This research work suggests that a control system architecture that relies on IoT Sensors and ML Optimisation can enhance vehicle decision-making tactical control and real-time navigation. LIDAR, cameras, radar, GPS and others are used in the collection of the environmental information that is required for the proper functioning of the system. Artificial intelligence, more specifically the machine learning and reinforcement learning (RL), is used for path planning, obstacle avoidance and reward based decision making system. Experimental results indicate that the proposed framework outperforms traditional control systems in key metrics: optimal path decreases the time required for travelling by 15%, fuel consumption increases by about 12% and lastly through sensor fusion, obstacle detection becomes more accurate by 25%. The proposed reinforcement learning model shows 30% greater accuracy in decision making, along with better Q-values convergence. Real time responsiveness is improved through sensor data fusion and integration of machine learning, which cuts down the decision latencies by 20%. Edge computing was adopted to relieve the computational requirement of the framework in processing and decision making in real time. These results proved that the integration of IoT and ML is beneficial in improving the performance of autonomous vehicles when applied showing safety, efficiency, and flexibility.