Gradient-based methods of tuning noise covariance for an induction motor model
Daniel C. Lee · 2016
This paper proposes a stochastic gradient-like optimization for experimentally determining parameters to be used in the extended Kalman filter in an induction motor model. The gradient-like algorithm can be combined with an evolutionary algorithm in order to avoid being trapped in a low-performance local extremum. Recursive methods of computing the gradient is presented.