Autonomous 2D Mapping of Interior Spaces Utilizing LIDAR Data Used for Asia Pacific College RamiBot: Concierge Robot

RamiBot is one of the current projects of the researchers at Asia Pacific College (APC). Where, this project will be used as a type of robotic concierge which will be designated, and is needed to always be present, at the APC Lobby for better accessibility for everyone. As such, the robot needs a type of system or function that could help it to stay inside the lobby and define its area limits. One function that the researchers have created to solve this issue is a type of 2D LIDAR Mapping system, used as a way create a 2D map of the lobby for RamiBot to define its boundaries. Where, it uses Garmin LIDAR-Lite V3 as its main LIDAR sensor; and the PyGame Python library to create the 2D map itself. Such a system also utilizes the DBSCAN algorithm in order to filter out any outliers in the data, for the system to create a more accurate map of the lobby. After testing such a system, the researchers were able to determine the system's accuracy and precision, together with the effectiveness of the filtering technique used to remove outliers in the data set. With this, they were able to define that the accuracy of the system was adequate, as it had a Mean Absolute Error of 7.4 cm, Root Mean Squared Error of 8.7 cm, and percent accuracy of 94.52%. While its precision was very satisfactory as it was found to be 96% to 98% precise, due to its Structural Similarity Index being computed to be from 0.96 to 0.98. Lastly, through visual inspection, the researchers have observed that the DBSCAN algorithm was able to satisfy its function of being able to filter or remove the outliers in the data. Although, some improvements are needed as some of the features of the area were either thinned out or removed by the filter. Although most of the features were still present.