Budimir Lutovac

In this paper we present a two-integrator, Bi- quadratic CCII filter, which consists of all-positive CCIIs with grounded components, which makes it suitable for integrated circuit realization. The circuit has low sensitivity to component tolerances, low output thermal noise, and the capability of orthog- onal tuning of filter parameters. A step-by step design procedure for a fourth-order band-pass filter with optimum dynamic range is proposed. Breadboard laboratory designs displayed excellent results.

The reconstruction of non-stationary signals with missing samples is a particularly challenging topic. The compressed sensing (CS) reconstruction requires that signals exhibit sparsity in a transformation domain. We perform the CS reconstruction of the linear frequency modulated (LFM) signals with a common chirp rate, usually appearing in ISAR imaging. To this aim, we apply the genetic algorithm (GA).

This paper presents simple, flexible and effective approximation method for fractional order lead-lag compensators. The proposed method relies on a Padé approximation of linear fractional order transfer functions, giving rational approximations of order N accurate enough for control applications as soon as N is greater than 2 or 3. An example of feedback loop incorporating this approximation is adopted from the well-known car suspension problem, wherein an iso-damping property of the closed loop response is achieved, with regard to a variation of the vehicle mass.

In this paper output noise is investigated for several fractional-order filters. Examples are given using the low-pass Butterworth current-mode active-RC filter approximations. It is demonstrated experimentally that as with integer-order filters, output noise decreases with decreasing order of fractional-order filters. The considered circuits use CCIIs and have two integrators in the feedback loop. The fractional-order filter has one fractional capacitance realized by a second Foster circuit. Other components are of integer order. The noise comparison is simulated with Pspice.

The paper presents analysis of the second order band-pass and notch filter with a dynamic damping factor βd of fractional order. Factor βd is given in the form of fractional differentiator of order a, i.e. βd=β/sa , where β and a are adjustable parameters. The aim of the paper is to exploit an extra degree of freedom of presented filters to achieve the desired filter specifications and obtain a desired response in the frequency and time domain. Shaping of the frequency response enables achieving a better phase response compared to the integer-order counterparts which is of great concern in many applications. For the implementation purpose, the paper presents a comparison of four discretization techniques: the Osutaloup's Recursive Algorithm (ORA+Tustin), Continued Fractional Expansion (CFE+Tustin), Interpolation of Frequency Characteristic (IFC+Tustin) and recently proposed AutoRegressive with eXogenous input (ARX)-based direct discretization method.

This paper presents an efficient optimization approach for the placement of electric vehicles charging stations within the road network. The approach is based on the integer linear pro ...

The paper introduces a novel the second order band-pass and notch filter with dynamic damping factor β _d of fractional order. Indeed, the factor β _d has a form of fractional differentiator of order α, i.e. β _d =β/s ^α , where β and α are adjustable parameters. Shaping of the frequency response enables achieving better phase response compared to the integer-order counterparts which is of great concern in many applications. The aim of the paper is to exploit an extra degree of freedoms of presented filters to achieve the desired filter specifications and obtain desired response in the frequency and the time domain.

Rounding error in the discrete Fourier transform calculated with fast fixed-point algorithms is considered. It is shown that the variance of the rounding error depends on frequency index. Theoretically obtained results for error variance are statistically checked on decimation-in-time fast Fourier transform with two rounding methods.

A method for a resistive circuit analysis based on graph spectral decompositions is proposed. It is shown that the Laplacian matrix can be used in order to calculate node potentials. Based on the Laplacian eigenvalues and eigenvectors it is possible to decompose a complex resistive circuit into smaller, weakly connected sub-circuits.