Coincidence and Auxiliary Channel Veto

Although the x2 veto reduces the rate of triggers from glitches, some glitch triggers survive. However, there are further tests that can be used to eliminate them. Most importantly, if more than one interferometer is involved in the search, one can require consistency between their triggers. LIGO is especially well designed in this regard. Because LIGO's three detectors are almost co-aligned, they should all be sensitive to the same signals (although the 2km H2 will only be half as sensitive to them as the 4km instruments). Thus, any signal that appears in one should appear in all. On the other hand, there is no reason for glitches in one instrument to be correlated with glitches in another, especially between a detector located in Washington state and the Louisiana instrument. Thus, one way to distinguish between triggers generated by actual gravitational waves and those generated by glitches is to demand coincidence between triggers in different instruments [85,86].

The most fundamental coincidence is coincidence in time. The timing precision of the matched filtering for properly conditioned interferometric data is ms. The larger effect is the time it could take the gravitational wave to traverse the distance between detectors (i.e. the light travel time between them). The maximum time delay for this is also measured in ms (e.g. 10ms between the Washington site and the Louisiana site for LIGO). Thus, triggers at one site which are not accompanied by triggers at another site within the light travel time plus 1 ms are likely not gravitational waves and can be discarded. For triggers from instruments which are sufficiently well aligned, there are several other quantities for which one could required coincidence. Of these, the only one which had been applied to date as a trigger veto is the template which generated the trigger - for the LIGO S2 search the same template was required to have generated all coincident triggers (or represent all coincident clusters of triggers) or they were discarded.

The final hurdle that a trigger may have to overcome to remain viable is that it not be associated with a known instrumental disturbance. Auxiliary channels which monitor the instruments and their environment contain information about many potential disturbances. Those channels most likely to correspond to spurious disturbances which would be manifest in the gravitational wave channel have been studied intensively. To date, studies have revealed that channels which allows for safe and useful auxiliary channel vetoes are not forthcoming for most instruments. However, for LIGO's second science run it was discovered that a channel which measures length fluctuations in a certain optical cavity of the L1 interferometer had glitches which were highly correlated with glitches in that instrument's gravitational wave channel [87]. Thus, triggers which occurred within a time window 4 seconds earlier to 8 seconds later than a glitch in this auxiliary channel in Louisiana were also discarded for that analysis.

Finally, triggers which survive all of these cuts must be examined individually to determine if they are genuine candidates for gravitational wave signals. The flow chart for the procedure we have just describe is shown in Fig. 5. A more comprehensive and detailed discussion of such an algorithm can be found in [88]. Details of the pipelines used by LIGO for the S1 and

Fig. 5. A diagram showing a typical work-flow to look for signals from neutron star binaries. This is a two instrument work-flow for LIGO involving one detector from Washington and the Louisiana detector. It is similar to the one used for the S2 analysis. There would a slightly more complicated work-flow diagram when all three instruments were used. Note that the Washington data is only analyzed at times and for binary templates that correspond to a trigger in Louisiana, thus minimizing both false alarms and computing time. The + and - signs by the two bottom-most arrows indicated that the addition and subtraction of coincident triggers respectively - i.e. passing the SNR and \2 thresholds adds coincident triggers, while occurring during an auxiliary channel glitch removes them.

Fig. 5. A diagram showing a typical work-flow to look for signals from neutron star binaries. This is a two instrument work-flow for LIGO involving one detector from Washington and the Louisiana detector. It is similar to the one used for the S2 analysis. There would a slightly more complicated work-flow diagram when all three instruments were used. Note that the Washington data is only analyzed at times and for binary templates that correspond to a trigger in Louisiana, thus minimizing both false alarms and computing time. The + and - signs by the two bottom-most arrows indicated that the addition and subtraction of coincident triggers respectively - i.e. passing the SNR and \2 thresholds adds coincident triggers, while occurring during an auxiliary channel glitch removes them.

S2 analyses can be found in [89] and [90] respectively. In the next section, we discuss how the surviving candidates are analyzed and how upper limits are set from them.

Was this article helpful?

0 0

Post a comment