Autonomous materials handling robot for reconfigurable manufacturing systems.
Butler, Louwrens Johannes.
MetadataShow full item record
The concept of mass producing custom products, though extremely beneficial to the commercial, and retail industries, does come with some limitations. One of these is the occurrence of bottlenecks in the materials handling systems associated with reconfigurable manufacturing systems tasked with achieving the goal of mass customisation manufacturing. This specific problem requires the development of an intervention system for rerouting parts and materials waiting in line, around bottlenecks and/or work flow disruptions, to alternative destinations. Mobile robots can be used for the resolution of bottlenecks, and similar disruptions in work flow, in these situations. Embedding autonomy into mobile robots in a manufacturing environment, releases the higher level production management systems from routing of parts and materials. The principle of the inverted pendulum has recently become popular in mobile robotics applications, and is being implemented in research projects around the world. The use of this principle produces a two-wheeled mobile robot that is able to actively stabilise itself while in operation. The dissertation is focused on the research, design, assembly, testing and validation of a two-wheeled autonomous materials handling robot for application in reconfigurable manufacturing systems. This robot should be dynamically or statically stable during different phases of operation. The mechatronic engineering approach of system integration has been used in this project in order to produce a more reliable robotic system. The application of the inverted pendulum principle requires that a suitable control strategy be formulated. It also necessetates the ues of sensors to track the state of the robot. Control engineering theory was used to develop an optimal control strategy that is robust enough to cope with varying payload characteristics. The Kalman filter is employed as state estimation measure to improve sensor data. For a mobile robot to be deemed autonomous, one of the requirements is that the robot should be able to navigate through its environment without colliding with obstacles in its path, and without human intervention. A navigation system has been designed, through field specific research, to enable this. The robot is also required to communicate with remote computers housing production management systems as well as with mobile robots that form part of the same materials handling system. Performance analysis and testing proves the feasibility of a mobile robot system.