Multi-tone signals are discrete-spectrum signals with spectral components located at non-harmonically related frequencies. Their spectral analysis can be performed in short observation intervals by means of a method that combines the use of a virtual time domain and statistical techniques. The contributions to uncertainty that specifically arise from the implementation of the method are discussed in the paper and a study is developed to optimize the measurement technique in the case of bi-tone, periodic or not, signals. Finally, experimental results, which confirm theory, are reported and...
This paper presents a measurement method for measuring temperature values of test gears on a test bed. For better understanding of the wear behaviour of worm gears and for analysing the contact conditions between the flanks of worm and worm wheel in presence of lubricating oil, it is necessary to know the temperature of the contact area. The oil between the flanks is heated by the hot flanks and therefore the viscosity of the lubricating oil is reduced. For this reason the temperature curve of the contact area and thus the oil temperature between the flanks was measured and afterwards...
The universal solution for precise virtual instrument system realisation permitting to connect any frequency-time domain (quasi-digital) sensor to IBM PC compatible computers is described in the paper. The virtual instrument system for smart sensors with frequency-time domain output is based on the novel adaptive program-oriented method for frequency-to-code conversion – method of dependent count. The main advantages of this method for virtual sensor instrumentation applications are: high accuracy, constant quantization error in all frequency range, non-redundant conversion time, minimum...
In this paper we present an experimental method that is suitable to ascertain the degree of nonlinearity present in a given signal as well as its adequacy to be used as a stimulus signal in a statistical method, such as the histogram test of ADCs. We show one possible implementation of the method as well as its theoretical substrate and some preliminary experimental results. Finally we show the importance of determining the distribution of the stimulus signal when using noise.
Signal-processing applications frequently use least-square sine fitting algorithms. The parameter estimation provided by these algorithms is exposed to errors due to different causes. Good results in sinewave estimation may be achieved when the frequency is unknown by applying the new method presented in this paper.
A simple technique for the restoration of color images degraded by lossy compression is presented. The proposed approach deals with the YIQ color space and processes the luminance component by means of a recursive algorithm based on fuzzy models. This design choice combines effectiveness and simplicity. Indeed, the method can satisfactorily reduce quantization errors produced by lossy compression such as the popular JPEG technique. On the other hand, the tuning process is very fast because it requires one parameter only.
A virtual instrument for the remote control of measurement instruments has been realised. It enables the user to control the characteristic parameters of the instruments, to start the measurement process and to acquire the obtained results. At this step the implemented approach can enhance the availability of the lab.
The paper describes a virtual instrument designed to fast and accurately identify systems working in a large range of frequencies. It comprises both the hardware aimed on input-output data generation and acquisition and the software dedicated to signal processing and all the calculus ended with the model estimation. The instrument offers all the facilities provided by the LabVIEW environment, being also interfaced with the powerful toolboxes of Matlab.
The paper presents the design and implementation of an automated system for the analysis of gas mixtures. Using low-cost, non-selective gas sensors in combination with signal processing algorithm based on artificial neural networks, the prototype system is able to correctly classify three combustible gases (methane, isobutane, and hydrogen) and to indicate when the total gas concentration in the air exceeds a preset alarm point. The system is designed using LabVIEW and virtual instrumentation concept and is capable of performing online analysis of gas mixtures.
The possibility to predict the iron losses under nonsinusoidal waveform of magnetic flux density using neural network based magnetic models is analyzed in this paper. Two different types of model structures (predictors with and without feedback) have been used. The predictions are compared to the measured output for different magnetic materials and induction waveforms.