| MS TSA Meeting

Name of the Speaker: Mr. HARIVIGNESH A (EE20S048)
Guide: Prof. K. Giridhar
Online meeting link: meet.google.com/rbz-ycwc-xcj
Date/Time: 21st June 2024 (Friday), 2:00 PM
Title: On reducing the search complexity for PMI selection in Codebook-based 5G NR downlink systems.

Abstract :

To obviate the need for full Channel State Information (CSI) feedback in a Downlink (DL) Multiple Input Multiple Output (MIMO), 5G New Radio (5GNR) employs codebook-based limited feedback techniques. These techniques require the User Equipment (UE) to select the best precoder for downlink transmission from a set of codebook entries and send the corresponding Precoding Matrix Indicator (PMI) as feedback to the Base Station (BS). Various methods have been proposed in the literature to choose the best possible PMI. The Exhaustive Search (ES) method that examines all possible entries of the codebook is the conventional method. However, the computational complexity of searching all codebook entries is significantly high. Also, with the increase in MIMO technologies, the codebook size grows, resulting in even longer search times.

This work presents a low-complexity PMI selection method that leverages the spatial characteristics of the codebook entries. In this method, the UE approximates the transmit beam to estimate the coarse Angle of Departure (AOD), which is then mapped to PMI using a novel PMI-to-angle mapping look-up table (LUT). This coarse estimate is then used to reduce the codebook size, and the final PMI is obtained by searching the reduced codebook. The proposed method significantly reduces the number of codebook entries to be searched, and simulations show that it lowers the search complexity by about 94%, while maintaining the same performance as the conventional exhaustive search scheme.