On-board feature extraction and selection from high frequency electricity consumption data
View Patent ↗Power consumption related data for the power source supplying an entire household or other group of appliances is captured at a relatively high data rate. The raw data is analyzed to remove redundancy and to identify pertinent features, and then the more informative of these features are then selected. The analysis and selection of features are pre-processing tasks that are performed on-site, on the same piece of hardware that acquires the raw data and transmits the pre-processed data to the server. By transmitting selective data to a remote server, only a low bandwidth is required for the transmission.
1. A method for monitoring an electricity supply comprising:
receiving raw data from a sensor that senses the electricity supply;
extracting features from the raw data by calculating feature vectors, wherein the extracting is performed over successive predetermined time periods and the feature vectors each comprise:
a timestamp identifying the time period during which the feature vector was extracted; and
a number indicating a cycle of electricity that was supplied within the identified time period and to which the feature vector relates;
selecting feature vectors of interest from the feature vectors; and
transmitting the feature vectors of interest to a server.
2. The method of claim 1 wherein the raw data is received at a data rate of 100 Hz or above.
3. The method of claim 1 , wherein the electricity supply is a household electricity supply.
4. The method of claim 1 , wherein the electricity supply is a business electricity supply.
5. The method of claim 1 , wherein the raw data is electrical current data.
6. The method of claim 1 , wherein the raw data includes one or more of power factor; reactive power; apparent power; voltage; and energy.
7. The method of claim 1 , wherein said receiving, extracting, selecting and transmitting are performed on a common piece of hardware.
8. The method of claim 1 , wherein the feature vectors of interest are transmitted at a rate between once per minute to 1 Hz.
9. The method of claim 1 , wherein said transmitting requires a bandwidth no greater than two orders of magnitude less than a bandwidth required to receive the raw data.
10. The method of claim 1 , wherein the feature vectors each comprise at least one extrema.
11. The method of claim 1 , wherein:
the feature vectors of interest are selected by choosing those that are different from each other by at least a minimum threshold.
12. The method of claim 1 , wherein the selection of feature vectors of interest is performed according to an instruction received from the server.
13. The method of claim 1 , wherein the data is received from two or more sensors, each sensor providing raw data from a different phase of the electricity supply.
14. The method of claim 1 , wherein the extraction of feature vectors is performed by carrying out discrete cosine transformations on the raw data.
15. A device for monitoring an electricity supply comprising:
an input module configured to receive raw data from a sensor that senses the electricity supply;
an extractor module configured to extract features from the raw data by calculating feature vectors, wherein
the extractor module extracts features over successive predetermined time periods; and
the feature vectors each comprise:
a timestamp identifying the time period during which the feature vector was extracted; and
a number indicating a cycle of electricity that was supplied within the identified time period and to which the feature vector relates;
a selector module configured to select feature vectors of interest from the feature vectors; and
a communication module configured to transmit the feature vectors of interest to a server.
16. The device of claim 15 wherein:
the raw data is electrical current data;
the input module receives the raw data at a data rate of 100 Hz or above;
the electricity supply is a household electricity supply; and
the feature vectors each comprise at least one extrema.
17. The device of claim 15 , wherein:
the selector module selects feature vectors of interest by choosing those that are different from each other by at least a minimum threshold.
18. A system for monitoring an electricity supply comprising:
a device comprising:
an input module configured to receive raw data from a sensor that senses the electricity supply;
an extractor module configured to extract features from the raw data by calculating feature vectors, wherein
the extractor module extracts features over successive predetermined time periods; and
the feature vectors each comprise:
a timestamp identifying the time period during which the feature vector was extracted; and
a number indicating a cycle of electricity that was supplied within the identified time period and to which the feature vector relates;
a selector module configured to select feature vectors of interest from the feature vectors; and
a communication module configured to transmit the feature vectors of interest; and
a server comprising:
a processor configured to receive the feature vectors of interest; and
non-transitory computer readable media configured to store said feature vectors of interest;
wherein the processor is configured to analyze said feature vectors of interest.