Adaptive Neural Network Finite-Time Control for Uncertain Robotic Manipulators

Adaptive neural finite-time control for a class of switched nonlinear systems

“…It has been proved that some inherently nonlinear systems, which cannot be stabilized by any smooth feedback control method, may be stabilized by using finite-time control methods [23,24]. Recently, some results on finite-time control for robotic manipulators by using neural network have been reported in [28,29]. Based on the terminal sliding mode control method, by using a radial basis function neural network (RBFNN) based compensator to approximate the uncertain nonlinear function, a neural network based robust finite-time control strategy was proposed for the trajectory tracking of robotic manipulators with structured and unstructured uncertainties in [28].…”

“…Based on the terminal sliding mode control method, by using a radial basis function neural network (RBFNN) based compensator to approximate the uncertain nonlinear function, a neural network based robust finite-time control strategy was proposed for the trajectory tracking of robotic manipulators with structured and unstructured uncertainties in [28]. In [29], a second-order uncertain nonlinear dynamical system was considered, where the uncertainties and external disturbances w(x) were unknown and modeled by a RBFNN. An adaptive neural network finite-time controller was proposed by using the backstepping method.…”

On Training Flexible Robots using Deep Reinforcement Learning

“…In such cases, reinforcement learning and policy search algorithms that can learn from a robot's experience have been shown to be successful [8], [9] for tasks such as object manipulation [10], [11], [12], locomotion [13], [14], [15], [16] and flight [17]. However, most of this work involves using a model-free component to approximate features of the robot or the world that cannot be modeled while still using model-based controllers for other parts of the system [12], [18] In work where flexibility is taken into consideration, learning is still based either on building a more complex model [6], [19], [20], an approximate model [21] or plugging in a learned-model component into a model-based controller. Recently, work involving end-to-end model-free methods using deep reinforcement learning have been demonstrated successfully in rigid real robots [22], [23], [16].…”

Adaptive fuzzy finite-time control for a class of switched nonlinear systems with unknown control coefficients

“…It has been proved that some inherently nonlinear systems, which cannot be stabilized by any smooth feedback control method, may be stabilized by using finite-time control methods [26,27]. Up to now, a large number of studies have been done on finite-time stability and stabilization of nonlinear systems or switched nonlinear systems [25][26][27][28][29][30][31][32][33][34][35][36][37][38][39].However, unknown nonlinearities in the systems are usually restricted or even unconsidered in the existing finite-time stabilization results. For example, in [26,31,36], the considered systems do not include unknown nonlinearities.…”

“…In [29,30,38], unknown nonlinearities were allowed in the considered systems and finite-time controllers were constructed under some assumptions that the unknown nonlinearities were dominated by the products of unknown linear parameters and known functions, which is very restricted and some practical systems may not satisfy these conditions. In [35,37], some results on finitetime control for robotic manipulators by using neural network have been reported. It is worth pointing out that the control coefficients of the systems studied in the aforementioned papers are required to be either exactly known [26, 27, 29-31, 36, 39] or unknown but limited in a known constant interval [32].…”