Basestation Switch On/Off Strategy for Energy Efficiency in Ultra-Narrowband IoT Networks

In this work, we examine ultra-narrowband IoT networks in the case of basestation (BS) over-deployment, where we necessitate smartly turning on/off BSs for an energy-efficient network. We consider a dynamic environment and varying network traffic, where the optimal choice of BSs changes frequently. Therefore, we want to often re-evaluate if BSs should be on/off. Addressing this problem, we find a low time-complexity algorithm minimizing the number of active BSs while maintaining a high network coverage, measured by the network's packet decoding probability (PDP). The proposed strategy is based on solving an integer convex optimization problem maximizing an approximate term for PDP using collected signal-to-interference-and-noise ratio (SINR) values and constrains the number of BSs remaining on. Therefore, the comprehensive proposed algorithm includes our development of a model to estimate how many BSs should minimally be on to satisfy our high PDP value, denoted the “coverage constraint.” Our algorithm has a complexity scaling linearly with the BSs on, O(B), far better than the exhaustive search complexity to find the optimal choice of BSs, O(2B), and the greedy algorithm complexity used as an on/off strategy in the static case of cell-free MIMO wireless networks, which has O(B
³
). The proposed strategy's low complexity not only aids in the dynamic case-our proposed algorithm performs closer to the optimal when the number of available BSs is high, which is also when greedy and exhaustive methods become infeasible due to their complexity. As a result, our algorithm is also advantageous specifically for excessive over-deployment of BSs.