Localization provides the fundamental capabilities for the various autonomous mobile robots. It is due to the fact that it provides uncertainty in acting, sensing and environmental factors of the Robot and hence it loses localization. If we lose the Localization, Deep Reinforcement Learning helps to find the patterns in the particles. The patterns can be mixed with the weak classifiers from the particle set and observe the sensors to estimate the localization using boosted learning. In order to provide an efficient navigation, the robots are used to implement the localization strategy effectively. The comparative analysis has been analyzed for various approaches in mobile robots. We propose an extended Kalman filter (EKF) based localization algorithm scheme for robot localization, focusing on its performance and practicability. Our algorithm takes advantage of the EKF’s ability to handle non-linear motion and measurement models. By incorporating these models into the EKF framework, we can effectively estimate the robot’s pose (position and orientation) in real-time. The localization performance and practicability of the developed robot localization algorithm, extended kalman filter based localization algorithm scheme is proposed.