Improving Connectivity in Urban IoT-Based Wireless Sensor Networks: A Nonorthogonal Cognitive-Based Power Allocation

Wireless sensor networks (WSNs) have been recently deployed to support various Internet-of-Things (IoT) applications, including sensing in urban environments. Accordingly, several research efforts have been conducted to support the unprecedented massive connectivity requirements, where a massive number of urban sensors are expected to be connected to the Internet. To enable such large-scale urban connectivity, cognitive radio (CR) technology has been identified as an appealing solution. This can be achieved by opportunistically exploiting the under-utilized licensed spectrum. However, most existing CR-based communication protocols were designed under imposing the exclusive-channel occupancy constraint, in which an idle channel cannot be assigned to multiple users at a time. This constraint can significantly limit the number of served sensor devices, which negatively impacts spectral efficiency. This article proposes a batch-based power-controlled nonorthogonal spectrum sharing protocol for CR-enabled WSNs, referred to as the distance-fading factor MAC (DFF-MAC) protocol, aiming to maximize the number of per-channel served transmissions while minimizing the overall transmit power. Specifically, we develop a power-minimization framework, and thus we derive a closed-form expression for the required transmit power for each transmission. This can be achieved by employing a novel multistage power allocation that considers channel gain and interference constraints. Simulation results show that DFF-MAC outperforms the benchmark, namely the exclusive-occupancy protocol, by serving up to three times more transmissions. Furthermore, DFF-MAC achieves a level of performance (in terms of serviced sensors and power consumption) within 4% of the optimal one obtained through exhaustive search while maintaining low complexity.