Systems and methods for machine learning zone-based foreign object detection for wireless power transmission
An example method is provided for detecting and classifying foreign objects. The method includes after a predetermined period of time, for a zone of a plurality of zones, enabling a beacon signal, obtaining a set of electrical measurements, and disabling the beacon signal. The method also includes determining, based on the set of electrical measurements and a trained classifier, a status of the zone. The status of the zone indicates whether there is a presence of foreign objects within the zone of the plurality of zones. The method further includes in accordance with an indication that there is the presence of the foreign objects, forgo transmitting wireless power to one or more wireless power receivers. The method further includes in accordance with an indication that there is an absence of the foreign objects, transmitting wireless power to the one or more wireless power receivers.
1. A non-transitory, computer-readable storage medium storing instructions that, when executed by one or more processors of a wireless-power transmitter including a wireless-power-transmitting antenna, cause the wireless-power transmitter to perform operations, comprising:
after the wireless-power transmitter has been enabled for a predetermined period of time, for each zone of a plurality of zones:
enabling a beacon signal,
obtaining a set of electrical measurements, wherein the set of electrical measurements includes measurements of a plurality of electrical features of an object;
selecting, via a feature selection algorithm, two measurements from the set of electrical measurements having a highest object categorization performance when plotted against one another, wherein the two measurements are measurements of different electrical features;
disabling the beacon signal, and
determining, using a trained classifier configured to receive the set of electrical measurements, a zone status indicating presence of a foreign object within a respective zone of the plurality of zones, wherein the trained classifier is a clustering algorithm trained to classify known and unknown objects based on the two selected measurements;
in accordance with a first zone status indicating that a foreign object is present within a first zone of the plurality of zones, forgo transmitting wireless power to one or more wireless power receivers within the first zone of the plurality of zones; and
in accordance with a second zone status indicating an absence of a foreign object within a second zone of the plurality of zones, transmitting wireless power to one or more wireless power receivers within the second zone of the plurality of zones.
2. The non-transitory computer readable storage medium of claim 1 , wherein obtaining the set of electrical measurements for each zone of the plurality of zones further comprises forming a respective feature vector for each set of electrical measurements.
3. The non-transitory computer readable storage medium of claim 2 , wherein the instructions, when executed by the one or more processors of the wireless-power transmitter, cause the wireless-power transmitter to perform operations, further comprising:
for each zone of a plurality of zones:
continuing to (i) obtain the set of electrical measurements, (ii) form the respective feature vector, and (iii) determine a respective zone status by inputting the respective feature vector to the trained classifier; and
waiting to transmit wireless power to the one or more wireless power receivers until the trained classifier stops detecting that the one or more foreign objects are present.
4. The non-transitory computer readable storage medium of claim 2 , wherein forming the feature vector further comprises calculating a respective feature data for the set of electrical measurements, and aggregating the respective feature data to obtain the feature vector.
5. The non-transitory computer readable storage medium of claim 1 , wherein the trained classifier includes a machine-learning model and applies reinforcement learning to adjust one or more weights used in the machine-learning model upon determining that (i) the trained classifier has detected the presence of the one or more foreign objects and (ii) wireless power received by the one or more wireless power receivers indicates that no foreign object is present.
6. The non-transitory computer readable storage medium of claim 1 , wherein the trained classifier is further trained to classify the one or more foreign objects into one or more categories.
7. The non-transitory computer readable storage medium of claim 1 , wherein the set of electrical measurements is obtained at predetermined time intervals.
8. The non-transitory computer readable storage medium of claim 1 , wherein the set of electrical measurements include measurements of one or more of: load impedance, reflective power, forward power, drive current, drive voltage, magnetics, and temperature.
9. The non-transitory computer readable storage medium of claim 1 , wherein the trained classifier detects one or more of: coins, credit cards, and radio-frequency identification (RFID) tags.
10. The non-transitory computer readable storage medium of claim 1 , wherein the beacon signal is transmitted as a low power burst for a short period of time.
