In this paper, we present two DCAL-like (Desired Compensation Adaptation Law) controllers for link position tracking of n-link, rigid, revolute robot manipulators. First, we use a simplified stability analysis to illustrate global asymptotic link position-velocity tracking for a DCAL-like controller with nonlinear feedback. The proof is simplified by employing a different structure for the nonlinear feedback than that originally proposed and by making use of the nonlinear damping control design tool. We then use the nonlinear damping tool to show that a DCAL-like controller with linear feedback can guarantee semi-global asymptotic link position-velocity tracking. The proposed nonlinear and linear feedback DCAL-like controllers are experimetally tested and compared using the Integrated Motion Inc. 2-link direct drive robot manipulator.

The minimum-time and subminimum-time joint trajectories of manipulators with geometric path constraints are planned in consideration of physical constraints based on kinematics and dynamics. The idea of time scaling is introduced, i.e. a time scale factor k(t) and a set of joint trajectories, called reference trajectories, are used to describe all the sets of trajectories tracing the specified geometric path. The desirable factor k(t) which makes the travelling time as short as possible is obtained by two proposed methods: the first one is an iteratively improving method using B spline, and the second one is a directly minimizing method. These two methods are preferably applied to a geometric collision-free path of a manipulator.

Learning control is a new approach to the probelm of skill refinement for robotic manipulators. It is considered to be a mathematical model of motor program learning for skilled motions in the central nervous system.

This paper proposes a class of learning control algorithms for improving operations of the robot arm under a geometrical end-point constraint at the next trial on the basis of the previous operation data. The command input torque is updated by a linear modification of present joint velocity errors deviated from the desired velocity trajectory in addition to the previous input. It is shown that motion trajectories approach an e-neighborhood of the desired one in the sense of squared integral norm provided the local feedback loop consists of both position and velocity feedbacks plus a feedback term of the error force vector between the reactive force and desired force on the end-point constrained surface. It is explored that various passivity properties of residual error dynamics of the manipulator play a crucial role in the proof of uniform boundedness and convergence of position and velocity trajectories.

When a free-ranging walking machine carries its own fuel and power source as well as a payload, it is important to maintain high mechanical efficiency. If the masses of the legs are negligible then, in a straight and level walk, the only power to be generated would be that needed to overcome ground and wind resistance and internal frictions. The paper shows this to be the case when one-input straight-line generating mechanisms are used. However, when two-input mechanisms are used the actuators can work against each other resulting in finite, sometimes large, power requirements. The paper determines the effects of limb lengths and ground clearance on step length and quantifies the power requirements of serial-operated and parallel-operated legs. It shows that parallel-operation offers greater efficiency.

A motion controller for an overhead Cartesian crane in three-dimensional Euclidean (3-D) space is designed under the constraint that the control action belong to a discrete set of assigned values. The design approach rests upon a two-step procedure: first, a constraintfree motion controller is determined that satisfies the required dynamic specifications; second, this controller is replaced with an equivalent controller satisfying the discrete action constraint. The first step is implemented by means of a heuristic 3-D extension of a well-proven 2-D controller, the second step by applying recent sliding mode results. Numerical simulations illustrate the properties of the resulting feedback system under both nominal and perturbed operating conditions.

This paper describes an adaptive control system for an articulated robot with n joints carrying a variable load. The robot is a complex nonlinear time-varying MIMO plant with dynamic interaction between its inputs and outputs. However, the design of the control system is relatively straightforward and does not require any prior knowledge about the plant. This is because the control system is based on using neural networks which can capture the dynamic characteristics of the plant automatically. Three neural networks are employed in total, the first to learn the dynamics of the robot, the second to model its inverse dynamics and the third, a copy of the second neural network, to control the robot.