A 4-Term Exponential-Quadratic Approximation for Gaussian Q or Error Functions Accurate to $1.65\times 10^{-4}$

Graham Pulford


Integrals on $[0,\infty)$ where the integrand is of the form $Q^m(a\sqrt{x})\,{\rm p}(x)$, where $Q$ is the Gaussian $Q$ function, ${\rm p}(\cdot)$ a Gamma PDF, $m$ a positive integer and $a>0$; or of the form $\erf^m(ax+b)\,x^n\exp(-c^2x^2+2dx)$, where $\erf(x)$ is the error function, and $n$ a non-negative integer, arise in performance modelling of communication and machine learning systems. Such integrals cannot be evaluated analytically in general, but they are reducible to a key integral whose integrand is $\erf(ax+b)\,{\rm N}(x;m,s)$ where ${\rm N}()$ is a Gaussian PDF with mean $m$ and variance $s$. Seeking an efficient and accurate evaluation method, we develop a new 4-term exponential quadratic approximator (EQA) for the error function that includes both linear and quadratic terms in its exponents. The EQA minimises a sum-of-squares cost function with two ``spline-type'' constraints, i.e., constraints on the function value and its first derivative. This 12-parameter constrained optimisation problem is reduced to an unconstrained 8-parameter one by inverting a 4-D linear system, then solved by gradient descent. The resulting approximator has a maximum absolute error of $1.65\times 10^{-4}$ on the real line, and outperforms many other exponential sum approximators for $\erf(x)$ on $x\in[0,1.5]$ and for $Q(x)$ on $x\in[0,2]$. Moreover, due to its functional form, the EQA leads to an analytical formula for the key integral that is accurate to 3 to 4 significant figures when compared with known special cases or computationally intensive Monte Carlo integration. The EQA can equally be used to obtain closed forms for the average symbol error probability of various modulation schemes on Rayleigh fading channels.

Full Text:



C. Hastings, Jr.

Approximations for Digital Computers, Princeton, NJ: Princeton University Press, pp. 167--169, 1955.

C. W. Clenshaw.

Chebyshev series for mathematical functions, National Physical Laboratory Mathematical Tables, vol. 5, Her Majesty's Stationery Office, London, 1962.

W. J. Cody.

``Rational Chebyshev approximations for the error function'', Mathematics of Computation, Vol. 23, No. 107, pp. 631--637, July 1969.

J. L. Schonfelder.

``Chebyshev expansions for the error and related functions'', Mathematics of Computation. vol. 32, no. 144, pp. 1232--1240, 1978.

M. Abramowitz and I. A. Stegun.

(eds.) Handbook of Mathematical Functions, National Bureau of Standards, Washington, D.C., 1964.

S. Chevillard.

The functions erf and erfc computed with arbitrary precision and explicit error bounds. Information and Computation, 216:72--95, 2012.

Marvin K. Simon.

Some new twists to problems involving the gaussian probability integral. IEEE Trans. on Communications, 46(2):200--210, 1998.

M. K. Simon.

``Single integral representations of certain integer powers of the Gaussian Q-function and their application'', IEEE Commun. Letters, vol. 6, no. 12, pp. 532--534, 2002.

M. K. Simon.

``A simpler form of the Craig representation for the two-dimensional joint Gaussian Q-function'', IEEE Commun. Letters, vol. 6, no. 2, pp. 49--51, 2002.

R. M. Radaydeh and M. M. Matalgah.

Results for integrals involving m-th power of the Gaussian Q-function over rayleigh fading channels with applications. Int. Conf. on Communications, pages 5910--5914, 2007.

I. S. Gradshteyn and I. M. Ryzhik.

Table of Integrals, Series and Products. Academic Press, 7th edition, 2007.

H. A. Fayed and A. F. Atiya.

``An evaluation of the integral of the product of the error function and the normal probability density with application to the bivariate normal integral'', Mathematics of Computation, vol. 83, no. 285, pp. 235--250, 2013.

C.R. Selvakumar.

``Approximations to complementary error function by method of least squares'', Proceedings of the IEEE, vol. 70, no. 4, pp. , 1982.

