Presenter(s)
Baminahennadige Rasitha Fernando
Files
Download Project (1.8 MB)
Description
Control algorithms are used in almost all mechanical and electrical systems for controlling movements and activities. This includes robots, automobiles, aircrafts, industrial machines, and power systems. For mobile systems, the use of complex control algorithms – in particular adaptive control algorithms – would allow for much more refined performance. Unfortunately, these complex control algorithms are highly computationally intensive, requiring the use of high powered computers. This makes their use in mobile platforms (especially robots) almost impossible. This is achieved by using a completely new class of computing circuits developed at the University of Dayton over the last several years. This paper presents the developed novel compute circuits and systems that allow adaptive control algorithms to be implemented at high speeds and several orders of magnitude lower power than traditional computers using nanoscale devices known as the memristor. Keywords– Adaptive Controls, Low power architecture; Memristor crossbars; Deep neural network
Publication Date
4-5-2017
Project Designation
Graduate Research - Graduate
Primary Advisor
Tarek M. Taha
Primary Advisor's Department
Electrical and Computer Engineering
Keywords
Stander Symposium project
Recommended Citation
"MEMRISTOR-BASED NEURAL LEARNING FOR ADAPTIVE CONTROL SYSTEMS" (2017). Stander Symposium Projects. 985.
https://ecommons.udayton.edu/stander_posters/985
![MEMRISTOR-BASED NEURAL LEARNING FOR ADAPTIVE CONTROL SYSTEMS](https://ecommons.udayton.edu/stander_posters/1985/thumbnail.jpg)