Proposed joint propagation and reinforcement learning-based television white space ledger

• Main Authors: Pakzad, Armie E., Pakzad, Abbas Ali, Materum, Lawrence
• Format: text
• Published: Animo Repository 2020
• Subjects: Radio wave propagation; Television broadcasting; Reinforcement learning; Electrical and Electronics
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/436
• Institution: De La Salle University

Television white spaces (TVWSs) are vacant television (TV) channels allocated to TV broadcasting. The usability of TVWS can be classified in terms of frequency, time, and location. One way of implementing TVWS communications is by using the TVWS database (TVWSDB). The purpose of this database is to secure the primary users (PUs) from interference from secondary users (SUs). Existing TVWSDBs do not have a prediction feature that provides short-term, medium-term, and long-term forecast data for secondary TVWS users, that could be useful for government and industry stakeholders. This paper proposes an improved TVWSDB that incorporates the reinforcement learning (RL) technique providing short-term, medium-term, and long-term forecast data on the available channels at a given location, time, and frequency for secondary TVWS users. RL is to be used to provide the prediction feature of the TVWSDB. The prediction feature is seen to lessen the number of queries per transaction instance and the search duration for the availability of channels would be lessened and is seen to be beneficial to TVWS users. © 2020, World Academy of Research in Science and Engineering. All rights reserved.