A phenomenological dynamic model of a magnetorheological damper using a neuro-fuzzy system

Abstract

A magnetorheological (MR) damper is a promising appliance for semi-active suspension systems, due to its capability of damping undesired movement using an adequate control strategy. This research has been carried out a phenomenological dynamic model of two MR dampers using an adaptive-network-based fuzzy inference system (ANFIS) approach. Two kinds of Lord Corporation MR damper (a long stroke damper and a short stroke damper) were used in experiments, and then modeled using the experimental results. In addition, an investigation of the influence of the membership function selection on predicting the behavior of the MR damper and obtaining a mathematical model was conducted to realize the relationship between input current, displacement, and velocity as the inputs and force as output. The results demonstrate that the proposed models for both short stroke and long stroke MR dampers have successfully predicted the behavior of the MR damper with adequate accuracy, and an equation is presented to precisely describe the behavior of each MR damper.

Decentralized control for stabilization of nonlinear multi-agent systems using neural inverse optimal control

“…to be inverse optimal (globally) stabilizing if: -(i) It achieves (global) asymptotic stability of χ i = 0 for system (15);…”

“…for control law (16) in order to ensure stability of the equilibrium point χ i,k = 0 of system (15), which will be achieved by defining an appropriate matrix P i . Moreover, it will be established that control law (16) with (19), which is referred to as the inverse optimal control law, optimizes a meaningful cost function of the form (2).…”

“…Lemma 1 Consider the affine discrete-time nonlinear system (15) with i = 1, If there exists a matrix P i = P T i > 0 such that the following inequality holds: (20); then, the equilibirum point χ i,k = 0 of system (15) is globally exponentially stabilized by the control law (20), with CLF (19).…”

“…(20); then, the equilibrium points χ i,k = 0 of system (15) is globally exponentially stabilized by the control law (20), with CLF (19).…”

“…This fact motivates the need to derive a model which can really consider all system dynamics, soft computing methodologies have been extensively used for system modeling and identification as it is exemplified by multiple applications [10], [11], [12], [13], [14], [15], [16], [17], [18]. However, artificial neural networks, through their multiple application, have been demonstrated to be ideal for modeling complex nonlinear systems, due to their approximation capabilities, easy implementation, robustness against noise and online training, which make them very adequate for modeling complex nonlinear systems [19].…”