Direction-of-arrival estimation of a gravitational wave by correlations between quadrupole moments of pulsar timing data

arXiv:2602.21607 · gr-qc, astro-ph.HE, hep-th · Submitted 2026-02-25 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: General Relativity and Quantum Cosmology. The day in general relativity and quantum cosmology.

Kai: We're looking at the paper: "Direction-of-arrival estimation of a gravitational wave by correlations between quadrupole moments of pulsar timing data".

Mira: Here's the gist of the paper. Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations?

Kai: The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timings due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background.

Mira: Correlations between the quadrupole moments are discussed, where the correlations between pulsar pairs over the full sky are taken into account.

Kai: The correlations turn out to be in the form of a three-dimensional traceless matrix with rank 2 that can be closely related with a projection tensor for the GW.

Title and authors: Kai: So we're diving into a paper called "Direction-of-arrival estimation of a gravitational wave by correlations between quadrupole moments of pulsar timing data". It looks like it tackles how we can figure out where a gravitational wave is coming from just by looking at the timings of pulsars.

Mira: Exactly. The authors are looking at the problem in reverse, trying to estimate the direction of arrival for a gravitational wave signal, specifically considering scenarios like a binary system of supermassive black holes over some background noise.

Lev: So what they're doing is taking those timing measurements and looking at how they correlate across different pulsar pairs. It's an inverse problem; we know the effect on the data, but we want to find the source direction from it.

Kai: Right, and it focuses on these quadrupole moments of pulsar timings. They argue that by looking at correlations between these moments—and accounting for all those different pulsar pairs across the sky—we can get a handle on where that gravitational wave is pointing.

Mira: The paper goes into the specifics of how these correlations turn out, showing they form this three-dimensional traceless matrix with rank two, which they then connect to what's essentially a projection tensor for the gravitational wave itself.

Lev: That connection to a projection tensor is key because it means that this rank two matrix isn't just some abstract correlation; it directly relates to the polarization of the gravitational wave, which is what we need for direction estimation.

Kai: So, can this actually work? The main point they make is that this rank two matrix lets them estimate the direction of arrival in principle. It’s not a magic trick, but a mathematical path to getting there.

Mira: Yes, and they show the derivation of an equation called Q sub ij in Eq (two point one one), which simplifies beautifully as we take the limit where the observation time goes to infinity and the time lag between pulses becomes very small relative to how long a gravitational wave cycle is <ref:2602.21607#pg1>.

Title and authors: Lev: That limit makes things much cleaner, leading them to this simplified form for Q sub ij, which they call KP ij, and they define K as one-seventy-two times the square of the GW plus some other terms related to the polarization.

Kai: So they take that KP ij and use it in Eq (two point two one) to estimate the source direction of a gravitational wave by relating it back to R xy, which is derived from those observed signals sa and sb over time <ref:2602.21607#pg1,of a gravitational wave by>.

Mira: They rearrange that relationship into an equation for the direction vector Omega sub y GW, which is just the negative of R xy divided by K times Omega sub x GW, assuming the direction in x has a non-zero component.

Lev: It’s a direct way to calculate it, but we have to consider how noisy these inputs are. They discuss angular resolution and estimation errors separately in Section four point three and four point two six of the paper <ref:2602.21607#pg2>.

Kai: The angular resolution part shows that the uncertainty in pulsar direction isn't just about time; it's also limited by how many pulsars we actually have, which they show scales with the number of pulsars squared times some factor related to Np, which is the number of pulsars.

Mira: And then there’s the noise contamination aspect, where they look at how much random noise affects this directional estimate. They find that the error in a component of the gravitational wave direction vector scales roughly with one over N p root N obs times the square of h plus, which is related to the GW intensity.

Lev: So for someone building real hardware, that means if we want good angular resolution, we need a lot of pulsars and clean data; otherwise, the error in our estimated direction can be quite large.

Kai: It sounds like they’re saying that while the math works to connect the timing correlations to a directional estimate, the practical performance is heavily dependent on having enough pulsars and managing noise effectively.

