T. Hesketh

Adaptive learning control of robot manipulators in task space

Adaptive sliding mode control is considered, together with an iterative learning control scheme for tracking control of robot manipulators in Cartesian space coordinates. It is shown that the algorithm is globally convergent in the presence of external disturbances and modelling uncertainties. Owing to the robustness of the algorithm, a large gain for learning control can be used to achieve fast convergence of tracking errors. Moreover, the control scheme is rather simple and the inverse of the Jacobian matrix is not required.

Trajectory control of manipulators using adaptive sliding mode control

Considers an adaptive control algorithm for robot manipulators. The presented algorithm uses an adaptive sliding mode control law to ensure robustness. The proposed algorithms eliminate the requirement for inversion of the inertia matrix, the requirement for measurement of joint acceleration, and the requirement for calculation of a regressor. Computer simulation results show good properties of the proposed algorithm under large manipulator parameter uncertainties and disturbances.<<ETX>>

Piecewise output feedback controller synthesis of discrete time fuzzy systems

This paper presents an observer-based output feedback controller design method for discrete time Takagi-Sugeno fuzzy systems, using a discrete time piecewise Lyapunov function. The proposed piecewise output feedback fuzzy controller is made up of a piecewise state feedback fuzzy controller and observer for each subspace. The validity of such a scheme involving the separation principle is verified by a piecewise Lyapunov function. It is shown that the state feedback control laws and observer gains can be obtained by solving a set of linear matrix inequalities (LMI) that is numerically tractable with commercially available software. An example is given to demonstrate the advantage of the proposed method.

High Order Sliding-mode Control of Uncertain Linear Systems

Abstract High order sliding control of linear systems with unmatched perturbations is considered. A closed-loop system matrix assignment design procedure is proposed. The method is a modification of back-stepping design to take the advantage of scaling. A new high order sliding mode controller is presented, in which a nonlinear scaling function is used to achieve sliding mode.

H/sub /spl infin//-optimal control of multi-rate sampled-data systems

The hybrid multi-rate sampled-data system is considered which consist of a continuous time plant and a digital lifted controller. The paper describes an explicit solution to the multi-rate H/sub /spl infin//-optimal control problem using Neharis theorem. The solution is based on a lifting method, where an equivalent LTI system is obtained for the given periodic multi-rate problem. The causality constraint on the controller structure in lifting based solutions is automatically satisfied. By solving a standard Nehari problem, we obtain the optimal parameterization of a linear stabilizing controller in a worst case l/sub 2/ to l/sub 2/ sense and hence the optimal h-periodic controller for the multirate systems.

A novel 3D low voltage electrostatic RF MEMS switch with two movable electrodes

This paper will report on the design and fabrication of a novel 3D electrostatic RF MEMS switch, which uses two movable electrodes. The concept of two movable electrodes represents a unique feature of this device and is introduced to the RF MEMS community for the first time. Since the operating principle of the switch is based on electrostatic actuation, this unique feature results in a lower operating voltage. Combining the special bulk and surface micromachining techniques has enabled the realization of this new 3D RF MEMS switch. There are two main configuration for the device structure: 1) in the first device structure all parts are made up of bulk-micromachined free-structures. 2) In the second device structure the lower part is made up of a movable bulk-micromachined cantilever and the upper section is made up of surface micromachined movable thin film structures. By applying a DC voltage between movable plates, they come in touch and provide a pass for the RF signal (on-state of the switch) and as the DC voltage is removed, electrodes will be separated and disconnect the RF signal (off-state). The substrate can be used as a third electrode to separate beams in case of stiction. The monolithic nature of this switch technology makes it possible to develop various switch configurations like SPNT, C-type, and R-type switches, and switch matrices monolithically. This switch can be used as the basic building blocks for microwave switch matrices, multiplexers / demultiplexers, and phase shifters operating at microwave frequencies. The aim is to use the new features of this switch to achieve an acceptable low switching voltage, a better RF performance and particularly reliable switching operation. In this paper design considerations, HFSS simulation and the preliminary fabrication results of the switch are demonstrated.

Output feedback sliding-mode control of linear systems

The paper considers output feedback sliding-mode control of uncertain linear (nonminimum-phase) SISO systems. A new observer-free method is proposed, based on the transformation of the output feedback control problem to a state-feedback (scaled H/sub /spl infin//) control problem.

Betterment learning control of nonlinear systems

This paper considers betterment learning control of a class of uncertain nonlinear systems. We present a simple learning algorithm that exhibits uniform boundedness and ultimate uniform boundedness of the trajectories of the states. Moreover, if the initial condition errors are zero, the trajectories converge to zero.

Toward a Robust Model Predictive Controller Applied to Mobile Vehicle Trajectory Tracking Control

Abstract This paper presents an improved formulation of tube-based (i.e. a tube is a sequence of sets) robust model predictive control (MPC) paradigm to be applied to time-varying systems. In particular it is concerned with the trajectory tracking control problem. The improved formulation of tubes employs a tighter online prediction for the evolution of difference between nominal disturbance-free and actual states of the system, therewith ensuring improved performance. The proposed robust control approach is applied to the trajectory tracking control problem of UGVs (Unmanned Ground Vehicle) subject to bounded disturbance. Simulation results illustrate that the proposed control law yields suitable robust trajectory tracking with low level of conservativeness.

Non-minimal state-space realisation in H/sub /spl infin// control design

A non-minimal state-space realisation is used in this paper to represent linear uncertain systems described with input-output relationships. By this method, a dynamic output feedback control problem can be transformed to a state-feedback control problem. This will simplify the optimal I/O scaled H/sub /spl infin// output control problem.