Estimating Capture Probabilities for Complex Topologies in 2D Molecular Communication via Diffusion Channel using Artificial Neural Networks
2021 IEEE International Black Sea Conference on Communications and Networking, BlackSeaCom 2021, Virtual, Bucharest, Romanya, 24 - 28 Mayıs 2021, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/blackseacom52164.2021.9527790
- Basıldığı Şehir: Virtual, Bucharest
- Basıldığı Ülke: Romanya
- Anahtar Kelimeler: artificial neural network, channel impulse response, diffusion, Molecular communication
- Boğaziçi Üniversitesi Adresli: Evet
Özet
Nanonetworking is a way of communication that focuses on the communication between entities where (at least) a component is at a nano-scale. Molecular communication via diffusion (MCvD) is one of the fundamental communication mechanisms that is used in nanonetworks. There are two conventional approaches to model and analyze an MCvD channel: (1) analytical approach, (2) simulation based approach. Unsurprisingly, the analytical approach is the most efficient one. Yet, it is not largely applicable (i.e., provides solutions for only few simple topologies). On the other hand, simulation based solutions are much more applicable to different scenarios. However, it has another major shortcoming, it takes time and requires excessive processing power. In our study, the aim is to estimate the channel impulse response of complex two-dimensional (2D) MCvD channels by using artificial neural networks, in order to improve the overall time-based quality of service compared to simulation based approach. Different modeling approaches are compared to find the best approach in terms of error metrics. Consequently, an artificial neural network, which can estimate the channel coefficients of complex 2D MCvD channel coefficients is obtained.