Ponente
Descripción
Obstacle avoidance in robotics enables a robot to detect and avoid collisions with objects in its environment, ensuring safe and autonomous movement toward its goal. The primary objective is to guide the end-effector to its desired position while avoiding collisions, even in the presence of sudden or unavoidable threats, by minimizing damage through preventive actions.
Traditional methods such as Artificial Potential Fields (APF), collision maps, proximity sensors, or computer vision, which are reactive in nature, provide immediate solutions but exhibit limitations in complex dynamic environments. This work proposes a proactive approach using Model Predictive Control (MPC), which anticipates future trajectories and optimizes avoidance actions in real time while considering dynamic constraints.
A simulation has been implemented for a planar 2-DOF rotational robot with revolute joints (RR), where the end-effector follows a predefined trajectory and avoids an obstacle that disrupts its path. The results demonstrate the effectiveness of MPC in generating smooth and optimal maneuvers, surpassing the limitations of reactive methods.