S. A. Dyer and J. S. Dyer.

``Approximations to Error Functions'', IEEE Instrumentation & Measurement Magazine, vol. 10, no. 6, pp. 45--48, 2007.

S. Aggarwal.

``A Survey-cum-Tutorial on Approximations to Gaussian Q Function for Symbol Error Probability Analysis Over Nakagami- m Fading Channels'', IEEE Communications Surveys & Tutorials, vol. 21, no. 3, pp. 2195--2223, 2019.

V. N. Q. Bao, L. P. Tuyen, and H. H. Tue.

A Survey on Approximations of One-Dimensional Gaussian Q-Function. REV Journal on Electronics and Communications, 5(1-2):1--14, 2015.

P. Borjesson and C. E. Sundberg.

``Simple Approximations of the Error Function Q(x) for Communications Applications'', IEEE Trans. on Communications, vol. 27, no. 3, pp. 639--643, 1979.

W. M. Jang.

``A simple upper bound of the Gaussian Q-function with closed-form error bound'', IEEE Communications Letters, vol. 14, no. 2, p. 157, Feb. 2011.

G. K. Karagiannidis and A. S. Lioumpas.

``An improved approximation for the Gaussian Q-function'', IEEE Commun. Lett., vol. 11, no. 8, pp.644--646, Aug. 2007.

P. Loskot and N. C. Beaulieu.

``Prony and polynomial approximations for evaluation of the average probability of error over slow-fading channels'', IEEE Trans. Vehicular Technol., vol. 58, no. 3, pp. 1269--1280, Mar. 2009.

L. M. Benitez and F. Casadevall.

``Versatile, accurate, and analytically tractable approximation for the gaussian Q function'', IEEE Trans. on Communications, vol. 59, no. 4, pp. 917--922, Apr. 2011.

Marco Chiani, Davide Dardari, and Marvin K. Simon.

New exponential bounds and approximations for the computation of error probability in fading channels. IEEE Trans. on Wireless Communications, 2(4):840--845, 2003.

P. C. Sofotasios and S. Freear.

``Novel expressions for the Marcum and one dimensional Q-functions'', Proc. Int. Conf. on Wireless Info. Techn. & Systems (ICWITS'10), pp. 736--740, 2010.

A. Annamalai, E. Adebola and O. Olabiyi.

``Simple Closed-Form Approximations for the ASER of Digital Modulations over Fading Channels'', International Conference on Wireless Networks, pp. 1--7, 2012.

P. Dao Ngoc, U. Nguyen Quang, H. Nguyen Xuan, and R. McKay.

``Evolving approximations for the gaussian Qfunction by genetic programming with semantic based crossover'', Congress on Evolutionary Computation, pp. 1--6, 2012.

P. Van Halen.

``Accurate analytical approximations for error function and its integral'', Electronics Letters, vol. 25, no. 9, pp. 561--563, 1989.

R. M. Howard.

``Arbitrarily Accurate Analytical Approximations for the Error Function'', Technical Report, School of Elec. Eng. Comp. & Math. Sci., Curtin University, Perth, Australia, pp.1--44, Dec. 2020.

I. M. Tanash and T. Riihonen.

``Global minimax approximations and bounds for the Gaussian Q-function by sums of exponentials'', IEEE Trans. on Communications, vol. 68 no. 10, pp. 6514--6524, 2020.

C. de Boor.

A Practical Guide to Splines, Springer-Verlag, 1978.

E. W. Ng and M. Geller.

A table of integrals of the error functions. J. Research of the National Bureau of Standards, 73B(1):1--20, 1968.

A. P. Prudnikov, Y. A. Brychkov, and O. I. Marichev.

Integrals and Series, Volume 2 - Special Functions. Gordon & Breach Science Publishers, New York, 1986.

N. E. Korotkov and A. N. Korotkov.

Integrals Related to the Error Function. Chapman and Hall, 1st edition, 2020.

DOI: http://dx.doi.org/10.21553/rev-jec.276

Copyright (c) 2022 REV Journal on Electronics and Communications

Copyright © 2011-2022
Radio and Electronics Association of Vietnam
All rights reserved