Towards ultra-reliable M2M communication: scheduling policies in fading channels

Abstract

Design of the scheduling algorithm that satisfies the stringent timing and reliability requirements of control systems, while considering the delay and packet errors in the wireless channel, and limited battery resources of the sensors is the main challenge of ultra-reliable machine-to-machine (M42M) control applications. Transmission delay and energy consumption of a sensor node are determined by the transmission power and rate of that sensor node and the concurrently transmitting nodes. Hence, the transmission schedule, and the power and rate parameters of the nodes should be jointly optimized. Joint optimization problem for M2M control applications has been previously formulated and solved by adopting a static channel. in this paper, we extend this framework for fading channels. First, we propose an offline scheduling algorithm that exploits the periodic nature of the transmissions under a priori known channel. then we propose an online scheduling algorithm based on the prediction of the channel information by using previous measurements, scheduling of the nodes using offline algorithm, and modification of the schedule and the transmission parameters of the nodes under the actual channel information as time evolves. Extensive simulations demonstrate the near optimal performance of online and offline scheduling algorithms for different network sizes and densities.