METAHEURISTIC OPTIMIZATION FOR UPLINK TRANSMISSION IN CELL FREE UAV MASSIVE MIMO COMMUNICATION SYSTEM

• Main Author: Hidayat, Benny
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84837
• Institution: Institut Teknologi Bandung

The use of an Unmanned Aerial Vehicle (UAV) as a base station (BS) is expected to overcome coverage problems and have better flexibility than terrestrial BS. The limitations of terrestrial systems provide potential for the development of cell-free based UAV communication systems. A cell-free communication system using a UAV is used as a model for the system to be developed. High Altitude Platform Station (HAPS) is used as a backhaul that connects UAVs. The transmission system between the UAV and the user uses sub-6 GHz frequencies and HAPS uses sub-THz to avoid interference. The flexibility of the UAV offers the possibility to optimize the position and power allocation of the UAV in serving communications for each user. Metaheuristic optimization methods such as Particle Swarm Optimization (PSO), Flower Pollination Algorithm (FPA), and Simulated Annealing (SA) are used as optimization techniques to achieve optimal and more even data rates between each user. The results of optimization carried out with 2 to 12 iterations with PSO showed an increase in the average data rate reaching 10,723 bps/Hz in 6,928 seconds to 11,591 bps/Hz in 34,571 seconds. FPA achieved an increase in average data rate from 10,652 bps/Hz to 11,387 bps/Hz while requiring computing time from 4.95 seconds to 17.35 seconds. Meanwhile, SA has the fastest optimization in carrying out the computing process with a time of 3.8276 seconds to achieve an average data rate from 10,306 bps/Hz to 11,172 bps/Hz in 11,168 seconds. The metaheuristic optimization method was compared to the CVX method with the results showing that the metaheuristic method had a faster computing time and a better increase in data rate, while the CVX method produced a more stable and optimal data rate between users even though it required longer computing time.