Title and authors: Mira: And they do bring up how this method can be used for any direction of arrival, which is a big statement because it suggests this framework isn't tied to one specific sky location.

Lev: But the caveat they give is that the whole derivation relies on certain assumptions about the source being dominant and isolated over a stochastic background, which is what they are trying to estimate in Eq (two point one two) > <ref:2602.21607#pg1>

Kai: So, if we think about what this means for us right now, it means if future pulsar timing arrays get really big, like the square kilometer array we're looking at for the near future, we can actually start using these correlations not just to search for signals, but to try and pinpoint where they come from.

Mira: They show that this method allows for an estimation of the gravitational wave direction even when you are dealing with a single dominant source against a background of many others. That’s important because we often look for those dominant signals.

Lev: The overhead here is that the analysis has to handle these complex correlations between all those quadrupole moments, which means it's not a simple measurement; it's a lot of data processing.

Kai: So, to wrap up on this paper, "Direction-of-arrival estimation of a gravitational wave by correlations between quadrupole moments of pulsar timing data," they show that the rank two matrix approach based on the quadrupole moment correlations allows them to estimate the DOA in principle.

Mira: It’s a way to use the structure inherent in those timing patterns to map out where a gravitational wave is coming from, and they laid out how that estimation depends on things like pulsar count and noise levels >

Lev: And for real hardware running this, you need to be careful about the noise floor because that error term scales with one over N p root N obs times the square of h plus >

Kai: It’s a solid theoretical framework showing how pulsar timing data can move from just finding evidence of a signal to actually estimating its source direction. We'll see what this means when the SKA comes online.

The paper's summary: Kai: So, to recap, these authors are looking at how we can use the correlations between different timing measurements from pulsars to figure out where a gravitational wave is actually coming from in space.

Mira: Right, it’s about taking all those timing data points and turning them into a specific mathematical structure—a three dimensional traceless matrix with rank two—that they link to what's basically the projection of the gravitational wave itself.

Lev: So, the core idea is that this specific way these timing moments correlate across different pulsar pairs gives us enough information to calculate where the source direction is pointing, in principle.

Kai: It’s interesting because it shows a way to use existing data patterns—the correlations—to solve an inverse problem, which is always tricky when you're trying to find something that's happening far away.

Mira: Yeah, and they show how this rank two matrix actually lets you estimate the direction of arrival for the gravitational wave signal. It’s not just a correlation; it’s directly tied to what we know about the gravitational wave's polarization.

Lev: From an engineering side, that means if you have enough pulsars and good data, you can use this structure to get a direction estimate without needing to know everything about the source beforehand.

Kai: The paper then digs into how much uncertainty is involved—the angular resolution and the noise contamination. They show how the number of pulsars limits your ability to resolve the direction, which scales with something involving N sub p squared times some factor related to N sub obs factorial two point six.

Mira: And they also give you a way to estimate the error from random noise, and that error term for a component of your direction vector scales roughly with one over N sub p root N sub obs times the square of the gravitational wave plus intensity.

Lev: So, if you want a decent angular resolution on real hardware, you need more pulsars and less noise than those scaling factors suggest. It tells you exactly what the practical limitations are for this method.

Kai: The implication here is that as we get bigger arrays like the Square Kilometer Array online, these correlations aren't just useful for searching; they become a tool to actually map out where we find those signals.

Mira: It gives us a framework that works whether you're dealing with one dominant source or a background of many sources, because this method is designed to extract the signal direction even in those complex scenarios.

Lev: But the authors are careful; they establish that this whole estimation relies on assuming the source is really dominant and isolated against some kind of stochastic background noise. That’s a big assumption we have to keep in mind.

Kai: So, while it’s a strong theoretical path to DOA estimation, the real challenge now is making sure the assumptions hold up when we actually run experiments with real pulsars and all that noise.

The paper's improvements: Tom: So, we're looking at how these authors suggest ways to make their DOA estimation method even better than what's on this page right now.

Kai: They point out that the current setup is limited by the number of pulsars you have, which means the angular resolution has a ceiling based on N sub p squared times some factor related to N sub obs factorial two point six.

