Title :
Analog computing for real-time solution of time-varying linear equations
Author_Institution :
Nat. ICT Australia Ltd., Canberra, ACT, Australia
Abstract :
An implicit recurrent neural network model (IRNN) is proposed for solving on-line time-varying linear equations. Such a neural network can be implemented as analog circuits or VLSI. Excellent convergent properties have been revealed by careful theoretical analysis. In the specific case where the linear equation is obtained from a time-varying Sylvester equation, the proposed IRNN model coincides with some existing recurrent neural networks reported in recent literature, where simulation examples have been reported to demonstrate the effectiveness and efficiency.
Keywords :
VLSI; analogue computer circuits; analogue simulation; neural net architecture; recurrent neural nets; time-varying systems; VLSI; analog circuits; analog computing; implicit recurrent neural network model; neural network architecture; time-varying Sylvester equation; time-varying linear equations; Analog circuits; Analog computers; Australia; Convergence; Difference equations; Modeling; Neural networks; Recurrent neural networks; Systems engineering and theory; Very large scale integration;
Conference_Titel :
Communications, Circuits and Systems, 2004. ICCCAS 2004. 2004 International Conference on
Print_ISBN :
0-7803-8647-7
DOI :
10.1109/ICCCAS.2004.1346430
