On training optimization of the generalized ADLINE neural network for time varying system identification

On-line system identification of linear time-varying (LTV) systems whose system parameters change in time has been studied lately. One neural network based such on-line identification method was studied by the author with a generalized adaptive linear element (ADALINE). Since the ADALINE is slow in convergence, which is not suitable for identification of LTV system, one technique was proposed to speed up training, that is, to introduce a momentum term to the weight adjustment during convergence period. Experimental study was then performed to search for an optimal combination of the momentum term and the learning rate eta. The goal was to speed up convergence (or tracking) while keeping smooth tracking during any transient period. Simulation results show that several optimal combinations of the momentum factor and learning rate were found and the time varying parameters of LTV systems could be identified quite effectively; which, in turn, sows that the fined tuned GADLINE is quite suitable for online system identification and real time adaptive control applications due to its low computational demand.