Predictive Force Control of Robot Manipulators in Nonrigid Environments

Abstract

The application of robot manipulators in industry is in general related to tasks such as manipulation or painting that requires only position control of the arm. Nonetheless, there are other robotic tasks like pushing, polishing and grinding that require interaction between the manipulator and a contact surface or environment. This fact leads to the desire of controlling the interaction between the robot and the environment. Although a lot of different control schemes has been proposed in the literature, as surveyed by (Zeng & Hemami, 1997 ; De Schutter et al., 1997), the major force control approaches can be classified as hybrid control (Raibert & Craig, 1981) or impedance control (Hogan, 1985). The hybrid control separates a robotic force task into two subspaces: a force controlled subspace and a position controlled subspace. Two independent controllers are then designed for each subspace. In contrast, impedance control does not attempt to control force explicitly but rather to control the relationship between force and position of the end-effector in contact with the environment. Furthermore, when the environment is rigid with known characteristics it is possible to plan a virtual trajectory, such that a desired force profile is obtained (Singh & Popa, 1995). However, the same does not hold in the presence of nonrigid environments, which disables a reliable application of the classical impedance controller. This problem has motivated the development and design of more sophisticated force control methodologies which usually take into consideration the dynamics of the environment. In (Love & Book, 1995) it is shown that the performance of an impedance controlled manipulator increases when the desired impedance includes some modeling of the environment. Another possible solution to tackle this problem is to use a modelbased control scheme like predictive control, which incorporates the manipulator and environment models in a force optimization-based strategy (Wada et al., 1993). Recently, a force control strategy for robotic manipulators in the presence of nonrigid environments combining impedance control and a model predictive control (MPC) algorithm in a force control scheme has been proposed (Baptista et al., 2000b). In this force control methodology, the predictive