Deadzone compensation in motion control systems using neural networks

Function approximation-based sliding mode adaptive control

“…(2) Calculate the value of sliding function s(t) according to (6). in [−10, 10] and arrives to peak value of −20 sometimes.…”

“…Furthermore, dead zone uncertainties' bounds remain unknown in many practical DC motor control systems. This problem cannot be coped with conventional sliding mode controller [8,9] and general adaptive controller [1][2][3][4][5][6][7]. In order to deal with nonlinear systems with unknown bound time-varying uncertainties, adaptive control schemes combined with sliding mode technique have been developed [10][11][12][13].…”

“…In general, there are two usual methods treating the systems with uncertain timevarying dead zone characteristics caused by uncertain nonlinear frictions in DC motor position control systems. The first one is to separate the unknown dead zone from the original DC motor systems and construct an adaptive dead zone inverse, and then compensate the effects of unknown dead zone characteristics [1][2][3][4]; The second method is to deal with both the unknown dead zone characteristics and all the other uncertainties as one uniform uncertainty, thereupon design proper compensator [5] or adaptive controller which can counteract the effects of uncertainty [6,7]. Furthermore, dead zone uncertainties' bounds remain unknown in many practical DC motor control systems.…”

Constructive Approximation of Discontinuous Functions by Neural Networks

“…piecewise-continuous). Examples include friction, deadzones, backlash, and so on [28,29]. It is found that attempts to approximate piecewise continuous functions using smooth activation functions require many hidden nodes (neurons) and many training iterations, and still do not yield very good results [27].…”

Adaptive fuzzy output feedback control of uncertain nonlinear systems with nonsymmetric dead-zone input

“…However, the results motivated above in [6-16, 20-25, 27, 28] did not consider the affection of unknown dead-zone inputs. Recently, based on the state feedback control strategy the adaptive control schemes were proposed in [17,18] for nonlinear systems with unknown functions and dead-zone inputs by using the approximation property of the neural networks. More recently, based on the universal approximation property of the fuzzyneural networks, an adaptive fuzzy-neural observer design algorithm is studied in [19] for a class of nonlinear SISO systems with both a completely unknown function and an unknown dead-zone input.…”