Temperature sensor calibration using recurrent network
A method for determining ambient temperature of a device in a building control system. The method includes obtaining a measurement of ambient temperature from a first temperature sensor of the device and obtaining a measurement of temperature of a processing circuit of the device. These temperature measurements along with a brightness level of a display of the device and a state of a relay of the device are provided as input to a neural network. The neural network is trained to recognize these inputs in order to more accurately determine the ambient temperature of the device. The reading from the first temperature sensor may otherwise be inaccurate due to heat generated by electronic components of the device.
1. A method for determining an actual ambient temperature value associated with a measured ambient temperature in a building control system, the method comprising:
obtaining a first temperature measurement from a first temperature sensor of a device, the first temperature sensor configured to sense the measured ambient temperature, wherein heat generated by electronic components of the device affects the measured ambient temperature causing the first temperature measurement to be inaccurate relative to the actual ambient temperature value;
obtaining a second temperature measurement from a second temperature sensor of the device, the second temperature sensor configured to measure a temperature of a processing circuit of the device;
determining a brightness level associated with a display of the device;
determining a state of a relay associated with the device, the state of the relay indicative of a level of electrical energy powering the device;
providing the first temperature measurement, the second temperature measurement, the brightness level, and the state of the relay as input to a neural network; and
determining the actual ambient temperature value associated with the measured ambient temperature based on an output of the neural network.
2. The method of claim 1 , further comprising training the neural network with a time series dataset comprising the first temperature measurement, the second temperature measurement, the brightness level, the state of the relay, and a measurement of the actual ambient temperature generated by a third temperature sensor that is separate from the device.
3. The method of claim 1 , further comprising determining, by the neural network, an internal heat index associated with the device using the second temperature measurement, the brightness level, and the state of the relay, wherein the internal heat index comprises an estimation of the heat generated by the electronic components of the device.
4. The method of claim 3 , wherein determining the internal heat index comprises determining a temperature of the display based on the brightness level and an amount of time since the brightness level has changed.
5. The method of claim 4 , wherein determining the internal heat index further comprises determining a temperature of the relay based on the state of the relay and an amount of time since the state of the relay has changed.
6. The method of claim 5 , wherein the internal heat index comprises a weighted sum of the second temperature measurement, the temperature of the display, and the temperature of the relay.
7. A method for determining an actual ambient temperature value associated with a measured ambient temperature in a building control system, the method comprising:
obtaining a first temperature measurement from a first temperature sensor of a device, the first temperature sensor configured to sense the measured ambient temperature;
obtaining a second temperature measurement from a second temperature sensor of the device, the second temperature sensor configured to measure a temperature of a processing circuit of the device;
determining a brightness level associated with a display of the device;
determining a state of a relay associated with the device, the state of the relay indicative of a level of electrical energy powering the device;
providing the first temperature measurement, the second temperature measurement, the brightness level, and the state of the relay as input to a neural network;
determining the actual ambient temperature value associated with the measured ambient temperature based on an output of the neural network; and
determining, by the neural network, an internal heat index associated with the device using the second temperature measurement, the brightness level, and the state of the relay, wherein the internal heat index comprises an estimation of heat generated by electronic components of the device, wherein the output of the neural network comprises a weighted sum of the first temperature measurement and the internal heat index.
8. The method of claim 1 , wherein the neural network comprises a recurrent neural network, and wherein an activation function associated with at least one node in the neural network comprises a hyperbolic tangent function.
9. A device in a building control system, the device comprising:
a first temperature sensor configured to generate a first temperature measurement that is indicative of a measured ambient temperature near the device, wherein heat generated by electronic components of the device affects the measured ambient temperature causing the first temperature measurement to be inaccurate relative to an actual ambient temperature value;
a relay configured to control an amount of electrical energy powering the device;
a display configured to provide a user interface to a user of the device; and
a processing circuit comprising a second temperature sensor configured to generate a second temperature measurement that is indicative of a temperature of the processing circuit, the processing circuit configured to:
obtain the first temperature measurement from the first temperature sensor;
obtain the second temperature measurement from the second temperature sensor;
determine a brightness level associated with the display;
determine a state of the relay;
provide the first temperature measurement, the second temperature measurement, the brightness level, and the state of the relay as input to a neural network; and
determine the actual ambient temperature value associated with the measured ambient temperature near the device based on an output of the neural network.
10. The device of claim 9 , wherein the neural network is trained with a time series dataset comprising the first temperature measurement, the second temperature measurement, the brightness level, the state of the relay, and a measurement of the actual ambient temperature generated by a third temperature sensor that is separate from the device.
11. The device of claim 9 , wherein the neural network is configured to determine an internal heat index associated with the device using the second temperature measurement, the brightness level, and the state of the relay, wherein the internal heat index comprises an estimation of heat generated by electronic components of the device.
12. The device of claim 11 , wherein the neural network is configured to determine a temperature of the display based on the brightness level and an amount of time since the brightness level has changed.
13. The device of claim 12 , wherein the neural network is configured to determine a temperature of the relay based on the state of the relay and an amount of time since the state of the relay has changed.
14. The device of claim 13 , wherein the internal heat index comprises a weighted sum of the second temperature measurement, the temperature of the display, and the temperature of the relay.
15. The device of claim 11 , wherein the output of the neural network comprises a weighted sum of the first temperature measurement and the internal heat index.
16. The device of claim 9 , wherein the neural network comprises a recurrent neural network and an activation function associated with at least one node in the neural network comprises a hyperbolic tangent function.
17. A method for determining an actual ambient temperature value in a building control system, the method comprising:
obtaining a first temperature measurement from a first temperature sensor of a device, the first temperature sensor configured to measure a measured ambient temperature, wherein heat generated by electronic components of the device affects the measured ambient temperature causing the first temperature measurement to be inaccurate relative to the actual ambient temperature value;
obtaining a second temperature measurement from a second temperature sensor of the device, the second temperature sensor configured to measure a temperature of a processing circuit of the device;
determining a brightness level associated with a display of the device;
providing the first temperature measurement, the second temperature measurement, and the brightness level as input to a neural network; and
determining the actual ambient temperature value associated with the measured ambient temperature based on an output of the neural network.
18. The method of claim 17 , further comprising training the neural network with a time series dataset comprising the first temperature measurement, the second temperature measurement, the brightness level, and a measurement of the measured ambient temperature generated by a third temperature sensor that is separate from the device.
19. The method of claim 17 , further comprising determining, by the neural network, an internal heat index associated with the device using the second temperature measurement and the brightness level, wherein the internal heat index comprises an estimation of heat generated by electronic components of the device.
20. The method of claim 19 , wherein the output of the neural network comprises a weighted sum of the first temperature measurement and the internal heat index.