The project's goal was to program a robot to be placed in an unknown environment and explore the entirety of the environment while running a version of online SLAM. Simultaneous Localization And Mapping The Turtlebot was equipped with a LiDAR sensor and encoders on the wheels allowing it to estimate its displacement from reading to reading. The robot would identify frontiers to search and would continue to plan and follow paths to the frontiers until there were no more frontiers to explore. Utilizing The A* algorithm the robot would quickly plan a path from its current position to the target, searching quickly prioritizing nodes that travel towards the target while avoiding obstacles. The searchable space was shrunk adjacent too walls to guarantee safety as the robot navigates the environment. Furthermore a gaussian kernel was applied to the map and used as a heuristic in the path planning. This worked as fast lookup table for the heuristic values and updated when new map information was received. This was essential because during the path planning process the panner will look at hundreds of points and precise methods take too much time. The robot would visit the frontiers based on its proximity to the robot, the length of the frontier, the density of the frontiers, and the distance to the start. These heuristics were chosen to prioritize searching the maze as fast as posiable and prioritize gaining the most information quickly. Using ROS Robot Operating System over the network the turtlebot was able to communicate with a laptop to handel all of the algorithms. It implements a pure pursuit controller to follow the paths, ensuring the robot follows a continuous trajectory. A Pure pursuit controller works by looking ahead on the path a set horizon then moving towards that goal. The robot runs obstacle avoidance to ensure safety, especially since the pure pursuit controller has a tendency to round corners by nature. The robot uses an online SLAM library and particle filter to predict the state of the environment and localize it by minimizing the lidar's re-projection error. Once the robot is done mapping it can be relocated to a different part of the maze and utilizing Adaptive Monte Carlo Localization (AMCL) it can localize itself and travel to designated target location.

This Project focuses on advanced navigation techniques along with applying modern tools methods too navigate in an unknown environment. Collaborated in a four-person team utilizing Agile methodologies to distribute and delegate tasks strategically, aligning assignments with the complexity and workload of each task.