| MS Seminar

Name of the Speaker: Ms. BHUKYA KALYANI (EE21S124)
Guide: Dr. Radha Krishna Ganti
Venue: ESB-244 (Seminar Hall)
Online meeting link: http://meet.google.com/wmu-mfdd-zbk
Date/Time: 9th July 2025 (Wednesday), 2:00 PM
Title: Precoder Implementation and Optimization in 5G NR Massive MIMO Radio.

Abstract :

Precoding/Beamforming plays a pivotal role in enabling the high throughput and spatial multiplexing capabilities of 5G New Radio (NR) Massive MIMO systems. However, implementing precoding in real-time systems poses significant challenges due to the computational complexity of large matrix-vector multiplications and the strict timing constraints of the 5G NR physical layer.

In this work, a hardware-efficient architecture for precoding has been designed and optimized to support various antenna configurations and user layers. A codebook-based approach has been adopted, and fixed-point matrix-vector multiplication has been implemented using Karatsuba’s technique to reduce the number of real multipliers. To facilitate continuous data flow and parallel processing, ping-pong buffering has been employed for efficient storage and retrieval of precoding matrices. Additionally, a caching strategy has been proposed for CSI-based beamforming, allowing frequently used CSI matrices to be stored and reused, thereby eliminating redundant computations.

Through these techniques, low-latency and resource-efficient precoding and beamforming have been achieved, with scalability across different system configurations. The broader objective is to ensure that physical layer algorithms can be effectively translated into deployable hardware modules suitable for real-time 5G transmission.