Linus Michaeli

ANN-based error reduction for experimentally modeled sensors

A method for correcting the effects of multiple error sources in differential transducers is proposed. The difference in actual characteristics of the sensing elements of the differential scheme, and an easily controllable auxiliary quantity (e.g. supply voltage of conditioning circuit) provide independent information for the correction. This is carried out by a nonlinear multidimensional inverse model of the transducer based on an artificial neural network. Experimental results of the correction of a variable-reluctance displacement transducer, subject to the combined interference of structural and geometrical parameters, highlight the effectiveness of the proposed method.

Influence of the architecture on ADC error modeling

The influence of the architecture on analog-to-digital converter modeling is investigated for the three most widespread architectures: integrating, successive approximations, and flash. The effects of main error sources are analyzed in terms of integral and differential nonlinearity with the aim of setting up a unified error model. Such a model is useful both to economically generate a look-up table for error correction and to quickly produce diagnosis models for fault detection and isolation. Numerical simulations aimed to show the model effectiveness and experimental tests carried out to validate the model are discussed.

Automatic and accurate evaluation of the parameters of the magnetic hysteresis model

This paper presents a method based on both Artificial Neural Networks (ANNs) and on a multidimensional optimisation procedure in order to significantly reduce the time taken and to improve the accuracy in evaluating parameters of the Jiles-Athertons model of magnetic hysteresis. The main steps of the method can be individualised as: (i) data acquisition of the experimental hysteresis loop of the magnetic material under test, (ii) evaluation of the models parameters by means of the ANN, and (iii) parameter accuracy improvement by means of a multidimensional optimisation procedure. In order to highlight the methods effectiveness, the results of numerical and experimental tests are also given.

Measurement of the exponential signal distortion

In the ADC test is convenient to use as stimulus the real exponential voltage than the sinusoidal one. Indeed, the real exponential voltage is much more similarity to the theoretical one than the sinusoidal signal. For the estimation of the final uncertainty in the ADC test, the knowledge of the distortion of the exponential stimulus needs. The distortion is caused by the exponential signals superimposed to the waited one. With this aim, in the paper is pointed out (i) the stand for the accurate acquisition of the real exponential voltage generated by capacitor discharge, and (ii) the numerical procedure to evaluate the parameters of the assigned number nP of exponential functions that best fit the acquired samples. Results of tests assess that the reconstruction error is lower than the resolution of the acquisition stand.

An Improved ADC Error Correction Scheme Based on a Bayesian Approach

The paper presents an improved method for the ADC nonlinearity correction based on a Bayesian filtering approach. In particular, the dependence of the previous method version on the statistic characterization of the input signal has been removed. The proposed improvement has been validated in simulation using behavioural models provided by an ADC producer and on actual ADCs

Systematic error correction for experimentally modeled sensors by using ANNs

The paper deals with the compensation of the systematic uncertainty of sensors subject to nonlinear and combined influence parameters. The compensation is based on a second sensor and a digital artificial neural network (ANN). This heuristic fully a-posteriori approach allows the twofold problems of (i) the complex mathematical modeling of the influence on the measurement, and (ii) the effective solution of the nonlinear model to be simultaneously bypassed. Experimental results of the characterization of a variable-reluctance proximity transducer highlight the effectiveness of the proposed compensation scheme.

BandPass Sigma-Delta Modulator for Capacitive Pressure Sensor

The paper deals with bandpass sigma-delta modulator (BP SDM), which is used for conversion of signal from capacitive pressure sensor. This approach is absolutely new and unique, because this kind of modulators is utilized only for wireless and video applications. The main advantage of BP SDM is due to its defined band. That is why it is resistant against offsets of its subcuircits. Another important advantage is low power consumption, since the BP SDM digitizes only narrow band instead of whole Nyquist band with similar dynamic range. The paper shows basic ideas of this approach and simulation results with ideal and real modulator. The main stages are implemented in switched-capacitor (SC) technique. The modulator layout is presented as well.

Integral Nonlinearity Correction Algorithm Based on Error Table Optimizing and Noise Filtering

Integral Nonlinearity Correction Algorithm Based on Error Table Optimizing and Noise Filtering The main purpose of this paper is to present the external correction of analog to digital converters (ADC) integral nonlinearity and quantization noise based on the look up table method (LUT) combined with the averaging and Wiener filtering and dithering method. The LUT compression and LUT precision effect are also studied.

Compensation of intrinsic nonlinearity of SAR ADCs

In successive approximation analog-to-digital conversion principle, in case of input signal variation less than the least significant bit (LSB) during the conversion time, an intrinsic dynamic nonlinearity arises. In this paper, a method for compensating such a nonlinearity is proposed. Theoretical fundamentals of the method are reported, by paying particular attention to the accuracy of the compensation. With this aim, the dynamic phase distortion related to the intrinsic dynamic nonlinearity is modeled firstly. Then, a compensation based on the maximization of the signal-to-noise ratio is applied. Simulation and experimental results show the method effectiveness in compensating the intrinsic dynamic nonlinearity and in incrementing the signal-to-noise ratio in actual working conditions.

Error Models of the Analog to Digital Converters

Abstract Error models of the Analog to Digital Converters describe metrological properties of the signal conversion from analog to digital domain in a concise form using few dominant error parameters. Knowledge of the error models allows the end user to provide fast testing in the crucial points of the full input signal range and to use identified error models for post correction in the digital domain. The imperfections of the internal ADC structure determine the error characteristics represented by the nonlinearities as a function of the output code. Progress in the microelectronics and missing information about circuital details together with the lack of knowledge about interfering effects caused by ADC installation prefers another modeling approach based on the input-output behavioral characterization by the input-output error box. Internal links in the ADC structure cause that the input-output error function could be described in a concise form by suitable function. Modeled functional parameters allow determining the integral error parameters of ADC. Paper is a survey of error models starting from the structural models for the most common architectures and their linkage with the behavioral models represented by the simple look up table or the functional description of nonlinear errors for the output codes.