Carrier frequency offset modeling for radio frequency fingerprinting
A method comprising operating at least one hardware processor for: receiving, by a radio frequency (RF) receiver, a plurality of training RF transmissions from an RF device, wherein each of said training RF transmissions is temporally associated with operational parameters and ambient parameters of said RF receiver and said RF device; at a training stage, training a machine learning classifier based, at least in part, on a training set comprising: (i) a Carrier Frequency Offset (CFO) value calculated for each of said training RF transmissions, and (ii) labels associated with said operational parameters and said ambient parameters; and at an inference stage, applying said machine learning classifier to determine whether one or more runtime RF transmissions originate from said RF device.
1. A method comprising operating at least one hardware processor for:
receiving, by a radio frequency (RF) receiver, a plurality of training RF transmissions from an RF device, wherein each of said training RF transmissions is temporally associated with operational parameters and ambient parameters of said RF receiver and said RF device;
at a training stage, training a machine learning classifier based, at least in part, on a training set comprising:
(i) a Carrier Frequency Offset (CFO) value calculated for each of said training RF transmissions, and
(ii) labels associated with said operational parameters and said ambient parameters; and
at an inference stage, applying said trained machine learning classifier to determine whether one or more runtime RF transmissions originate from said RF device.
2. The method of claim 1 , wherein each of said training RF transmissions is labelled with said temporally associated operational parameters and ambient parameters.
3. The method of claim 1 , wherein said training further comprises generating one or more CFO profiles for said RF device, wherein each of said CFO profiles comprises said calculated CFO values, and wherein said determining is based, at least in part, on calculating a distance metric between a measured CFO value of said runtime RF transmissions and said one or more CFO profiles.
4. The method of claim 1 , wherein said operational parameters and said ambient parameters of said RF receiver and said RF device are selected from the group consisting of: Operating temperatures, ambient temperature, ambient relative humidity, RF channel used for transmitting said RF transmissions, and a connectivity state of one or more wireless transceivers.
5. The method of claim 4 , wherein said operating temperatures include at least some of Central Processing Unit (CPU) temperature, Graphic Processing Unit (GPU) temperature, and battery temperature.
6. The method of claim 1 , wherein at least some of said plurality of operational parameters are initiated in said RF device, based, at least in part, on instructions transmitted from said RF transceiver, wherein said instructions further cause said RF device to transmit at least one training RF transmission during said initiating.
7. The method of claim 1 , wherein said machine learning classifier is further trained based, at least in part, on a training set comprising:
(i) CFO values calculated for at least some of said runtime RF transmission, and
(ii) labels associated with operational parameters and ambient parameters temporally associated with each of said at least some runtime RF transmissions.
8. A system comprising:
at least one radio frequency (RF) receiver;
at least one hardware processor; and
a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to:
receive, by a radio frequency (RF) receiver, a plurality of training RF transmissions from an RF device, wherein each of said training RF transmissions is temporally associated with operational parameters and ambient parameters of said RF receiver and said RF device;
at a training stage, train a machine learning classifier based, at least in part, on a training set comprising:
(i) a Carrier Frequency Offset (CFO) value calculated for each of said training RF transmissions, and
(ii) labels associated with said operational parameters and said ambient parameters; and
at an inference stage, apply said trained machine learning classifier to determine whether one or more runtime RF transmissions originate from said RF device.
9. The system of claim 8 , wherein each of said training RF transmissions is labelled with said temporally associated operational parameters and ambient parameters.
10. The system of claim 8 , wherein said training further comprises generating one or more CFO profiles for said RF device, wherein each of said CFO profiles comprises said calculated CFO values, and wherein said determining is based, at least in part, on calculating a distance metric between a measured CFO value of said runtime RF transmissions and said one or more CFO profiles.
11. The system of claim 8 , wherein said operational parameters and said ambient parameters of said RF receiver and said RF device are selected from the group consisting of: Operating temperatures, ambient temperature, ambient relative humidity, RF channel used for transmitting said RF transmissions, and a connectivity state of one or more wireless transceivers.
12. The system of claim 11 , wherein said operating temperatures include at least some of Central Processing Unit (CPU) temperature, Graphic Processing Unit (GPU) temperature, and battery temperature.
13. The system of claim 8 , wherein at least some of said plurality of operational parameters are initiated in said RF device, based, at least in part, on instructions transmitted from said RF transceiver, wherein said instructions further cause said RF device to transmit at least one training RF transmission during said initiating.
14. The system of claim 8 , wherein said machine learning classifier is further trained based, at least in part, on a training set comprising:
(i) CFO values calculated for at least some of said runtime RF transmission, and
(ii) labels associated with operational parameters and ambient parameters temporally associated with each of said at least some runtime RF transmissions.
15. A computer program product comprising a non-transitory computer-readable storage medium having program instructions embodied therewith, the program instructions executable by at least one hardware processor to:
receive, by a radio frequency (RF) receiver, a plurality of training RF transmissions from an RF device, wherein each of said training RF transmissions is temporally associated with operational parameters and ambient parameters of said RF receiver and said RF device;
at a training stage, train a machine learning classifier based, at least in part, on a training set comprising:
(i) a Carrier Frequency Offset (CFO) value calculated for each of said training RF transmissions, and
(ii) labels associated with said operational parameters and said ambient parameters; and
at an inference stage, apply said trained machine learning classifier to determine whether one or more runtime RF transmissions originate from said RF device.
16. The computer program product of claim 15 , wherein each of said training RF transmissions is labelled with said temporally associated operational parameters and ambient parameters.
17. The computer program product of claim 15 , wherein said training further comprises generating one or more CFO profiles for said RF device, wherein each of said CFO profiles comprises said calculated CFO values, and wherein said determining is based, at least in part, on calculating a distance metric between a measured CFO value of said runtime RF transmissions and said one or more CFO profiles.
18. The computer program product of claim 15 , wherein said operational parameters and said ambient parameters of said RF receiver and said RF device are selected from the group consisting of: Operating temperature of a Central Processing Unit (CPU), operating temperature of a Graphic Processing Unit (GPU) temperature, operating temperature of a battery, ambient temperature, ambient relative humidity, RF channel used for transmitting said RF transmissions, and a connectivity state of one or more wireless transceivers.
19. The computer program product of claim 15 , wherein at least some of said plurality of operational parameters are initiated in said RF device, based, at least in part, on instructions transmitted from said RF transceiver, wherein said instructions further cause said RF device to transmit at least one training RF transmission during said initiating.
20. The computer program product of claim 15 , wherein said machine learning classifier is further trained based, at least in part, on a training set comprising:
(i) CFO values calculated for at least some of said runtime RF transmission, and
(ii) labels associated with operational parameters and ambient parameters temporally associated with each of said at least some runtime RF transmissions.