Time domain amplitude and frequency detection of gravitational waves from coalescing binaries

Abstract

We propose a multistep procedure for the on-line detection and analysis of the gravitational wave signals emitted during the coalescence of compact binaries. This procedure, based on a hierarchical strategy, consists of a rough on-line analysis of the gravitational wave signal using adaptive line enhancers filters and a fast off-line parameter estimate, using the controlled random search optimization algorithm. A more refined off-line analysis using the classic matched-filtering technique, with a greatly reduced computational burden, can follow to further improve the parameter estimate. The results of simulations for the rough analysis are quite promising both for the relatively small computational power needed and for the robustness of the algorithms used, so that it could be very helpful for gravitational wave detection with very large baseline interferometric detectors such as LIGO and VIRGO

Adaptive filters for detection of gravitational waves from coalescing binaries

“…We briefly note here that this filter structure is different from the one reported in our previous works [9,10] that exhibited a magnitude response larger than 1 around the central frequency, which worsened its noise reduction capability.…”

Genetic approach helps to speed classical Price algorithm for global optimization

“…Moreover the analysis of clusters evolution during the optimization iterations gives more valuable information on problem features. As referred in [2,1,18] where numerical results are reported original Price's algorithm performs well in the global search and points generated by the algorithm clusters around minima of the problem. Nevertheless it is not able to perform fast enough local optimization to find global minimum starting from a good estimate generated at an intermediate step.…”

“…Such property is also retained in heuristic algorithms that preserve a random generation of new points. In the case of Price's algorithm (see [18] for a simple proof of this property of standard Price algorithm) and also in the improved version [5] new points are generated along a direction randomly chosen on a finite set depending on past evaluated points, so that global convergence cannot be assured. In practical instances the ability of CRS methods to find the global minimum relies in a sufficiently large size of the initial set of points chosen at random and on the hope that the global optimum point do not have a too small region of attraction.…”

“…In such instances, as pointed out in [3,18] and in [5], most adapted methods are the controlled random search (CRS) ones. Such family of methods tends to solve the global optimization problems by following at the same time two strategies.…”

Dynamic matched filters for gravitational wave detection

“…This is very useful to get out of the local maxima. The original Price algorithm does not define precise stop criteria, which can be defined in an effective way by the user [13].…”