Autonomous robot navigation in a populated environment

Abstract

One of the fields that is greatly benefitting from the advances in technology related to electronics, computing capabilities and portable electrical power storage is robotics. These advances in technology made robots more financially accessible to industry and the general public. The applications of robotics are diverse and range from the use of drones by military and intelligence services for surveillance and air attacks, to the use of robotic systems in hospitals in what is known as robot-assisted surgery. This dissertation focuses on the application of robotics on the ground, and aims to solve the problem of autonomous navigation of a wheeled mobile robot in a populated environment that is also cluttered with both static and dynamic obstacles. The environment in which this autonomous navigation takes place is known a priori by the robot through a preloaded map. Therefore, this dissertation encompasses the analysis of algorithms that can be used for map construction, the methods employed for the robot to be able to localize itself within the map, and the actual path planning and obstacle avoidance algorithms that are required for the robot to find its way to a goal destination autonomously, without colliding with obstacles and people along its course. In order to design such system, the problem of Simultaneous Localization and Mapping (SLAM) was tackled in the mapping stage. The Adaptive Monte Carlo Localization (AMCL) approach was then used to localize the pose of the robot within the preloaded map. Finally, Dijkstra’s algorithm was used in the path planning stage and was coupled with a local planner, the Dynamic Window Approach (DWA), for dynamic obstacle avoidance. The robot used in this project is a commercial research robot named PowerBotTM which was developed by Adept MobileRobots. This robot was installed with the Robot Operating System (ROS), which is an open-source, meta-operating system for robots, and it contains a set of tools that facilitate the development of robotic applications. The performance of the complete autonomous navigation system was tested and evaluated using PowerBot and the results achieved show a good level of performance. Robots with the capabilities mentioned above find applications as delivery robots in urban environments and warehouses, robotic guides in museums and city centres, and possibly as emergency guide/rescue robots. PowerBot can also serve as the mobile base of a robotic arm which provides added functionality.