Name
Title | ||
|---|---|---|
Nonlinear system identification using a Bayesian–Gaussian neural network for predictive control |
Abstract | ||
|---|---|---|
A Bayesian–Gaussian neural network (BGNN) is presented for nonlinear system identification, as well as for model predictive control. The topology and connection weights of this network can be set immediately when the training samples are available, and the output of it is a fusion of multiple pieces of information. The training of this network is a minimization process to optimize the input factors, rather than the connection weights plus thresholds of the back-propagation neural network (BPNN) or its variations, and therefore could save a large amount of time in training. The self-tuning ability of this network can easily be achieved in an optimal way so that it can on-line adapt to the shift of system dynamics. |
Year | DOI | Venue |
|---|---|---|
1999 | 10.1016/S0925-2312(98)00113-1 | Neurocomputing |
Keywords | Field | DocType |
Bayesian-Gaussian neural network,Back-propagation neural network,Non-linear system identification,Network training,Self-tuning ability | Pattern recognition,Computer science,Stochastic neural network,Model predictive control,Nonlinear system identification,Probabilistic neural network,Time delay neural network,Gaussian,Artificial intelligence,System dynamics,Artificial neural network,Machine learning | Journal |
Volume | Issue | ISSN |
28 | 1-3 | 0925-2312 |
Citations | PageRank | References |
2 | 0.53 | 0 |
Authors | ||
2 | ||