Advantage of Open Architecture for Modular Mobile Robotic Platform in Research of Complex Control Algorithms Abdulah Akšamović, Samim Konjicija, Bakir Lačević, Jasmin Velagić and Branislava Peruničić University of Sarajevo, Faculty of Electrical Engineering Zmaja od Bosne bb, 71000 Sarajevo, Bosnia and Herzegovina E-mail: nimsaj @bih.net.ba, samim.konjicija@etf.unsa.ba, bakir.lacevic@etf.unsa.ba, jasmin.velagic@etf.unsa.ba, branislava.perunicic@etf.unsa.ba Abstract – This paper presents the open hierarchical architecture of mobile robot realized at the Faculty of Electrical Engineering, University of Sarajevo. The basic idea during development was to enable simple implementation of algorithms for control of mobile robot, without large interventions on hardware, and to avoid complex software implementation of algorithms. MATLAB was used as basic development tool at the highest hierarhical level, considering its flexibility, as well as numerous high-level functions offered by MATLAB. As an illustration, the results achieved in development of fuzzy algorithm for navigation of mobile robot in unknown environment with obstacle avoidance, and the simplicity of its interactive modification will be demonstrated. I.INTRODUCTION Intensive development of mobile robotics was propelled, among other disciplines, by advancements in microelectronics, embedded systems and simulation tools. Each of the mentioned disciplines achieved considerable level [1, 2], and their development is still in progress. The wide scientific community spends resources in order to acquire applicable results in the area of mobile robotics. The expectations of effects of research in the field of mobile robotics for the human well-being are large. Most of universities in the world teach courses which are directly or indirectly related to the topics of robotics. This process shows noticeable disproportion between the development of simulation models, and the development of prototype solutions. The experimental development, due to its complexity and the required resources, is inferior to simulation modes, what is also the consequence of the fact that simulation tools like MATLAB are widely used. The development of physical model is much more expensive and hard job, which requires specific knowledge, application of various technologies, and more time to realize. That makes each attempt of realization of physical models of mobile robotic systems to be especially challenging. The knowledge necessary to realize such a prototype are usable in much industrial applications (microcomputer- based applications for special purposes, embedded systems, digital control systems, real-time systems, mathematical modeling and computer simulation, control algorithms, software development, data acquisition and processing, etc.). The basic goal of development of mobile robot presented in this paper was: 'To realize hardware and software platform of mobile robot which is open, modular, compatible to the existing knowledge and available technology, as well as easy to use for experimentation and education'. Modularity and openness offer possibility for further development and improvement of capabilities of the realized platform, as well as concurrent development of its segments by more development teams. On the other hand, the use of MATLAB as tool for development of algorithms, offers the possibility for participation of more researchers, especially students. The main goal for the first phase of the project was to realize the following task: 'Enable the direct use of MATLAB for implementation of algorithms for the prototype platform'. This paper briefly presents the hardware of the realized prototype at the end of the first phase of the project, and to describe the implementation of fuzzy algorithm for motion control of mobile robot in unknown environment. II. DESCRIPTION OF THE PROTOTYPE AT THE PHYSICAL LEVEL The prototype represents mobile platform which moves in two dimensional space, at the flat horizontal surface. The platform is driven by two independent wheels. The third wheel is passive, and it is not controllable. The active wheels are actuated by two DC motors. The geometry of the platform is depicted on Fig. 1. Figure 1. Platform geometry (bird's eye view) The flat metal boards, which are fixed at the vertical support bars, carry physical components of the system. In this first phase, three such metal boards are mounted. The first board carries DC motors (from lower side), power and control electronics for motors, IR sensors with controller active wheels castor wheel