The International Journal of Robotics Research

This article presents some results of an ongoing research project to investigate the components and modules that are necessary to equip a robot with exploratory capabilities. Of particular interest is the recovery of certain material properties from a surface, given minimal a priori information, with the intent to use this information to enable a robot to stand and walk stably on a surface that is unknown and unconstrained. To this end, exploratory procedures (EPs) have been designed and implemented to recover penetrability, compliance, and surface roughness by exploring the surface using a compliant wrist sensor. A six-degree-of-freedom compliant wrist sensor that combines passive compliance and active sensing has been developed to provide the necessary flexibility for force and con tact control, as well as accurate position control. This article describes the compliant wrist and sensing mechanism design along with a hybrid control algorithm that utilizes the sensed information from the wrist to adjust the apparent stiffness of the end effector as desired.

The high level of skill required to maneuver present-day, flexible endoscopes that are utilized in esophagoscopy and colonoscopy has prompted us to propose a new class of safer, more easily controlled devices. These devices are mechanical teleoperators with few degrees of freedom based on the famil iar bead chain with internal control cables. The study of the physics associated with the flexible control devices is carried out with respect to the design parameters of both the practical system and its envaronment. The suitable geometries for lock ing such a mechanism are presented herein. We introduce the "slide motion" scheme, which provides a smooth movement of the endoscope stem with only three continuously controlled degrees of freedom. Such movement is accomplished by imple menting a series of relative motions between an outer tube and intermittently stiff spines inside. The strength and stability of the locking mechanism, which exhibits intermittent stiffness, is also discussed.

This article introduces an invariant measure to quantify the motion of a rigid body attached to the end effector of a serial manipulator. The measure arises as a natural consequence of the properties of the Euclidean group. It is invariant with respect to the position and orientation of the fixed and moving coordinate systems used to parameterize the motion of the manipulator. It is independent of the type and number of joints in the chain and the dimensions used to parameterize the chain. The singularities of the manipulator are also predicted by the measure. Finally, it is shown that, when applied to redundant manipulators, the measure quantifies the importance of each joint of the manipulator.

In this article, a control methodology is proposed that ad dresses the problem of control of robotic manipulators during a general class of task that requires the manipulator to make a transition from (I) noncontact motion to contact motion and (2) contact motion to noncontact motion. Specifically, the problem of control of a general n-degree-of-freedom rigid-link robotic manipulator during a general class of tasks, as described ear lier, is treated. It is assumed that during the contact phase of the assumed task, frictionless point contact is made with fixed objects in the manipulator work environment, which is modeled as a linear mechanical impedance. Furthermore, it is assumed that exact knowledge of the manipulator and work environment kinematic and dynamic parameters is available.

The following closed-loop behavior is achieved with the pro posed control law: (1) the closed-loop system exhibits global asymptotic stability; (2) asymptotic trajectory tracking of gen eralized force and position inputs is achieved; and significantly, (3) on inadvertent loss of contact by the manipulator, contact is reestablished, and generalized forces and positions are again asymptotically achieved. This closed-loop behavior is achieved in the presence of collisions between the manipulator and ob jects in the manipulator work environment; such collisions may include multiple impacts as a result of bouncing on initial contact. A general mathematical framework is established to prove that the closed-loop robotic system, with a discontin uous control applied, is asymptotically stable. Experimental results performed on a two-degree-of-freedom direct-drive robot support the theoretical claims made in this work.

This article discusses the numerical solution of the forward dynamic equations of an n-degree-of-freedom manipulator with friction. Also discussed are the modeling and experimen tal identification of friction. It is shown that the inclusion of Coulomb-type friction in the dynamic equations introduces two difficulties in the forward dynamic solution. The differential equations are shown to be discontinuous in the highest-order derivative terms. In addition, the load dependency of this type of friction typically causes the equations to be implicit in the joint accelerations. For the important case of load-dependent transmission friction, the equations can be explicit. Techniques for the forward solution are described through the example of a roller screw transmission. Experimental and simulation results are used to show the importance of load-dependent friction in a particular robot.

Manipulator regressors are n x l matrix functions in the dynamic expressions τ = Y(q, q, q)ζ_r or (1/D + σ)τ = W(q, q)ζ_r, which linearize the robot dynamics with respect to a properly defined inertia parameter vector ζ _r ∈ R^l. Many modern adaptive controllers require on-line computation of a regressor to estimate the unknown inertia parameters and ensure robustness of the closed-loop system.

Although the computation of Y(q, q, q) has been studied by Atkeson et al. (1985), Khosla and Kanade (1985), and Khosla (1989), the computation of W(q, q) for a general n-link robot has not been reported in the literature. This article presents an algorithm to compute W(q, q) for a general n-link robotic manipulator The variables used to construct the regressor matrix are directly available from the outward iteration of a Newton-Euler algorithm; some additional arithmetic operations and first-order, low-pass filtering are needed. The identification of unknown inertia parameters is also discussed.

One of the problems in real-time control of redundant manip ulators is considerably increased computational complexity compared with nonredundant robots. This is due to the rela tively high dimensions of the Jacobian matrix (6 x n, n ≥ 7) and the fact that its pseudoinverse is to be computed. On the other side, analytical solutions for nonredundant robots offer minimal numerical complexity (if such exist). The main idea in this article is to reduce computational complexity by com bining the analytical and the pseudoinverse solution. Namely, some of the joint angles that do not actively participate in the redundancy are evaluated analytically, while the group of joint angles that are actually redundant are solved using the pseu doinverse of the Jacobian matrix. In this way the dimensions of Jacobian are considerably reduced yielding several times less computational complexity. Further reduction of computational time is achieved by applying the gradient projection method (Dubey et al. 1991) to the actually redundant subrobot. In this article the solution is derived for a 7-DOF manipulator, and the possibility of application to other robots is discussed. At the end, simple performance criteria for avoiding wrist singularity that take into account the knowledge of robot geometry are proposed.