الفهرس 
Localization enhancement and particle swarm optimization for resource allocation in visible light communications
SummaryOnly 14 pages are availabe for public view
 |  | from | 102 |  |  |
| | | from | 102 | | |
Abstract
Visible Light Communications (VLC) systems are considered a promising turning point and an alternative to indoor Radio Frequency (RF) communications systems. In this thesis, enhancement algorithms to VLC systems are studied; Re-generations for Joint Optimization Algorithms (JOA) and Separate Optimization Algorithms (SOA) are developed; as the building blocks for hybrid VLC/RF systems, hence applying network load balancing to the RF access point in case of deteriorated transmission data rates to the mobile users. For the later mentioned algorithms, JOA optimizes Access Point Assignment (APA) and Resource Allocation (RA), whereas SOA marginally optimizes allocated time RA. Moreover, this thesis work integrates Particle Swarm Optimization (PSO) into the SOA resource allocation process; to optimize the time-resource allocated to each mobile user to maximize one or more of the system performance metrics: Fairness, Network Utility, and the Dual metric. Five proposed algorithms are introduced: SOA with PSO; a Dynamically adaptive PSO; three Dynamically Adaptive PSO each of Fairness, Network Utility, and Dual metric objective functions, respectively. Results show that SOA with PSO achieves throughput enhancements despite experiencing crashes due to system complexity. Accordingly, the Dynamically Adaptive PSO solves prior issues, in addition to achieving enhancements for different metrics. For instance, the Dynamically Adaptive PSO achieves better fairness by almost 1.5 times that of SOA, meanwhile, the Dynamically Adaptive PSO with Fairness objective function has an absolute edge for fairness. Moreover, the Dynamically Adaptive PSO with Network Utility objective function achieves enhances the total system throughput by about 20% of that achieved by SOA. Furthermore, the Dynamically Adaptive PSO with Dual objective function has supremum both Fairness and Dual metric. It enhances fairness by an order of magnitude and achieves a better Dual metric by almost double that of SOA.
In addition, localization error enhancement is achieved in this work, where parameter relaxation is introduced to the Received Signal Strength (RSS)-based localization system of equations. The location error enhancements include three points: Localization performance being enhanced by four times than that using Artificial Neural Networks (ANN); Reducing the number of iterations for solving the location system of equations; Processing time reduction for solving the system of equations.