چكيده به لاتين
Neuromorphic architectures are hardware systems designed to use the function principles of the brain neural networks as the basis of their mechanism of action. An advantage of neuromorphic systems is to create decentralized operations associated with abundant parallelism with simple processing units. Machine vision, audio processing, image recognition, and pattern recognition are some examples of the applications of these systems. The accurate function of electrical circuits has always been the focus of attention. Radiation appears as a factor capable of disabling the operation of these circuits. Accordingly, the development of circuits resistant to the reduced effects of radiation has always been a goal. These circuits have advanced with the growth of electronic devices.
This dissertation aimed to design a neuromorphic circuit to be used in robots and spacecraft operating outside the Earth's atmosphere or in other words, in harsh environments with radiation, which can function properly. It was found after examining research samples in launcher laboratories that radiation can cause changes in the threshold voltage, and thereby, increases the short circuit of MOSFET transistors. In this context, axon-hillock neuromorphic circuit parameters were selectively measured before and after radiation. According to the results, the presence of radiation leads to changes in the threshold voltage of this transistor. The increased threshold voltage enhances the simultaneous On-state time of pull-up network and pull-down network transistors. The simultaneous On-state of the pull-up network and pull-down network transistors increases the short circuit current in the circuit. The increased short circuit current will lead to enhanced dynamic power. Thus, the short circuit current reduction techniques were studied in the dissertation. In doing so, the technique of placing a resistor between the pull-up network and the pull-down network transistors was used to reduce the short circuit current.
