The primary goal of our project is to leverage machine learning in conjunction with the Frenet-RRT-Cartesian framework to create a more advanced navigation system for urban autonomous vehicles (AVs). This framework facilitates environmentally conscious research and provides optimized routing within dense urban traffic by integrating Frenet coordinates with Rapidly Exploring Random Trees (RRT) in Cartesian space. Our solution incorporates machine learning algorithms to enhance traffic law compliance, collision avoidance, and trajectory optimization, emphasizing safety, reliability, and adherence to regulations. Our approach validates resilience and efficacy through extensive simulations in diverse urban scenarios, demonstrating that our system can adeptly navigate complex cityscapes while observing safety standards. Additionally, our modular design allows for seamless integration with modern AV architectures, enabling rapid deployment in real-world settings. The computational efficiency of our machine learning-enhanced algorithm improves AV responsiveness in dynamic environments. This initiative has the potential to significantly advance AV route planning in urban areas, accelerating the widespread adoption of AV technology. Future research could involve real-time optimization and on-road validation to further enhance the efficiency and safety of urban transportation systems.