System and method for identifying appliances under recall

1. An appliance operation signal processing system comprising:

an input for receiving an input appliance operation signal, the input appliance operation signal comprising information relating to operation of an appliance;

an output for outputting information relating to a recall status of the appliance; and

a processor configured to:

monitor the input appliance operation signal for a change in magnitude;

if the change in magnitude is detected:

identify a component event corresponding to the change in magnitude;

determine whether the component event is a designated sequence start component event;

if the component event is determined to be the designated sequence start component event:

extract one or more features from the input appliance operation signal between the designated sequence start component event and a sequence stop point;

determine information relating to the recall status of the appliance by inputting the one or more features into a first algorithm.

2. The system of claim 1 , wherein the first algorithm comprises a first classifier and wherein determining the information relating to the recall status of the appliance comprises determining a probability that the appliance corresponds to an appliance model designated as recalled, by applying the first classifier to a feature vector comprising the one or more features.

3. The system of claim 2 , wherein the first classifier is configured to determine a probability that the appliance corresponds to each of a set of appliance models designated as recalled, and wherein the processor is further configured to normalise the probabilities, determine a probability of the normalized probabilities with a highest value, determine whether the probability with the highest value is greater than a threshold value, wherein if the probability with the highest value is greater than the threshold value, the appliance is identified as corresponding to an appliance model designated as recalled.

4. The system of claim 1 , wherein the processor is further configured to, if the component event is determined to be the designated sequence start component event:

monitor the input appliance operation signal for a further change in magnitude;

identify a second component event corresponding to the further change in magnitude;

wherein extracting the one or more features comprises:

extracting a first time location information, relative to the designated sequence start component event, for the component event;

extracting a second time location information, relative to the designated sequence start component event, for the second component event; and

wherein:

a first feature vector comprises information relating to the component event and the first time location information; and

a second feature vector comprises information relating to the second component event and the second time location information.

5. The system of claim 4 , wherein:

the first time location information and the second time location information are relative to a start time of the designated sequence start component event, and wherein:

the first feature vector comprises a start time of the component event and a stop time of the component event; and

the second feature vector comprises a second start time of the second component event and a second stop time of the second component event.

6. The system according to claim 5 , wherein extracting the one or more features further comprises, for the component event:

extracting a first section of the input appliance operation signal, the first section being a section of the input appliance operation signal from a location prior to the start time of the component event to a location after the start time of the component event and before an end time of the component event;

extracting a pre-component event portion of the first section, the pre-component event portion being at least a part of the first section of the input appliance operation signal prior to the start time of the component event;

extracting a post-component event portion of the first section, the post-component event portion being at least a part of the first section of the input appliance operation signal after the start time of the component event;

transforming the pre-component event portion and the post-component event portion into a frequency domain;

subtracting the frequency domain pre-component event portion from the frequency domain post-component event portion to give a resultant frequency spectrum, wherein the first feature vector further comprises information relating to the resultant frequency spectrum for the component event.

7. The system according to claim 6 , wherein the information relating to the resultant frequency spectrum for the component event comprises a plurality of amplitude values, wherein each amplitude value in the plurality of amplitude values corresponds to a harmonic frequency in a plurality of harmonic frequencies, wherein the plurality of harmonic frequencies are positive integers multiples of a fundamental frequency of the input appliance operation signal.

8. The system according to claim 4 , wherein the processor is further configured to determine whether the component event is a designated stop component event, wherein the sequence stop point corresponds to a time location of a designated stop component event.

9. The system according to claim 1 , wherein identifying the component event corresponding to the change in magnitude comprises:

extracting a second feature from the input appliance operation signal; and

applying a second classifier to a feature vector comprising the second feature, wherein the second classifier is configured to determine:

a first probability that the change in magnitude corresponds to a first component event corresponding to an appliance category type; and

a second probability that the change in magnitude corresponds to a second component event corresponding to the appliance category type; and

the component event is identified as the first component event responsive to determining that the first probability is higher than the second probability.

10. The system according to claim 1 , wherein the sequence stop point is a fixed time period after the designated sequence start component event.

11. An appliance operation signal processing method comprising:

receiving an input appliance operation signal, the input appliance operation signal comprising information relating to operation of an appliance;

monitoring the input appliance operation signal for a change in magnitude;

if the change in magnitude is detected:

identifying a component event corresponding to the change in magnitude;

determining whether the component event is a designated sequence start component event;

if the component event is determined to be the designated sequence start component event:

extracting one or more features from the input appliance operation signal between the designated sequence start component event and a sequence stop point;

determining information relating to a recall status of the appliance by inputting the one or more features into a first algorithm;

outputting information relating to the recall status of the appliance.

12. The method according to claim 11 , wherein the first algorithm comprises a first classifier and wherein determining information relating to the recall status of the appliance by inputting the one or more features into the first algorithm comprises determining a probability that the appliance corresponds to an appliance model designated as recalled, by applying the first classifier to a feature vector comprising the one or more features.

13. A non-transitory computer readable medium comprising computer readable code configured to cause a computer to perform the method of claim 11 .