Optimizing traffic flow remains a central objective in transportation research. With the continuous growth in vehicle numbers, the limited scalability of existing infrastructure, and the inherently nonlinear, dynamic, and stochastic nature of traffic systems, signal control at road junctions has become an increasingly complex control challenge. Traditional signal control methods are predominantly rule-based, designed for deterministic environments, and often fail to adapt to real-time traffic variations and unexpected disruptions. This study proposes a distributed traffic signal control framework built upon a Machine Learning (ML) paradigm utilizing Reinforcement Learning (RL). Owing to their non-model-based architecture and computational efficiency, RL-based methods exhibit strong adaptability to changing traffic conditions and robustness against environmental uncertainties. In the proposed system, each junction is independently managed by an autonomous learning agent capable of interacting with its environment, refining its control policy over time, and making localized, real-time decisions. Simulation results demonstrate the effectiveness of the proposed approach, with vehicle queue lengths and average waiting times reduced by 35% to 66.4% on roads leading to the junctions, compared to conventional rule-based systems. These findings highlight the potential of distributed, learning-enabled control strategies in achieving scalable, adaptive, and efficient urban traffic management.