Extended Kalman Filter Line-Based Simultaneous Localization and Mapping for Autonomous Robotics

Technological progress requires humans to perform hazardous tasks. This led tothe creation of the Occupational Safety and Health Administration (OSHA). OSHA classifies as extremely dangerous occupations related to search and rescue, leak detection in pipes, and landmine detection and recovery. Therefore, it is imperative to limit human involvement in these activities for safety reasons. Autonomous mapping of environments considered potentially life-threatening for humans is one of those areas that require attention and further development. To investigate solutions to these problems in a wide variety of resource environments, this work explores the state of the art of combined Extended Kalman Filter - Simultaneous Localization and Mapping (EKF-SLAM) algorithms using a microcontroller robot platform equipped with a Light Detection and Ranging (LiDAR) sensor. Specifically, we develop and implement a type of EKF-SLAM that utilizes lines instead of points (Line-EKF-SLAM), with its corresponding line specific data association. The advantages over Point-EKF-SLAM are discussed in depth. With Line-EKF-SLAM, we demonstrate how to map an environment. The results of this thesis are restricted to 2-Dimensional environments. A full software implementation is provided as well as its deployment on a real autonomous vehicle. The robot will autonomously traverse a hallway while identifying the number of variable length line segments present in the environment. The numbered line segments are updated and plotted in real-time over a wireless connection to show the utility and versatility of the technique for the identification of walls in the areas of interest. Through this implementation, we hope to improve the lives of people affected by these safety issues