Digital computing offers ease of use and flexible programming while analog
computing enables fast real time processing. For calculus functions such
as integration, unlike digital computing, analog computing does not suffer
from step size limitations. Hence, this project combines the best of both
domains in order to effectively model large scale dynamic systems in
applications such as neural networks and real time simulation of hardware
controller. It finds immense utility in fields like automotive systems,
radar, robotics and power systems.
We have built a prototype of a digitally programmable analog computer (DPAC) to solve a system of first order ordinary linear differential equations through analog integrators and summers. The desired configuration of the circuit for each system can be digitally programmed from one’s laptop or PC via an integrated microcontroller.
The talk will cover the motivation for an analog computer, the challenges involved, and the methods used by two independent teams to build the digitally programmable analog computer.