Using analog neutral networks for control sensor linearization

Dempsey, G.L.; Alig, J.S.; Redfield, D. E.

doi:10.1109/mwscas.1995.504381

Cited by 3 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Artificial neural network (ANN) based curve fitting techniques invite the measurement community to make use of them for linearization. ANN-based hardware and software linearizers have also been reported [16][17][18][19]. Among software-based techniques recently proposed [20][21][22][23][24], not many of them are found to be simple in operation, yielding poor efficiency, and some of them are efficient in performance but suffer from severe computational burden, which thereby lead to difficulties in implementation.…”

Section: Introductionmentioning

confidence: 99%

An evolutionary optimized nonlinear function to improve the linearity of transducer characteristics

Abudhahir¹,

Baskar²

2008

Meas. Sci. Technol.

View full text Add to dashboard Cite

This paper proposes a nonlinear optimal-function-based algorithm which can be utilized to replace electronic circuitry traditionally employed to linearize the characteristics of commonly used temperature transducers such as resistance temperature detectors, thermistors and thermocouples. The function exploits ratiometric-logarithmic operation for linearization. The optimal parameters of the function are determined using a covariance matrix adopted evolutionary strategy (CMAES) algorithm. Transducers' input–output data are derived from the Yokogawa handy calibrator model CA 150 and subjected to the proposed algorithm to evaluate the performance of the method. The performance measures such as full-scale error and mean square error are considered to compare the performance of the proposed technique with other methods reported for transducers. The present linearization algorithm was implemented using LabVIEW 7.1 Professional Development System in a personal computer that provides the facility to interface with the National Instruments data acquisition module NI DAQCard PCI-6221. Experimental results reveal that the proposed evolutionary optimized nonlinear function based software linearizer does its job efficiently in a better way than that of the conventional hardware and software methods. Also, the results obtained using the CMAES algorithm are compared with the results of a real-coded genetic algorithm. The comparison shows that the CMAES algorithm is more consistent in determining the best solution for the proposed ratiometric-logarithmic function with reasonable computation time.

show abstract