Monte Carlo methods are used to compare the performance of four non-linear estimators when applied to the logarithmic pH non-linearity. The estimators are: direct inversion of the non-linearity, Bayes'rule, extended Kalman filter, statistical linearisation. It is found that the unbiased optimal estimates of Bayes' rule are in many cases significantly better than the other estimates. It is concluded that the Bayes algorithm could be useful for practical purposes.
Get full access to this article
View all access options for this article.
References
1.
Bucy, R.S. and Senne, K.D.1971. 'Digital synthesis of non-linear filters', Automatica, 7,287-298.
2.
Gelb, A. (ed.). 1974. Applied optimal estimation, MIT Press.
3.
Jacobs, O.L.R. (1984). 'pH Regulation', Pergamon Encylcopaedia of Systems and Control (to be published).
4.
Jacobs, O.L.R., Badran, W.A. and Proudfoot, C.G.1984. 'CAD of systems for regulating pH', to be published in TheChemical Engineer, Mar, 19- 21.
5.
Jacobs, O.L.R., Bullingham, R.E.S., Lammer, P., McQuay, H.J., O'Sullivan, G. and Reasbeck, M.P.1983. 'On-line estimation in the relief of post-operative pain' , Oxford University Engineering Laboratory Report No. 1452/83, to be published in Automatica.
6.
Jacobs, O.L.R., Hewkin, P.F. and While, C.1980. 'On-line computer control of pH in an industrial process' , Proc. IEE, 127D, 161-168.
7.
Jazwinski, A.H.1970. Stochastic Processes and Filtering Theory, Academic Press.
8.
Orava, P.J. and Niemi, A.J.1974. 'State Model and Stability Analysis of a pH Control Process' , Int J Cont, 20,557-567.
9.
Sorenson, H.W. and Alspach, D.L.1971. 'Recursive Bayesian Estimation using Gaussian sums', Automatica, 7,465-479.
