implemented method of claim 29 , wherein:

the set of segments includes at least 10 elements; and

the respective set of physical configurations includes at least 10 distinct physical configurations.