Open-Loop Power Adaptation in Nanosensor Networks for Chemical Reactors

Due to their extremely small form factor, wireless nanosensor networks (WNSNs) have the potential to monitor and control chemical processes right at the molecule level lifting the process efficiency to a level not possible with conventional methods. However, sensor networking within a chemical reactor is challenging due to its time-varying chemical composition, which creates a time-varying radio channel. The nanosensors therefore need to continuously adapt their transmission power according to the chemical composition while maintaining a low overall power budget. We show that this problem can be modeled as a Markov decision process (MDP). However, the MDP solution requires the sensors to know the composition of the reactor at each time instance, which is difficult to realize due to computation and communication constraints at the nanoscale. We therefore choose to use open-loop methods for power adaptation, which allows nanosensors to dynamically adjust their powers based on a policy derived entirely from offline analysis of the expected channel variation over time. Using extensive simulations, we evaluate the performance of the proposed open-loop power adaptation method for improving the efficiency of a widely deployed gas-to-liquid conversion process known as Fischer-Tropsch (FT) synthesis. We find that the proposed method performs close to optimal achieving a three fold improvement in FT efficiency with only a 100 femto watt power consumption on average for the nanosensor communications. In terms of maximum achievable process efficiency, our method outperforms the previously proposed open-loop power adaptation policy by 30% and the nonadaptive power allocation by 52%.