With the flourishing growth of Internet-of-Things (IoT) applications, cellular-assisted IoT networks are promising to support the ever-increasing wireless traffic demands generated by various kinds of IoT devices. As one of the emerging technologies in the next-generation cellular systems, small cells, or so-called 'femtocells' in indoor environments, are expected to be densely deployed for the ubiquitous coverage and capacity enhancement in a cost-effective and energy-efficient way. However, unplanned femtocell deployment and complex interior layouts of buildings may lead to severe coverage and capacity problems and even possible negative impacts on the dense deployment. In this paper, we aim at developing a data-driven indoor diagnostic framework for fault detection in cellular-assisted IoT networks. The framework utilizes machine learning techniques to analyze crowd-sourced measurements uploaded from diverse end devices. Then, some well-trained prediction models are used to construct a global view of a radio environment, namely, radio environment map (REM), for diagnosis and management purposes. Moreover, REMs enable operators to detect existing coverage and capacity problems at any interested location. To verify the feasibility of indoor diagnoses, we collect a real trace data from an indoor testbed and then conduct a series of experiments to evaluate the performance of the machine-learning algorithms in terms of the accuracy, the time complexity, and the sensitivity to data volumes. The experimental results provide an insightful guideline to the indoor deployment of small cells in cellular-based IoT networks.
All Science Journal Classification (ASJC) codes
- Signal Processing
- Information Systems
- Hardware and Architecture
- Computer Science Applications
- Computer Networks and Communications