Information-Oriented Opportunistic Mobile Crowdsensing for Enhanced Heterogeneous Small Cell Network Operation
Abstract
Accurate onsite wireless environment information is crucial for effective heterogeneous small cell network (HetNet) operations, especially due to strong signal attenuation in millimeter wave (mmWave) compromising communication quality. Opportunistic mobile crowdsensing (OCS) offers a cost-effective solution by opportunistically recruiting ubiquitous active mobile devices for onsite wireless environment sensing. This thesis introduces two tailored OCS scheduling mechanisms: optimizing information quality and enhancing communication throughput.
The first study addresses the complex challenge of collecting real-time high-quality spectrum information for effective UAV network operation. Existing OCS approaches, due to a lack of sensing and transmission coordination, failed to gather spectrum information with sufficient quality in the data acquisition process, leading to deteriorated UAV network operation performance. To address this, we propose a novel dynamic scheduling mechanism that maximizes the quality of information (QoI) in the location-aware OCS scheme. Our proposed approach jointly allocates sensing and communication resources to maximize the collected information quality and reduce the OCS communication overhead, marking an improvement over existing approaches.
The subsequent study tackles the issue of onsite real-time interference information collection for successful mmWave-empowered SC network deployment. Conventional OCS methods follow a routine schedule for updating interference levels for each user at fixed time intervals, leading to unnecessary sensing overhead and degrading system performance. To resolve this, we propose a novel low-overhead situation-dependent OCS scheme that maximizes the system throughput in the indoor location-aware SC communication system by leveraging the interference spatial correlation to reduce the sensing frequency. Our proposed mechanism jointly optimizes the sensing scheduling and time allocation to maximize the total throughput in the SC network system, offering a more advanced solution than conventional approaches.
All algorithms presented in this thesis are demonstrated using MATLAB simulations that mirror real-world conditions and are compared against existing schemes.