Rapidly exploring random trees for motion planning of car-like vehicles

Autor: Katić, Rebeka
Přispěvatelé: Seder, Marija
Jazyk: chorvatština
Rok vydání: 2022
Předmět:
Popis: Ovaj rad bavi se implementacijom brzorastučih slučajnih stabala (eng.rapidly exploring random tree - RRT) na robota nalik automobilu. Zbog svoje efikasnosti u pretrazi prostora odabaran je Stable-Sparse RRT(SST) iz familiije RRT. Odlike takvih algoritama su pristranost prema neistraženim prostorima zbog čega se široko se koristi u autonomnom robotskom planiranju kretnje. Koristeći ROS (Robotic Operating System), Stage Simulator i ROS publishera izračuni algoritma su simulirani na robotu. Za uspješnu implementaciju problem bilo je potrebno prilagoditi mapu Stage Simulatoru i frekvencije objavljivanja brzina na /cmd_vel topic. This paper deals with the implementation of rapid-growing random trees (RRT) on a car-like robot. Stable-Sparse RRT (SST) from the RRT family was chosen for its efficiency in space search. Features of such algorithms are bias towards unexplored spaces, which is why it is widely used in autonomous robotic motion planning. Using ROS (Robotic Operating System), Stage Simulator and ROS publisher the algorithm calculations were simulated on a robot. For successful implementation of the problem, it was necessary to adjust the Stage Simulator folder and the frequency of publishing speeds on the /cmd_vel topic.
Databáze: OpenAIRE