ICANN '09 Proceedings of the 19th International Conference on Artificial Neural Networks: Part I
Superior-Order Curvature-Correction Techniques for Voltage References
Superior-Order Curvature-Correction Techniques for Voltage References
A built-in-test circuit for RF differential low noise amplifiers
IEEE Transactions on Circuits and Systems Part I: Regular Papers
A Gaussian synapse circuit for analog VLSI neural networks
IEEE Transactions on Very Large Scale Integration (VLSI) Systems
FGMOS transistor based low voltage and low power fully programmable Gaussian function generator
Analog Integrated Circuits and Signal Processing
| Hi-index | 0.00 |
An original realization of a CMOS Gaussian function generator is presented. The proposed method is based on a new approximation function that is able to fourth-order match the Gaussian function. The current-mode operation of the circuit strongly reduces the technological-caused errors and the errors introduced by temperature variations, with the result of an important increasing of the accuracy for the squaring circuit that represents the functional core of the Gaussian generator (0.1%). Additionally, the bandwidth of the generator is increased as a result of its current-mode operation. Because of the utilization of the new fourth-order approximation function, the deviation from the ideal Gaussian function is smaller than 1dB for an extended range of the input variable. The circuit is designed for implementing in 0.18@mm CMOS technology, its proposed architecture being compatible with a low-voltage operation (V"D"D=1V). The proposed Gaussian function generator based on the new approximation function allows to extend its capability of generating any continuous mathematical functions, this feature being obtained by changing the approximation function coefficients.