Point-to-point trajectory planning of wheeled mobile manipulators with stability constraint. Extension of the random-profile approach

We propose a flexible stochastic scheme for point-to-point trajectory planning of nonholonomic wheeled mobile manipulators subjected to move in a structured workspace. The problem is known to be complex, particularly if obstacles are present and if dynamic stability constraint is considered. The proposed method consists of extending to wheeled mobile manipulators the random-profile approach recently applied to wheeled platforms. This versatile method handles constraints on: (i) geometry (obstacle avoidance, bounded joint positions and path curvature); (ii) kinematics (bounded velocities and accelerations); (iii) dynamics (bounded torques, stability condition). It may be applied using various forms of cost functions involving travel time, efforts and power. Solutions are presented for planar and spatial nonholonomic wheeled mobile manipulators undertaking, in a constrained workspace, a point-to-point task defined either in generalized or operational coordinates.