Autonomous Navigation of Mobile Robot Using Modular Architecture for Unstructured Environment

Abstract

This article proposes a solution for autonomous navigation of mobile robot based on distributed control architecture. In this architecture, each stage of the algorithm is divided into separate software modules capable of interfacing to each other to obtain an effective global solution. The work starts with selection of suitable sensors depending on their requirement for the purpose and for the present work a stereo vision module and a laser range finder are used. These sensors are integrated with the robot controller via Ethernet/USB and the sensory feedbacks are used to control and navigate the robot. Using the architecture, an algorithm has been developed and implemented to intelligently avoid dynamic obstacles and optimally re-planning the path to reach the target location. The algorithm has been successfully tested with a Summit_XL mobile robot. The thesis describing the present research work is divided into eight chapters. The subject of the topic its contextual relevance and the related matters including the objectives of the work are presented in Chapter 1. The reviews on several diverse streams of literature on different issues of the topic such as autonomous navigation using various combinations of sensors networks, SLAM, obstacle detection and avoidance etc. are presented in Chapter 2. In Chapter 3, selected methodologies are explained. Chapter 4 presents the detail description of the sensors, automobile platform and software tools used to implement the developed methodology. In Chapter 5, detail view of the experimental setup is provided. Procedures and parametric evaluations are given in chapter 6. Successful indoor tests results are described in chapter 7. Finally, Chapter 8 presents the conclusion and future scope of the research work.