Robots are widely used in domestic and commercial applications. These applications typically involve robots built for a specific the task, thus leading to efficiency in task completion. However, with the increase in completion efficiency comes an increase in the time required for completion of the whole system. Specific tasks create the need for many different robots, all with differing capabilities, causing an increase in development cost and time needed.
This raises the issue of whether using an off the shelf system with minor modifications can accurately perform the same tasks. If so, more time can be spent on refining the process leading to completion of the task, resulting in less time spent developing the robot. Consequently, less cost in the life cycle of the system leads to less cost for the end user, thus allowing robots to be used for more applications.
This thesis explores using a commercially available robot, Acroname Inc.’s Garcia, to perform mapping and localization tasks. As the robot travels it gathers data about the environment. This data is processed in Matlab and the result of the algorithm is a map. In the creation of the map, mathematical morphology is explored as a means to reduce noise. When the robot has located a corner, Matlab provides the robot with a position estimate correction. This correction allows the robot to better estimate its location resulting in a more accurate map. As the results of this thesis illustrate, with very minor modifications, the robot is capable of accurately performing mapping and localization tasks. The results demonstrate that an off the shelf system is capable of accurately performing tasks for which it was not specifically designed.
