In this thesis, I will be presenting the specifics of developing and designing a mobile
robot platform from the ground up.
Before this, I will review a couple of readily available and popular mobile robot sys-
tems. As opposed to their centralised monolithic architecture, I will propose a different
approach, a decentralised one.
To demonstrate this architectural concept, I have designed a rover consisting of three dis-
tinct modules and a chassis. The modules work autonomously and separately. I designed
the motor controllers first, which achieve sensorless closed-loop speed control based on
back-EMF. To orchestrate these controllers, a central processing unit is needed with
the ability to measure its own attitude and heading to support an eventual autonomous
driving application. For the robot to move freely in its environment, these modules have
to be battery powered, therefore the third module is responsible for providing power
and also for protecting and charging the batteries. In my thesis, I will demonstrate the
conceptual, schematic and printed circuit board designs of each module. To house these
devices, a chassis was designed utilizing the rocker-bogie suspension system used on the
Mars rovers. This structure consists of 3D printed parts, which are assembled and held
together with off-the-shelf tighteners.
To utilize and demonstrate the capabilities of the mobile robot prototype, I wrote
firmwares for the central unit and the motor controllers as well as a high-level control
application which handles the correct Ackermann-steering and position estimation of the
rover. This application was implemented in Simulink and included into the software
framework via code generation.
However, the main focus of this work was to learn and acquire new skills and knowledge
through the project.