Integrated approach to simultaneous localization and mapping with path planning algorithms for indoor mobile robots

Abstract

This thesis investigates Simultaneous Localization and Mapping (SLAM) problem in integration with path planning algorithm for an indoor mobile robot implemented in computer simulation. The first focus is on SLAM solutions using both parametric and non-parametric Bayesian filtering approach. Comparative simulations then prove pros and cons of each method as well as the feasibility to be used for the proposed integration. The second part of this work studies A* path planning algorithm combined with a proportional controller to make SLAM robot be able to follow planned paths until reaching the goal without collision along the way. Thirdly, the information gain- and minimized traveling cost-based exploration is studied in order to provide the robot the decision making capability for where to go next. In addition, map representations in forms of feature-based and occupancy grid map play important roles in reasoning the environment, which are also taken into consideration. Putting these techniques together, we introduce the whole framework for an indoor autonomous mobile robot and evaluate its effectiveness via simulations.