Available at: https://digitalcommons.calpoly.edu/theses/3115
Date of Award
6-2025
Degree Name
MS in Electrical Engineering
Department/Program
Electrical Engineering
College
College of Engineering
Advisor
Payam Nayeri
Advisor Department
Electrical Engineering
Advisor College
College of Engineering
Abstract
As autonomous vehicles continue to evolve, reliable and efficient real-time communication between vehicles is essential for safety and performance. This thesis explores Streamlined Intelligence: Resource Efficient machine learning for 5G NR V2V Channel Equalization, focusing on lightweight random forest decision tree models to address the challenges of channel equalization in 5G New Radio (NR) vehicle to vehicle (V2V) systems. Using orthogonal frequency division multiplexing (OFDM) with QPSK modulation, the study simulates data transmission in nonlinear channels characterized by obstructions, Doppler shifts, and fading. Decision trees are proposed as a computationally efficient alternative to other machine learning methods while being compared to traditional methods such as minimum mean square error (MMSE) equalizers. Using MATLAB, the performance of these models is evaluated on the basis of bit error rate (BER), training time, and delay under varying channel conditions. The random forest equalizer BER outperforms MMSE in nonlinear channel conditions. However, the random forest adds prediction time to the system and a vast amount of training time. This work aims to demonstrate the potential of resource-efficient machine learning in achieving high-performance channel equalization, paving the way for scalable and effective communication in next-generation autonomous systems.