The application of adaptive control algorithms for vibration suppression of smart structures is investigated in this paper. An accurate mathematical representation is not required in this approach. The controller adapts to the parameter variations of the structural system by updating the controller gains. When the desired performance of an unknown plant with respect to an input signal can be specified in the form of a linear or a non-linear differential equation, stable control can be achieved using model-reference adaptive control (MRAC) techniques. The conventional MRAC techniques have been successfully implemented on a smartstructure test article resulting in perfect model following. We have also investigated the design of neural-network-based adaptive control system for smart structures. Two neural networks, one for system identification and another for control purposes, are needed for the implementation of adaptive controilers. In this paper, a direct variable model identification technique using neural networks has been developed. An iterative inversion of a neural model of the forward dynamics of the plant h a s been utilized for the implementation of a neural controller. This algorithm generates a smooth control. We have also developed an adaptive neuron activation function for reducing the learning time of neural networks. The performance of the neurainetwork-based adaptive controllers has been verified using simulation studies.

This paper describes an integrated approach to design and implement robust controllers for smart structures. To demonstrate this procedure, we have designed and fabricated a structural test article incorporating shape memory &y !sup.! ac!ga!o=:, controllers with flexible structures. A neural-network-based structural identification method to determine a state space model of the system from its experimental inputloutput data is presented. To reduce the learning time required to train a neural network significantly, we have developed an accelerated adaptive learning-rate algorithm. The mathematical model derived using neural networks is compared with models obtained by more conventional and well known methods. Using this model, a modified linear quadratic Gaussian with loop transfer recovery (LQGLTR) controller is designed for vibration suppression purposes. This robust controller accommodates the limited control effort produced by SMA actuators. A multilayered feedfotward neural network is then trained to mimic this controller. These designs are all then realized as digital controllers and their closed-loop performances have been compared. In particular, the robustness properties of the controller have been verified for variations in the mass of the test article and the sampling time of the controller. gigge sen+or+, sign&proaes+ing cim& 2nd digi!~!

This paper details identification and robust control of smart structures using artificial neural networks. To demonstrate the use of artificial neural networks in the control of smart structural systems, two smart structure test articles were fabricated. Active materials like piezoelectric (PZT), polyvinylidene (PVDF) and shape memory alloys (SMA) were used as actuators and sensors. The Eigensystem Realization Algorithm (ERA), a structural identification method has been utilized to determine a minimal order discrete time state space model of the test articles. The ERA requires the Markov parameters of the physical system. A neural network based method has been developed to estimate the Markov parameters of a multi input multi output system from experimental test data. The accelerated adaptive learning rate algorithm and the adaptive activation function were utilized to improve the learning characteristics of the network and reduce the learning time. The identified models were used to design a robust controllers for vibration suppression of smart structures using a modified Linear Quadratic Gaussian with Loop Transfer Recovery (LQG/LTR) method. This control design methodology has better loop transfer recovery properties while accommodating the limited control force available from the SMA and the PZT actuators. This controller was copied into a feedforward neural network using the connectionist approach. This neural network controller was implemented using a PC based data acquisition system. The closed loop performance and robustness properties of the conventional and the neural network based controller are compared experimentally.

No abstract

No abstract