Robotic Trajectory Planning Using Computer Vision Based on Signaling Marker Identification and Graph Theory Applications

• Lucas Da Silva Carvalho IFSP
• Glauber Da Rocha Balthazar IFSP

Resumo

A mobile robot, called Sunrise Robot, was used in this work to develop an autonomous movement from one point to another. For this purpose, Graph Theory was used, through the Dijkstra's algorithm, which modeled a trajectory between vertices in a directed graph, determining a displacement trajectory. Traffic markers were associated with the vertices and computer vision through YoloV8 recognized the visual elements in these markers. An algorithm was developed to receive the results of computer vision and associate them with the mathematical model of the graphs. This algorithm made the decision for the autonomous movement of the robot, moving it in the correct direction determined by the graph trajectory. The results showed promise in the training and identification of vertices and traffic markers, as well as in the autonomous movement of the robot in detriment of the direction defined by the directed graph.

Palavras-chave: Computer vision, Visualization, Accuracy, Educational robots, Navigation, Computational modeling, Robot sensing systems, Mathematical models, Trajectory, Robots, autonomous robot movement, vertex robot, labeling plates, traffic sign identification, Dijkstra theory robotic