| Summary: | 碩士 === 國立臺灣大學 === 電信工程學研究所 === 106 === LTE in unlicensed band has been discussed in recent years and named Licensed- Assisted Access (LAA) which uses Listen-Before-Talk (LBT) mechanism. There are lots of works which only focus on system simulation, but few of them concern about analytical model. In this thesis, we probe into the measure of how to build an analytical model to evaluate the performance of LAA with Hybrid Automatic Repeat reQuest (HARQ) feedback. Therefore, we create a proposed Markov chain model and use it to help us to evaluate the performance of LAA with HARQ feed- back. Besides, we re-define some definition of probability to measure the accurate performance. In addition, because each LAA evolved NodeB can serve many users at the same time and some transmission would not be successful, it has to use new way to decide whether total transmission is successful or not and how to set the contention window size (CWS). According to 3GPP document, the CWS updated rule of LAA should be based on HARQ feedback. The CWS is increased if at least 80% of the HARQ-ACK feedback values for the first subframe of a down- link burst are NACK. The present papers still do not consider HARQ feedback in analytical model. Therefore, We take HARQ feedback into consideration in our proposed LAA Markov chain model. Then, we use MATLAB to simulate LAA protocol based on the flowchart in standard. We use MATLAB simulation results to verify our LAA Markov chain model with HARQ feedback. Our LAA Markov chain model can get higher than 70% accuracy to show the LAA performance in most situation which is better than other model. Moreover, we also show the effect of HARQ feedback. The difference between value with HARQ feedback and value without HARQ feedback is small for utility and throughput (smaller than 1%), but it is much large for percentage of dropped packet (90%). In the end, we discuss the relationship between parameter of HARQ feedback.
|
|---|
