For networks like Delay-Tolerant Networks (DTNs), ensuring limited and occasional communication opportunities are utilized effectively over     the unpredictable channels, disconnection periods, and varying conditions is an imperative aspect. Existing DTN routing protocols, which are mostly based on network-layer metrics like user mobility and social ties, do not at all take into account the quality of transmission resources in terms of physical-layer channel quality and, thus, suffer from low efficiency. A new predictive framework, Finite-State Markov Channel Predictive Transmission Optimizer (FSMCPTO), is proposed in this paper, utilizing a finite automaton for modelling and predicting the dynamic behaviour of wireless links.
The proposed model classifies channel quality according to a discrete set of quality states based on Signal-to-Noise Ratio (SNR) thresholds; thus, a DTN node can take an intelligent and forward-thinking decision whether to transmit a packet now or to delay it in anticipation of better channel conditions. The physical layer awareness incorporated into the store-carry-and forward paradigm fundamentally changes forwarding decisions from being a topological to an optimization problem across layers. Simulation results validating the FSMC-PTO framework have been carried out extensively; indeed, while improving the packet delivery ratio significantly, it reduces the network overhead as against the so-called traditional DTN routing methods which completely ignore the channel.

Wireless Channel Modelling,Finite Automata,Finite-State,Markov Channel (FSMC),Delay-Tolerant Networks (DTN),Opportunistic Communication,Adaptive Transmission