Abstract
In this paper, a new approach to smoothen the one-dimensional canonical piecewise-linear model is presented. Contrary to traditional smoothing strategies, the proposal preserves the algebraic form of the canonical representation and only adds terms (expressed by hyperbolic tangent and absolute value functions) that smoothly compensate the characteristic slope transitions in a piecewise-linear graph. In general terms, it is shown that the graph of any one-dimensional piecewise-linear function can be transformed into a smoother one by a graphical compensation at their knot points. Besides preserving the original form of the model, the proposal also allows controlling smoothness either globally (at all the breakpoints) or locally (by a different level at each breakpoint).
References
Chua L O and Deng A C 1988 Canonical piecewise-linear representation. IEEE Transactions on Circuits and Systems 35(1): 101–111
Julian P, Jordan M and Desagues A 1998 Canonical piecewise-linear approximation of smooth functions. IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications 45(5): 567–571
Jimenez-Fernandez V M, Jimenez-Fernandez M, Vazquez-Leal H, Filobello-Nino U A and Castro-Gonzalez F J 2016 Smoothing the high level canonical piecewise-linear model by an exponential approximation of its basis-function. Computacion y Sistemas 20(2): 227–235
Jimenez-Fernandez V M, Jimenez-Fernandez M, Vazquez-Leal H, Munoz-Aguirre E, Cerecedo-Nunez H H, Filobello-Nino U A and Castro-Gonzalez F J 2016 Transforming the canonical piecewise-linear modeling into a smooth-piecewise representation. SpringerPlus 5(1612): 1–7
Lazaro M, Santamaria I and Pantaleon C 2000 A smooth and derivable large-signal model for microwave HEMT transistors. In: Proceedings of the IEEE International Symposium on Circuits and Systems, pp. 713–716
Lazaro M, Santamaria I and Pantaleon C 2001 Smoothing the canonical piecewise-linear model: an efficient and derivable large-signal model for MESFET/HEMT transistors. IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications 48(2): 184–192
Klimaszewski J, Sklyar M and Korzen M 2017 Learning \(L^{1}\)-penalized logistic regressions with smooth approximation. In: Proceedings of the IEEE International Conference on Innovations in Intelligent Systems and Applications, pp. 126–130
Nicolae A 2018 PLU: the piecewise linear unit activation function. eprint arXiv:1809.09534
Baptista F D and Morgado-Dias F 2012 Automatic general-purpose neural hardware generator. Neural Computing and Applications 28(2017): 25–36
Carrasco-Robles M and Serrano L 2008 A novel CMOS current mode fully differential tanh(x) implementation. In: Proceedings of the IEEE International Symposium on Circuits and Systems, pp. 2158–2161
Nascimento I, Jardim R and Morgado-Dias F 2012 A new solution to the hyperbolic tangent implementation in hardware: polynomial modeling of the fractional exponential part. Neural Computing and Applications 23(2): 363–369
Chua L O and Den A C 1986 Canonical piecewise-linear modeling. IEEE Transactions on Circuits and Systems 33(5): 511–525
Chua L O, Desoer C A and Kuh E S 1987 Linear and nonlinear circuits. New York: McGraw-Hill
Leenaerts D and van Bokhoven W M G 1998 Piecewise linear modeling and analysis. Eindhoven: Kluwer Academic Publishers
Santoro R P 1995 Piecewise-linear modeling \(I-V\) characteristics with SPICE. IEEE Transactions on Education 38(2): 107–113
Sarmiento-Reyes A L, Hernandez-Martinez L, Gutierrez de Anda M A and Castro-Gonzalez F J 2010 Piecewise linear-based hybrid simulation of MOS and single-electron circuits. In: Proceedings of the IEEE Latin American Symposium on Circuits and Systems, pp. 69–72
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Jimenez-Fernandez, V.M. A graphical compensation approach for smoothing the one-dimensional canonical piecewise-linear model. Sādhanā 46, 175 (2021). https://doi.org/10.1007/s12046-021-01700-6
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12046-021-01700-6