Correlated Source Achievable Information Rate Analysis in Diffusive Channels with Memory

This paper explores the Achievable Information Rate (AIR) of a diffusive Molecular Communication (MC) channel. We consider a fully absorbing receiver that counts the absorbed particles during each symbol time interval (STI) and uses a binary concentration shift keying modulation. The diffusive MC channel experiences inter-symbol interference (ISI) due to the information particles' delayed arrival. We quantify the channel's memory as an integer multiple of the STI. Under a Gaussian approximation of the channel law, we maximize the mutual information at the output of a detector threshold. Two distinct scenarios are considered for the source symbols: (i) independent and (ii) correlated via a first-order Markov process. It is demonstrated that employing a correlated source enables attaining higher constrained capacity (CC). The results indicate that the CC-achieving input distribution is not necessarily uniform. Notably, for short STI (or strong ISI), the maximum AIR is not achieved through equiprobable symbol transmission.