Channel Estimation via Thermal Noises

Employing thermal noise for conveying information is a potential solution to achieve an extremely low power consumption. In this letter, we study the pilot-assisted channel estimation tailored for wireless communications utilizing thermal noise modulation. Specifically, the Cramér-Rao bound (CRB) for the channel estimation is first derived to reveal the relationship between the estimation accuracy and the pilot bits and length. Subsequently, we develop a modified maximum likelihood (ML) channel estimator in a closed form, in order to circumvent the invalidity issue of the conventional ML solution. Simulation results show that the mean squared error (MSE) of the proposed channel estimator is very close to the CRB. Additionally, given a few pilot bits, the proposed method achieves nearly the same bit error rate (BER) performance as the perfect channel state information (CSI) case.