RF Signal Classification with Synthetic Training Data and its Real-World Performance
This addresses the challenge of limited real RF data for signal classification, offering a practical solution for radio monitoring and communication systems, though it is incremental in optimizing synthetic data design.
The paper tackled the problem of neural networks trained on synthetic RF data performing poorly in real-world applications by investigating which signal impairments to include in synthetic datasets, achieving up to 95% accuracy in classifying 20 radio signal types using only synthetic training data.
Neural nets are a powerful method for the classification of radio signals in the electromagnetic spectrum. These neural nets are often trained with synthetically generated data due to the lack of diverse and plentiful real RF data. However, it is often unclear how neural nets trained on synthetic data perform in real-world applications. This paper investigates the impact of different RF signal impairments (such as phase, frequency and sample rate offsets, receiver filters, noise and channel models) modeled in synthetic training data with respect to the real-world performance. For that purpose, this paper trains neural nets with various synthetic training datasets with different signal impairments. After training, the neural nets are evaluated against real-world RF data collected by a software defined radio receiver in the field. This approach reveals which modeled signal impairments should be included in carefully designed synthetic datasets. The investigated showcase example can classify RF signals into one of 20 different radio signal types from the shortwave bands. It achieves an accuracy of up to 95 % in real-world operation by using carefully designed synthetic training data only.