The Effect of Ambient Temperature On Device Classification Based On Radio Frequency Fingerprint Recognition

Physical layer authentication is an important technique for cybersecurity, especially in military scenarios. Device classification using radio frequency fingerprinting, which based on recognizing device-unique characteristics of the transient waveform observed at beginning a transmission from device, promising method this context. In study, effect ambient temperature performance device RF fingerprinting investigated. The waveforms regions transmissions are recorded as images, and ResNet50 InceptionV3 networks image used to determine devices. devices study belong same brand, model, production date, making problem more difficult than classifying different brands or models. Our results show that high levels accuracy can be attained convolutional neural network models such when test data training collected temperature, whereas suffers values. We provide figures blended model uses taken various A comparison two also provided.