Visual Predictive Control Scheme for a Mobile Robot Navigating in a Cluttered Environment

• Adrien Durand-Petiteville UFPE
• Viviane Cadenat LAAS-CNRS / Université Paul Sabatier Toulouse III

Resumo

This work focuses on the use of Visual Predictive Control to drive a mobile robot in a cluttered environment. To efficiently and safely achieve the navigation task, one proposes to modify a classical VPC scheme by (i) relaxing some control input boundaries to extend the feasibility set and guarantee the stability, (ii) adding obstacle constraint over the pieces of trajectory, and (iii) modifying the calculated solution to avoid local minima. Simulated results highlight the efficiency of the proposed approach.

Palavras-chave: Robots, Visualization, Cameras, Trajectory, Robot vision systems, Navigation, Robot kinematics