Hannu Olkkonen

Efficient lifting wavelet transform for microprocessor and VLSI applications

Recently, wavelet analysis has gained an established role in signal and image processing applications. In this paper, we present a discrete wavelet transform algorithm based on the lifting scheme. The Haar wavelet transformed and decimated signals are lifted by the ladder-type network. A unique lifting filter is designed for in-place computation. The present algorithm is especially suitable for microprocessor and VLSI applications since it can be implemented by integer arithmetics using only register shifts and summations.

Hilbert transform assisted complex wavelet transform for neuroelectric signal analysis

In this work, we present a new approach for shift invariant complex wavelet analysis of neuroelectric signals. A key idea is to preprocess the signal with the Hilbert transformer to yield an analytic signal, which is then wavelet transformed using the linear phase complex scaling and wavelet filters. In different scales, the total energy of the wavelet transform coefficients is shift invariant. The decimated analytic wavelet coefficients suffer no aliasing effects, which are predominant in conventional wavelet analysis. We show the usefulness of the present method in multi-scale analysis of the neuroelectric signal waveforms.

Microcomputer-aided measurement of puff parameters during smoking of low- and medium-tar cigarettes

A portable microcomputer-assisted flow transducer was developed for analysing puff parameters during smoking of low- and medium-tar cigarettes. Smoke flow was determined by measuring pressure difference between two sites within an orifice flowmeter. According to the Bernoulli equation, the pressure difference is proportional to the square of flow. For calibration of the method, various sizes of air volumes were puffed through the flowmeter by a piston syringe. The calibration curve, which consisted of the flow as a function of the square root of pressure difference, was linear (r = 0.98). The automatic microcomputer analysis consists of the following variables: mean flow and mean volume of inhaled smoke gas, puff duration, time interval between two puffs, number of puffs and total volume inhaled. Eight volunteers smoked 10 low- and 10 medium-tar cigarettes during the cross-over experiments. The investigation indicated that the total inhalation volume of smokers in the case of low-tar cigarettes is twice as large as in the case of medium-tar cigarettes.

Measurement and Reconstruction of Impulse Train by Parallel Exponential Filters

In many areas of science and technology, the measurement of impulse trains is an important sampling scheme. For example, wireless data transmission has grown enormously due to the recent developments in ultra wide band technology, where the information is represented as impulse trains. In this letter, we present a new approach for sampling of the impulse train using parallel exponential filters (EFs), whose outputs are measured simultaneously. We show that with parallel EFs, it is possible to reconstruct impulses with different amplitudes and appearance times within one sampling period. The reconstruction algorithm is manageable on microprocessor and VLSI circuits. Several applications of the parallel EF network are outlined.

Fractional Delay Filter Based on the B-Spline Transform

Fractional delay filters (FDFs) have a key role in communication systems. FDFs produce a delay that is a fraction of the sampling period. In this letter, we introduce a framework based on the B-spline interpolation and decimation procedure for design of the FDFs. The method generates precise fractional delays and is easy to implement in microprocessor and VLSI environments

Sampling and Reconstruction of Transient Signals by Parallel Exponential Filters

This brief introduces a new method for sampling of transient analog waveforms based on the parallel exponential filters. The signal is fed to the parallel network consisting of resistor-capacitor (RC) circuits, outputs of which are simultaneously sampled. We show that N previous samples of the input signal can be reconstructed from single output samples of N parallel RC circuits. The parallel sampling method increases the sampling rate of the data acquisition system by a factor of N. In particular, the method is useful in increasing the sampling rate of the Flash-type analog-to-digital VLSI circuits. We present the parallel RC network, develop the reconstruction algorithm, and briefly describe a variety of applications such as measurement and reconstruction of pulses produced by ultrawideband transmitters, radiation detectors, and pulse lasers.

Fractional Time-Shift B-Spline Filter

The fractional time-shift operators are essential devices in processing and adjustment of signals in modern communication systems. This letter introduces the fractional time-shift B-spline filter. The filter produces a time-shift Delta isin [0,1], which is an arbitrary fraction of the sampling period. We describe the construction and parallel implementation of the fractional time-shift filter. We give several of its applications (differentiator, integrator, and correlator) in numerical and statistical signal processing and in adaptive time-shift adjustments.

Computation of Hilbert Transform via Discrete Cosine Transform

Hilbert transform (HT) is an important tool in constructing analytic signals for various purposes, such as envelope and instantaneous frequency analysis, amplitude modulation, shift invariant wavelet analysis and Hilbert-Huang decomposition. In this work we introduce a method for computation of HT based on the discrete cosine transform (DCT). We show that the Hilbert transformed signal can be obtained by replacing the cosine kernel in inverse DCT by the sine kernel. We describe a FFT-based method for the computation of HT and the analytic signal. We show the usefulness of the proposed method in mechanical vibration and ultrasonic echo and transmission measurements.

Discrete Lattice Wavelet Transform

The discrete wavelet transform (DWT) has gained a wide acceptance in denoising and compression coding of images and signals. In this work we introduce a discrete lattice wavelet transform (DLWT). In the analysis part, the lattice structure contains two parallel transmission channels, which exchange information via two crossed lattice filters. For the synthesis part we show that the similar lattice structure yields a perfect reconstruction (PR) property. The PR condition can be used to design half-band filters, which effectively eliminate aliasing in decimated tree structured wavelet transform. The DLWT can be implemented directly to any of the existing DWT algorithms

Half-Delay B-Spline Filter for Construction of Shift-Invariant Wavelet Transform

We introduce a general framework for the shift-invariant biorthogonal wavelet transform. The method is based on the two parallel wavelet transforms, where the wavelets form a Hilbert transform pair. This condition requires that the impulse responses of the scaling filters are half-delayed versions of each other:ho [n] and ho [n-1/2]. The ideal half-delay operator is constructed by the interpolation and decimation procedure based on the polyphase decomposition of the two-scale B-spline equation. The present method yields linear phase and shift-invariant wavelet transform coefficients and can be adapted to any of the existing biorthogonal DWT filter bank.