XPRT: Extracting Priors for Signal Reconstruction

As wireless technology advances towards the sixth generation (6G), new and emerging applications require everincreasing data rates and capacity to match the demand. Generally, wireless receivers are developed with the theoretical background of an unbiased estimator that can maximize the posterior likelihood of the transmitted message (modulated symbols) based on the received signal impaired by the wireless channel. This method however is limited by theoretical channel capacity. In this work, we propose XPRT that can efficiently extract spatiotemporally invariant priors of the waveform that can be made available as common knowledge to all the receivers at the design stage without increasing any overhead during communication. This prior increases the probability of correct detection of transmit symbols -in a simpler term, it increases effective SNR. Consequently, higher modulation orders can be used without compromising the receiver accuracy, hence increasing throughput. We analyzed the equivalence of XPRT to an ideal receiver along with defining theoretical boundaries. Finally, we implemented XPRT for different waveforms over simulation with AWGN and multipath faded channel where we observed up to 2.2 dB improvement in received SNR at a reference BER of 10⁻².