CMOS implementation of locally connected neural cells to solve the steady-state heat flow problem
R. Yentis, Mona Elwakkad Zaghloul · 2002
A locally connected neural network for solving a class of partial differential equations is designed and simulated with SPICE. Each neural cell is designed using active and passive components. An architecture is described to control the components. The CMOS VLSI implementation was fabricated and measured. A sixteen cell network is also simulated and measured to solve the steady-state heat flow problem and the results are compared to the numerical solution of the partial differential equation.