Mira: And they show how you can improve this by focusing on those correlations—specifically, using a three dimensional traceless matrix with rank two that is related to the projection tensor for the gravitational wave.

Lev: That rank two matrix approach, as we talked about before, is what allows them to estimate the direction of arrival in principle because it has that direct relationship with the polarization.

Kai: So, this suggests a way to move beyond just looking at simple correlations and use this more structured mathematical object to get a better answer.

Mira: Exactly, they also discuss how you can improve the estimation error from random noise. Instead of just accepting that error term scaling with one over N sub p root N sub obs times the square of the gravitational wave plus intensity, they show a way to tighten that bound.

Lev: That tightening means for real hardware, you can predict exactly how much noise you need to suppress or how many more pulsars you need to get that error term down further.

Kai: It sounds like they are giving us a roadmap for improving the practical performance, moving from just a theoretical possibility to a more robust estimation technique.

Mira: The implication is that if we can implement this rank two matrix approach, we might be able to achieve better angular resolution than what's currently achievable with simpler methods.

Lev: But there’s still the catch; they’re still relying on those initial assumptions about the source being dominant and isolated against a background of others for this whole estimation framework to work.

Kai: So, it’s a step forward in how we use timing data, but we still have to verify those underlying physical conditions are met before we can trust the results.

Conclusion: Kai: So, to wrap up this talk on "Direction-of-arrival estimation of a gravitational wave by correlations between quadrupole moments of pulsar timing data," these authors show that we can use the rank two matrix structure from the quadrupole moment correlations to estimate where a gravitational wave is coming from.

Mira: Right, so the main point is that this mathematical structure lets us map out the direction of arrival in principle, even when dealing with complex sources and backgrounds.

Lev: From my side, it’s important to remember that this method really needs those initial assumptions about the source being dominant and isolated to actually work on real hardware.

Kai: And they gave us some clear numbers on what limits the resolution and how much noise contamination we can expect in our direction estimate, which is helpful for planning.

Mira: It’s a solid theoretical framework showing that pulsar timing data can move from just finding evidence of a signal to actually estimating its source direction, depending on things like the number of pulsars and how much noise is present.

Lev: So yeah, the practical hurdle is definitely making sure those scaling factors don't blow up when we move from theory to reality with a real array.

Kai: It’s a good demonstration of how structure in the data itself can lead to new ways of looking at gravitational wave signals, which is exactly what we want to see more of.

Mira: Yeah, this gives us a concrete tool based on timing correlations that could be used to guide future searches for gravitational waves from things like supermassive black hole binaries.

Lev: For the error term scaling with one over N sub p root N sub obs times the square of the gravitational wave plus intensity, we need to keep pushing for cleaner data collection.

Kai: So, even with these limitations, it’s a useful piece of work that shows how we can extract directionality from pulsar timings.

Mira: This sets the stage for us to look at how other theoretical models might fit into this framework next time we discuss gravitational wave sources.

Taichi Ueyamaa Hodaka Tamuraa Hideki Asadaa

Graduate School of Science and Technology, Hirosaki University

gr-qc, astro-ph.HE, hep-th

Submitted: 2026-02-25

Updated: 2026-11-04

Comments: 19 pages, 2 figures, distinction from pulsar averaging mentioned, Voronoi tessellation used, realistic pulsar distributions considered, accepted for JCAP

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 81/100

The gist: Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations? The present paper addresses the inverse problem…

Key concepts

Quadrupole Moments
These are statistical measures derived from the timing residuals of pulsars that reflect the quadrupole nature of gravitational waves. They capture how a GW affects the timing measurements across different pulsar pairs, providing information about the source's orientation.
Hellings and Downs (HD) Correlation
This is a known correlation pattern between pulsar timings among pairs. It depends on the angle between the lines of sight to Earth and the pulsars. This correlation serves as a strong evidence for gravitational waves because it reflects the quadrupole nature of GWs.
Rank-2 Matrix and Projection Tensor
The correlations between pulsar quadrupole moments are found to be described by a specific three-dimensional traceless matrix with rank 2. This structure is closely related to the projection tensor of the gravitational wave, which is crucial because this rank-2 matrix enables the estimation of the GW's direction.
Angular Resolution Limit
The precision in estimating the GW direction is limited by two factors: pulsar noise (proportional to $1/N_p$) and observation noise. This results in an angular resolution bound, $\delta p \Omega_{\text{GW}}$, which scales with $\sqrt{3} p N_p! 4.13$. This shows that more pulsars improve the ability to pinpoint the source direction.

