Mixed-Mode Neural Circuit for Solving Linear Equations
A ‘mixed’-mode neural network is one in which the neuronal states are represented by voltages and the synaptic signals are conveyed by currents. This results in a lower complexity circuit since the synaptic resistances are not required. The present chapter discusses a mixed-mode variant of the voltage-mode linear equation solver presented in the previous chapter. The Differential Voltage Current Conveyor (DVCC) has been used as the analog building block to realize a voltage comparator with current outputs. Further, a digitally programmable version of the mixed-mode circuit is also presented and a mechanism to adjust the weights corresponding to the coefficients in the set of linear equations is also discussed. Effect of deviations from ideal device behaviour, like offsets in the DVCC and opamps, is also explored.
KeywordsEnergy Function Current Output Operational Amplifier Simultaneous Linear Equation Voltage Comparator
- 7.Rahman, S.A.: A nonlinear synapse neural network and its applications. PhD thesis, Department of Electrical Engineering, Indian Institute of Technology, Delhi, India (2007)Google Scholar
- 9.LMC7101A. National semiconductor inc. http://www.national.com/assets/en/tools/spice/LMC7101A.MOD. Last Accessed on 25 Oct 2012
- 11.Plummer, J.D., Deal, M.D., Griffin, P.B.: Silicon VLSI Technology: Fundamentals, Practice and Modeling. Prentice Hall, Upper Saddle River (2000)Google Scholar
- 12.Hu, J., Xu, T., Zhang, W., Xia, Y.: A CMOS rail-to-rail differential voltage current conveyor and its applications. In: Communications, Circuits and Systems, 2005. Proceedings. 2005 International Conference on, vol. 2. IEEE (2005)Google Scholar
- 13.Mahmoud, S.A.: Low voltage wide range CMOS differential voltage current conveyor and its applications. Contemp. Eng. Sci. 1(3), 105–126 (2008)Google Scholar