Man is separated from machines by the ability to think reason, imagine and create. All of these actions are made possible via a complex network of neurons, massed at the top of the brain stem. Artificial neural networks attempt to simulate the actions of these biological neurons. Like their biological counterparts, network feedback produces trained outputs for given inputs, that is, the system learns. Neural control has significant advantages in analog to digital conversion. During the course of this project, we have researched traditional neural schemes and developed a new line of attack for a 4 bit neural A/D converter. This converter offers significant advantages over traditional A/D in cost, speed and accuracy. Preliminary testing has shown an increase to 8 bit accuracy with a conversion time less than 90 microseconds.
 

• (No web site available)
• Return to 1994 Project List
• Return to main Senior Project page