Terminology

Summary

Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations? p. 0

Where the paper says it

jp Abstract. Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations?

The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timings due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background. p. 0
Where the paper says it

Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations? The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timings due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background.

Correlations between the quadrupole moments are discussed, where the correlations between pulsar pairs over the full sky are taken into account. p. 0
Where the paper says it

The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timings due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background. Correlations between the quadrupole moments are discussed, where the correlations between pulsar pairs over the full sky are taken into account.

The correlations turn out to be in the form of a three-dimensional traceless matrix with rank 2 that can be closely related with a projection tensor for the GW. p. 0
Where the paper says it

Correlations between the quadrupole moments are discussed, where the correlations between pulsar pairs over the full sky are taken into account. The correlations turn out to be in the form of a three-dimensional traceless matrix with rank 2 that can be closely related with a projection tensor for the GW.

Thereby, we demonstrate that the rank-2 matrix allows to estimate the DOA of the GW. p. 0
Where the paper says it

The correlations turn out to be in the form of a three-dimensional traceless matrix with rank 2 that can be closely related with a projection tensor for the GW. Thereby, we demonstrate that the rank-2 matrix allows to estimate the DOA of the GW.

Introduction to PTA and GW Detection

The idea of a search for gravitational waves (GWs) by using radio pulse timings can be dated back to Reference [1–3]. Hellings and Downs (HD) found that the sky correlation pattern of pulse timings among pulsar pairs depends upon the angle γ between the lines of sight to the pulsars viewed from Earth, and the correlation pattern can be a strong evidence of GWs [4], because it reflects the quadrupole nature of GWs [5–8]. p. 1

Where the paper says it

12) with CA 14 B Relationship between quadratic polarization tensors 16 1 Introduction The idea of a search for gravitational waves (GWs) by using radio pulse timings can be dated back to Reference [1–3]. In their pioneering work, Hellings and Downs (HD) found that the sky correlation pattern of pulse timings among pulsar pairs depends upon the angle γ between the lines of sight to the pulsars viewed from Earth, and the correlation pattern can be a strong evidence of GWs [4], because it reflects the quadrupole nature of GWs [5–8].

Eventually, several teams of pulsar timing arrays (PTAs) have reported a strong evidence of nano-hertz GWs [11–14]. p. 1
Where the paper says it

See also Reference [9, 10] for a review on detection methods of stochastic GW backgrounds. Eventually, several teams of pulsar timing arrays (PTAs) have reported a strong evidence of nano-hertz GWs [11–14].

A possible variance around the HD correlation as the mean has been discussed [28, 29]. p. 1
Where the paper says it

According to these papers, a superposition of supermassive black hole binaries (SMBHBs) are among possible GW sources for the nano-hertz GWs, though improved methods, other possible explanations, and new physics searches have been widely argued e. A possible variance around the HD correlation as the mean has been discussed [28, 29].

If a single SMBHB dominates in a certain frequency bin for PTA observations, can we estimate the direction of arrival (DOA) for the GW? p. 1
Where the paper says it

A possible variance around the HD correlation as the mean has been discussed [28, 29]. If a single SMBHB dominates in a certain frequency bin for PTA observations, can we estimate the direction of arrival (DOA) for the GW?

Currently, this issue is theoretical. p. 1
Where the paper says it

If a single SMBHB dominates in a certain frequency bin for PTA observations, can we estimate the direction of arrival (DOA) for the GW? Currently, this issue is theoretical.

