RESUMO
A deep learning channel estimation scheme in orthogonal frequency division multiplexing for X-ray communication (XCOM) is studied. The scheme uses simulated and detected data obtained with different working parameters and numbers of pilots as training and testing data, respectively, for the deep neural network (DNN) model. The bit-error-rate performance of the DNN model under various system operating parameters, numbers of pilot sequences, and channel obstructions is investigated separately. Experiment results showed that the deep-learning-based approach can address the distortion of the air-scintillator channel for XCOM, giving a performance comparable to those of least-squares and minimum-mean-square error estimation methods.
Assuntos
Aprendizado Profundo , Comunicação , Redes Neurais de Computação , Raios XRESUMO
Carbon nanotube (CNT) field-emission x-ray source has great potential in x-ray communication (XCOM) because of its controllable emission and instantaneous response. A novel voltage loading mode was proposed in this work to achieve high-frequency pulse x-ray emission. The characteristics of cathode current and pulse x-ray versus voltage, frequency, and pulse amplitude were studied, and XCOM data transmission experiment was carried out. Results showed that the CNT cold cathode x-ray source, as a communication signal source, could work in 1.05 MHz pulse emission frequency. When the grid voltage was higher than 470 V, the pulse x-ray waveform amplitude achieved peak, and the shape exhibited a pseudo square wave. The duty cycle of the x-ray waveform exceeded 50%, reaching 56% when the pulse frequency reached 1 MHz. In the XCOM data transmission experiment, the pulsed x-ray waveform was well consistent with the loading data signal voltage waveform under different pulse-emission frequencies. This work realized the x-ray high-frequency pulse emission of CNT cold cathode x-ray source and lays a foundation for the development and application of CNT cold cathode x-ray source in XCOM.