Implementation of Simple Neurons with Complete Hardware-Based Learning Capabilities. | ||||
| MEJ- Mansoura Engineering Journal | ||||
| Article 3, Volume 23, Issue 1, March 1998, Page 12-21 PDF (321.62 K) | ||||
| Document Type: Research Studies | ||||
| DOI: 10.21608/bfemu.2021.149609 | ||||
 View on SCiNiTO | ||||
| Author | ||||
Hassan H. Soliman 
| ||||
| Electronics and Communications Engineering Department., Faculty of Engineering., El-Mansoura University., Mansoura., Egypt. | ||||
| Abstract | ||||
| In artificial neural network , implementation of processing units (neurons) which have programmable connection weights is the most process that takes many research efforts. Most of these efforts are dedicated to the implementation using VLSI techniques . unfortunately, VLSI implementing are not available in most developing countries such as Egypt . in this paper , simple design for implementing ADALINE-like neurons in artificial neural network (ANN)with complete learning capabilities is presented. The proposed design is based on the commercially available electronic components, however , it can be easily extended to be implemented using mixed (digital / analog) VLSI technology. The used components are so minimized to get simple design . moreover, the number of input connections for the neuron could simply be increased by adding asmall resistor for each new connection input. While main forward blocks of the neuron, e.g. summing function , are implemented using analog circuitry, the control of connection weights is implemented using digitally controlled circuitry. | ||||
|
Statistics Article View: 62 PDF Download: 222 |
||||
