Affiliations: Instituto Tecnológico de Nuevo León, Departamento de Ingeniería Eléctricay Electrónica, Av. Eloy Cavazos 2001. Col Tolteca, 67170 Guadalupe, N.L., Mexico | Université de POITIERS, Laboratoire de Mécanique des Solides U.M.R. 6610 C.N.R.S, SP2MI Téléport 2 Bd. Marie et Pierre Curie BP 30179 86962 Futuroscope-Chasseneuil CEDEX, France
Note: [] Corresponding author P. Quintero-Alvarez, Instituto Tecnológico de Nuevo León, Departamento de Ingeniería Eléctrica y Electrónica, Av. Eloy Cavazos 2001, Col Tolteca, 67170 Guadalupe, N.L., Mexico. Tel.: 52-81-81570511; Fax: 52-81-81570306; Email: [email protected]
Abstract: In previous works, we treated the collision-free path-planning problem for a nonholonomic mobile robot in a cluttered environment. We used a method based on a representation of the obstacles on the robot's velocity space. This representation is called Feasible Velocities Polygon (FVP). Every obstacle in the robot's influence zone is represented by a linear constraint on the robot's velocities such that a collision between the robot and the obstacle could be avoided. These constraints define a convex subset in the velocity space, the FVP. Every velocity vector in the FVP ensures a safe motion for the given obstacle configuration. The path-planning problem is solved by an optimization approach between the FVP and a reference velocity to reach the goal. In this paper, we have extended our work to an articulated mobile robot evolving in a cluttered environment. This robot is composed of a differential mobile robot and one or several modules that together form the trailer which are linked by off-center joints. This kind of robot is a strongly constrained system. Even in a free environment, under some circumstances, the robot may be blocked by its trailers in its progression towards the goal. The proposed approach, compared to other methods, has the main advantage of integrating anti-collision constraints between the articulated robot itself and the environment, in order to avoid and resolve dead-lock situations. For moving to the final position, the articulated mobile robot uses the FVP and a reference control law, to formulate the constraints method as a problem of minimal distance calculation. This formulation is then solved with the algorithm of minimal distance calculation proposed by Zeghloul (Zeghloul and Rambeaud, 1996). When a dead-locking situation arises and according to the robot–obstacle configuration, we have developed three different modules to solve these conditions. Each module uses a different approach to resolve the blocking situation. In order to show the capabilities of our method to lead the articulated robot to the final position in a stable way, a numerical result is presented.
Keywords: Nonholonomic systems, collision-free path-planning, feasible velocities polygon, articulated mobile robot, differential mobile robot
