Implementation of Slam-based Assistive Robot for Hospital Navigation using ROS2

This paper presents the development of an assistive indoor navigation and docking system designed for autonomous robots operating in healthcare environments, leveraging the Robot Operating System 2 (ROS 2) for modularity and real-time communication. The proposed system integrates Simultaneous Localization and Mapping (SLAM) for real-time environment modeling, Adaptive Monte Carlo Localization (AMCL) for precise localization, and a feedbackbased control mechanism to ensure accurate docking. Sensor fusion techniques combining LiDAR and encoder data enhance the robot’s situational awareness and localization robustness in dynamic indoor spaces. A key contribution of this work is the implementation of an adaptive control strategy that dynamically adjusts the robot’s motion in response to real-time environmental feedback, improving resilience against unexpected obstacles and variable surface conditions. The system’s architecture supports seamless integration of advanced modules, including predictive path planning, AI-driven obstacle avoidance, and cloud-based remote control, enabling scalability and continuous learning. This makes the solution highly applicable in complex environments such as hospitals and logistics facilities. The proposed system achieves 94.2 % docking success, 34.6 s per 10 m navigation time, ±5.2 cm localization accuracy, and 68 % CPU utilization, demonstrating efficient and reliable real-world performance. The results demonstrate the effectiveness of ROS 2 in enabling intelligent and reliable autonomous navigation and docking. The research contributes to the field of robot navigation and intelligent systems by showcasing a scalable and adaptable approach for deploying autonomous service robots in real-world indoor scenarios. Future work will explore multi-robot coordination, energy-efficient navigation algorithms, and platform-independent deployment to expand the system’s applicability in broader robotics and automation domains.