Error avoidance in wireless networks using link state history

We address the problem of time varying connectivity, as would arise in a wireless communication system with channels occasionally becoming more error prone. Such channels have the property that they can only be used during intervals of variable duration as the devices´ connectivities with a centralized controller change unpredictably with time. We propose a link state history based scheme in which the centralized controller, (a master for a master slave kind of system) tries to identify at each scheduling instant, the devices seeing bad connectivity. We demonstrate the advantage of the scheme on top of a master driven frequency hopping system derived from the Bluetooth specification. We also analyze the scheme using Markov chains. Numerical results from analysis show the performance of the scheme. We show that, with the right tuning of parameters, we can achieve high accuracy in identifying the good and the bad periods of the channels. Simulation results are also shown indicating improvement in throughput and goodput