UCLA

Low-power wide-area (LPWA) networking is a recent paradigm in wireless networking, intended to support access from a massive number of Internet of Things (IoT) devices. Ultra-narrowband (UNB) LPWA solutions apply the ultra-narrowband transmissions, which enable demodulation at very low received power. In this work, we consider an ultra-narrowband (UNB) network architecture design for uplink transmission of massive number of Internet of Things (IoT) devices over multiple multiplexing bands. An IoT device can randomly choose any of the multiplexing bands and transmit its packet. Due to hardware constraints, a base station (BS) is able to listen to only one multiplexing band. Our main objective is to optimize the BS infrastructure in terms of the placement of the BSs and frequency assignment of BSs to multiplexing bands to maximize the packet decoding probability (PDP). We develop two approaches that adapt to the environment through training. The training is used to learn some key information about the environment in the form of parameters that are used to perform optimal placement and assignment of BSs. The parameters are estimated by counting successful and unsuccessful packet transmissions at the BSs in the network. The two approaches differ in the assumed models of the environment. The first approach is based on a model that the environment is expected to adhere to, while the second approach is model-free which makes it applicable to any environment. The benefit of the model-based approach is a lower training complexity. The simulation results show that our proposed approaches to band assignment and BS placement offer significant improvement in PDP over baseline random placement and assignment approaches, and in simulated cases closely matches the theoretical best performance.