2D visual servoing for a long range navigation in a cluttered environment

In this paper, we consider the problem of realizing a vision-based long range navigation in a cluttered environment. To this aim, we couple a topological environment representation and a supervision algorithm. The latter organizes several elementary tasks which allow to perform image based visual servoing, obstacle avoidance and to estimate on line several required parameters. Well known and new methods based on the tasks function approach are introduced. The dynamical sequencing formalism is used to guarantee the global control law smoothness. Finally, simulation results validate our approach.