Robust Neural Identification of Robotic Manipulators Using Discrete Time Adaptive Sliding Mode Learning

The problem of identification of uncertain nonlinear systems using feedforward neural networks is investigated. The weights of the neural identifier are updated on-line by a discrete-time learning algorithm based on the sliding mode control technique, which is well known with its robustness to uncertainties. The learning parameters are adjusted to force the error between the actual and desired neural network outputs to satisfy a stable difference error equation and a quasi-sliding mode on the zero learning error is established. The behaviour of the proposed discrete-time algorithm is illustrated by using it for the neural identification of an experimental robotic manipulator. The results show that the neural model inherits some of the advantages of the sliding mode control approach, such as high speed of learning and robustness.