Robust neural network tracking controller based on simultaneous perturbation stochastic approximation.
Kyaw, Minn Latt.
Date of Issue2003
School of Electrical and Electronic Engineering
the robust neural controller based on the SPSA has been developed to obtain the guaranteed stability with a normalized learning algorithm. A three-layered neural network is used for the simulation study with 30 hidden layer neurons and two output neurons, which was trained by the standard back-propagation and SPSA training algorithm.
DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics
Nanyang Technological University