11. The non-transitory computer readable storage medium of claim 1 , wherein the beacon signal is transmitted using transmission parameters selected from a predetermined set of transmission parameters specifying at least a power level and a frequency.
12. The non-transitory computer readable storage medium of claim 1 , wherein the trained classifier includes a K-means classifier trained to detect the one or more wireless power receivers and to distinguish the one or more wireless power receivers from the one or more foreign objects.
13. The non-transitory computer readable storage medium of claim 1 , wherein the trained classifier includes a k-nearest neighbor (k-NN) classifier trained to classify the one or more wireless power receivers and the one or more foreign objects into at least one category of a plurality of categories of objects.
14. The non-transitory computer readable storage medium of claim 1 , wherein the wireless-power-transmitting antenna is configured to match load impedance of the one or more wireless power receivers for optimal power transfer.
15. The non-transitory computer readable storage medium of claim 1 , wherein the wireless-power-transmitting antenna is a near-field transmitting antenna.
16. A wireless-power transmitter, comprising:
a wireless-power-transmitting antenna;
one or more processors; and
memory storing one or more programs that are configured for fer execution by the one or more processors, the one or more programs including instructions for:
after the wireless-power transmitter has been enabled for a predetermined period of time, for each zone of a plurality of zones:
enabling a beacon signal,
obtaining a set of electrical measurements, wherein the set of electrical measurements includes measurements of a plurality of electrical features of an object;
selecting, via a feature selection algorithm, two measurements from the set of electrical measurements having a highest object categorization performance when plotted against one another wherein the two measurements are measurements of different electrical features;
disabling the beacon signal, and
based on the two selected measurements, determining, using a trained classifier configured to receive the set of electrical measurements, a zone status indicating presence of a foreign object within a respective zone of the plurality of zones;
in accordance with a first zone status indicating that a foreign object is present within a first zone of the plurality of zones, forgo transmitting wireless power to one or more wireless power receivers within the first zone of the plurality of zones; and
in accordance with a second zone status indicating an absence of a foreign object within a second zone of the plurality of zones, transmitting wireless power to one or more wireless power receivers within the second zone of the plurality of zones.
17. The wireless-power transmitter of claim 16 , wherein obtaining the set of electrical measurements for each zone of the plurality of zones further comprises forming a respective feature vector for each set of electrical measurements.
18. The wireless-power transmitter of claim 17 , wherein the one or more programs stored in the memory further include instructions for:
for each zone of a plurality of zones:
continuing to (i) obtain the set of electrical measurements, (ii) form the respective feature vector, and (iii) determine a respective zone status by inputting the respective feature vector to the trained classifier; and
waiting to transmit wireless power to the one or more wireless power receivers until the trained classifier stops detecting that the one or more foreign objects are present.
19. A method of detecting and classifying foreign objects, comprising:
after a wireless-power transmitter including a wireless-power-transmitting antenna has been enabled for a predetermined period of time, for each zone of a plurality of zones of the wireless-power transmitter:
enabling a beacon signal,
obtaining a set of electrical measurements, wherein the set of electrical measurements includes measurements of a plurality of electrical features of an object;
selecting, via a feature selection algorithm, two measurements from the set of electrical measurements having a highest object categorization performance when plotted against one another, wherein the two measurements are measurements of different electrical features;
disabling the beacon signal, and
based on the two selected measurements, determining, using a trained classifier configured to receive the set of electrical measurements, a zone status indicating presence of a foreign object within a respective zone of the plurality of zones;
in accordance with a first zone status indicating that a foreign object is present within a first zone of the plurality of zones, forgo transmitting wireless power to one or more wireless power receivers within the first zone of the plurality of zones; and
in accordance with a second zone status indicating an absence of a foreign object within a second zone of the plurality of zones, transmitting wireless power to one or more wireless power receivers within the second zone of the plurality of zones.
20. The method of claim 19 , wherein obtaining the set of electrical measurements for each zone of the plurality of zones further comprises forming a respective feature vector for each set of electrical measurements.