Guidance of mobile robot through the maze

Tema ovog rada je vođenje mobilnog robota kroz labirint. Kako bi se moglo voditi robota kroz labirint prvo je potrebno isplanirati putanju, potom je potrebno napraviti sustav kontrole lokacije robota i prepreka te napraviti sustav kontrole gibanja. Planiranje putanje mobilnog robota se izvodi pomoću podražajnog digitalnog reakcijsko-difuzijskog sustava (DRDS). Modificiranjem postojećeg podražajnog DRDS-a je postignuta stabilnost sustava te mogućnost planiranja putanje i u kompliciranim prostorima. Nakon planiranja putanje pomoću DRDS-a prikazano je kako generirati referentnu trajektoriju analizom gibanja robota planiranom putanjom. Prilikom generiranja trajektorije uzima se način gibanja robota. Potom je napravljena simulacija kretanja mobilnog robota kroz prostor. Kako bi robot što vjernije i kvalitetnije slijedio referentnu trajektoriju u upravljački sustav robota je ugrađen analitički neizraziti regulator. Nakon dokazivanja uspješnosti metode u simulacijama, izrađen je algoritam za lociranje robota i prepreka pomoću kamere smještene iznad labirinta. U algoritmu za lociranje robota i prepreka su korištene celularne neuronske mreže. Potom je napravljen sustav vođenja robota labirintom. Svi algoritmi korišteni u ovom radu su napravljeni u MATLAB programskom paketu.The subject of this paper is the guidance of a mobile robot through the maze. The first step of guiding robot through the maze is path planning. The second step is to project the algorithm for controlling location of robot and obstacles. The third step is to make controlling system for guidance of robot. Path planning of mobile robot is performed by excitable digital reaction-diffusion systems (DRDS). Stability of the system and ability of path planning in complicated spaces are achieved by improving existing excitable DRDS. After path planning with DRDS, it was shown how to generate referent trajectory with analyzing the motion of the mobile robot through planed path. While generating trajectory, the way robot moving was considered. Afterwards, the simulation motioning of the mobile robot through the maze was shown. An analytical fuzzy control was built in the control system of the robot in order to make it more consistent with the referent trajectory tracking. When the success of the method was proven through the simulations, an algorithm for locating the robot and the obstacles with camera, placed above the maze, was made. Cellular neural networks were used in the algorithm of locating robot and obstacles. Lastly, a system of guidance of robot through the maze was made. All proposed algorithms where programmed in MATLAB