Artificial neural network based misorientation correction in molecular 4x4 MIMO systems

MIMO systems in molecular communication are proposed and exhaustively investigated to increase the data rate. However, the misalignment problem between antennas decreases the performance of the system and has not been investigated adequately. This paper aims to correct the received signal in misaligned and distance-varying molecular 4x4 MIMO systems. By using multiple transmitter antennas and receivers, we can increase the data rate by utilizing higher-order modulation techniques that exploit spatial diversity. For the spatial modulation techniques, if there is a misalignment among the corresponding antennas, the signal quality degrades and decoding errors occur more frequently. Therefore, our main goal is to decode the received signal, while considering and compensating the misalignment of the system. To do so, we first estimate the misalignment along the rotation axis of our system and the distance between transmitter antennas and receivers. Then using these values, we process the received signal to eliminate the effects of misalignment. Additionally, we propose and investigate a method to detect active transmitter antennas that can be utilized for molecular index modulation under misalignment. We compare the performance of our proposed methods with the existing misalignment correction techniques.