Direct adaptive neural control for affine nonlinear systems

Stability analysis and robustness design of nonlinear systems: An NN-based approach

“…There is a need to develop more robust and effective control algorithms for nonlinear structural systems. In recent years, neural-network modeling has become the most often used technique for developing nonlinear control systems [24][25][26]. For example, Kar and Behera [26] developed a direct adaptive neural control scheme for a class of affine nonlinear systems which are exactly input-output linearizable by nonlinear state feedback.…”

“…In recent years, neural-network modeling has become the most often used technique for developing nonlinear control systems [24][25][26]. For example, Kar and Behera [26] developed a direct adaptive neural control scheme for a class of affine nonlinear systems which are exactly input-output linearizable by nonlinear state feedback. Shoorehdeli et al [27] developed a novel hybrid learning algorithm with stable learning laws for an Adaptive Network-based Fuzzy Inference System (ANFIS) that could be used as a system identifier.…”

Neural networks in virtual reference tuning

“…Then, as NN M is fixed, its derivatives with respect to any parameter are known and easy to compute when training NN C . There are other neuro-control paradigms, such as reinforcement learning (Sutton and Barto, 1998), sliding controllers (Baruch, 2007;Kar and Behera, 2009), neural feedbacklinearisation, internal model control, model predictive control (Hagan and Demuth, 1999;Hagan et al, 2002), etc. Neural networks appear as well as elements in higher-level learning frameworks.…”

Neural network based feedback linearization control of a servo-hydraulic vehicle suspension system

“…The requirement of full-state measurements and lack of guarantee for robustness are among the drawbacks of FBL which motivate for its augmentation with appropriate intelligent control techniques: neural network (Yesildirek and Lewis, 1995;He et al, 1998;Boutalis, 2004;Kar and Behera, 2009;Poursamad, 2009), and fuzzy logic (Boukezzoula et al, 2007)). …”

“…NN modelling was based on a multi-layer perceptron NN model and trained off-line using the LevenbergMarquardt algorithm. The NNFBL controller designs documented by Poursamad (2009) as well as Kar and Behera (2009) are based on the RBFNN architecture.…”