cavalier fabien on behalf lal group orsay
DESCRIPTION
Reconstruction of Source Location using the Virgo-LIGO network. Presentation of the method Toy Simulation Influence of timing resolution and Effects of systematic errors Geometrical Properties of Virgo-LIGO network Effects of the Beam-Pattern functions. Cavalier Fabien - PowerPoint PPT PresentationTRANSCRIPT
Cavalier Fabienon behalf LAL groupOrsay
GWDAW 10December, 14th 2005
Reconstruction of Source Location using the Virgo-LIGO network
• Presentation of the method• Toy Simulation• Influence of timing resolution and Effects of systematic errors• Geometrical Properties of Virgo-LIGO network• Effects of the Beam-Pattern functions
• With n measured arrival times ti with associated errors i, compute the best estimation of and
• Use of 2 Minimization
• 2 defined as:
where t0 is the arrival time of the signal at the center of the Earth and ti
Earth(,) is the delay between the ith ITF and the center of the Earth
Reconstruction Method
n
i i
iiEarthttt
t1
2
20
0
)))(((,(
• Symmetrical Definition
• No reference detector to compute timing differences
How to choose it ?• detector with lower error• detector leading to highest time delays• detector which gives the best relative errors on time delays• …
• Uncorrelated Errors
• Easily expendable to any set of detectors
Reconstruction Method
• Put a source at a given location
• Computes arrival times tTrue[i]
• Computes timing errors in each detector • set [i] = 10-4 s in all detectors (no beam-pattern effect)• set <SNR> = 10 (beam-pattern effect)and use [i] = / SNR[i] ms (0 = 1 ms will be used by default)
• tMeasured[i] = tTrue[i] + GaussianRandom * [i]
• Compute the angular distance (Angular Error in the following plots) between reconstructed location and true one
Toy Simulation
Example of Reconstruction for Galactic Center
Estimated errors (through covariance matrix) in agreement with RMS values
Effect of Systematic Errors on Arrival Time
tMeasured[i] = tTrue[i] + GaussianRandom * [i] + Bias (for only one ITF)
• bias in the angular reconstruction when the timing bias and [i] have the same order of magnitude • bias in the angular reconstruction proportional to the timing bias• no significant difference between the three ITF
Influence of Timing Resolution
msmsorAngularErrorAngularErr
1)1()( 0
0
Reconstruction for GC (no beam-pattern)
Mean Angular Error: 1.9Median Angular Error: 0.95o
Minimal Angular Error: 0.8o
Maximal Angular Error: 4.3o
Reconstruction for =0o (no beam-pattern)
Mean Angular Error: 1.8o
Median Angular Error: 1.5o
Minimal Angular Error: 1.3o
Maximal Angular Error: 3.1o
Error increase when source crosses 3-detector plane
• 2 detectors located at (±d/2,0,0)• source in the (x,y) plane defined by its angle with x axis
• t21 = d/c cos() (1)• if error t on t21, then error on the angle: = c/d t / |sin()|
• when approaches 0, due to statistical errors, Eq.(1) cannot be inverted• = 1/ t2 (t21
measured – d/c cos())2
• minimal and equal to 0 for = acos(t21measured *c/d)
if |t21measured *c/d|≤1
• minimal and 1 for = 0 if |t21measured *c/d|>1
• Similar effect in the 3-detector case
Reconstruction for full sky (no beam-pattern)
Mean Angular Error: 1.6o
Median Angular Error: 1.1o
Minimal Angular Error: 0.7o
Maximal Angular Error: 4.5o
Reconstruction for GC (beam-pattern effect included)
Reconstruction for GC (beam-pattern effect included)
All events:
Mean Angular Error: 4.0o
Median Angular Error: 1.8o
Minimal Angular Error: 0.7o
Events with all SNR > 4.5 (56%):
Mean Angular Error: 1.8o
Median Angular Error: 1.25o
Minimal Angular Error: 0.7o
Reconstruction for =0o (beam-pattern effect included)
Reconstruction for =0o (beam-pattern effect included)
All events:
Mean Angular Error: 3.0o
Median Angular Error: 2.2o
Minimal Angular Error: 1.2o
Events with all SNR > 4.5 (79%):
Mean Angular Error: 2.1o
Median Angular Error: 1.7o
Minimal Angular Error: 1.2o
Reconstruction for full sky (beam-pattern effect included)
All events:
Mean Angular Error: 2.7o
Median Angular Error: 1.7o
Minimal Angular Error: 0.7o
Events with all SNR > 4.5 (60%):
Mean Angular Error: 1.8o
Median Angular Error: 1.3o
Minimal Angular Error: 0.7o
Conclusion
• 2 minimization works properly for position reconstruction
• In the case of 3 detectors, the method is able to find the two possible solutions
• The method is easily extendable to any ITF network
• Errors on parameters provided by the minimization procedure
• Accuracy of one degree can be achieved with good timing reconstruction (error proportional to timing error)
• Bias in timing estimator can easily introduce bias in reconstructed angles