Adaptive Monte Carlo Localization Combines Dijkstra's and DWA for Mobile Robots to Navigate Through Predefined Points Based on ROS

This paper presents an Adaptive Monte Carlo Localization (AMCL) approach for robot localization, which enables obstacle avoidance and navigation along predefined waypoints in an indoor environment. The model is implemented using the Robot Operating System (ROS) along with the Gazebo and RViz simulation environments. The AMCL algorithm estimates the robot's pose using a Particle Filter, representing it as a distribution of particles where each particle corresponds to a possible robot position. This allows the algorithm to estimate the robot's 2 D pose within a known environment. Additionally, the global path planning algorithm Dijkstra's is integrated to generate the robot's trajectory from the start to the target point. The Dijkstra's algorithm works in conjunction with the Dynamic Window Approach (DWA) for local path planning, ensuring the robot moves without colliding with obstacles in the static map. The simulation results in Gazebo and RViz demonstrate the robot's ability to follow the planned trajectory effectively while avoiding obstacles.