Yet, it is expected that the near future Square Kilometer Array (SKA) will drastically increase the number of pulsars that can be used to significantly improve the sensitivity of PTAs. p. 1
Where the paper says it

[30, 31] for methods for a single GW source detection by using PTAs. Yet, it is expected that the near future Square Kilometer Array (SKA) will drastically increase the number of pulsars that can be used to significantly improve the sensitivity of PTAs.

Derivation from Quadrupole Moments

The directional average of F A a q ij a for A = +, × becomes CAe ij A (omegaˆ GW), where the sum over A is not taken and CA = 1/6. By using Eqs. p. 3

Where the paper says it

10) mean that there are no dipole and quadrupole moments of noises. By using Eqs.

(2.3), (2.4) and (2.12), Eq.(2.11) is calculated as Q ij = 1/36Tobs Z Tobs 0 dt (h+)

cos(ωt) cos(ω(t + τ))e ik+ (omegaˆ GW)e kj+ (omegaˆ GW) + (h×) squared sin(ωt) sin(ω(t + τ))e ik× (omegaˆ GW)e kj× (omegaˆ GW) + h++h× h cos(ωt) sin(t + τ))e ik+ (omegaˆ GW)e kj× (omegaˆ GW)+ sin(ωt) cos(t + τ))e ik× (omegaˆ GW)e kj+ i. Substituting Eq. p. 4

Where the paper says it

15) See Appendix B for the derivation of Eq. Substituting Eq.

(14) into Eq.(13) leads to Q ij = 1/36Tobs Z Tobs 0 dt (h+)

cos(ωt) cos(ω(t + τ))P ij + (h×) squared sin(ωt) sin(t + τ))P ij − h++h× h cos(ωt) sin(t + τ))− sin(ωt) cos(t + τ))iϵ ijkomegaˆ k GW. Furthermore, we take the limit as Tobs → ∞, which leads to Q ij → 1/72 h(h+)

squared + (h×) squared i cos(ωτ)P ij − h++h× sin(ωτ)ϵ ijkomegaˆ k GW. From Eq.(16), we thus arrive at Q ij → KPij, where K ≡ 1/72 (h+) squared, and h+tot ≡ (h+) squared + (h×) 2 is the GW intensity. In addition, we suppose that the time lag is much shorter than the period of GWs (denoted as TGW), namely 2π(τ /TGW) = ωτ ≪ 1. p. 4

Where the paper says it

16), we thus arrive at Q ij → 1 72 h(h +) 2 + (h ×) 2 i cos(ωτ )P ij − h +h × sin(ωτ )ϵ ijkΩˆ k GW . 17) In addition, we suppose that the time lag is much shorter than the period of GWs (denoted as TGW ), namely 2π(τ /TGW ) = ωτ ≪ 1.

As an example, we can imagine τ = 2 weeks, and TGW = 12 months. p. 4
Where the paper says it

17) In addition, we suppose that the time lag is much shorter than the period of GWs (denoted as TGW ), namely 2π(τ /TGW ) = ωτ ≪ 1. As an example, we can imagine τ = 2 weeks, and TGW = 12 months.

So, we take the limit as ωτ → 0, which leads to cos(ωτ) → 1 and sin(ωτ) → 0. p. 4
Where the paper says it

13) leads to Q ij = 1 36Tobs Z Tobs 0 dt (h +) 2 cos(ωt) cos(ω(t + τ ))P ij + (h ×) 2 sin(ωt) sin(ω(t + τ ))P ij − h +h × h cos(ωt) sin(ω(t + τ )) − sin(ωt) cos(ω(t + τ ))i ϵ ijkΩˆ k GW . 16) Furthermore, we take the limit as Tobs → ∞, which leads to cos(ωt) cos(ω(t+τ )) → cos(ωτ )/2, sin(ωt) sin(ω(t + τ )) → cos(ωτ )/2, cos(ωt) sin(ω(t + τ )) → sin(ωτ )/2, and sin(ωt) cos(ω(t + τ )) → sin(ωτ )/2.

Eq.(17) thus becomes Q ij → KPij, where K ≡ 1/72 (h+) squared. p. 4
Where the paper says it

So, we take the limit as ωτ → 0, which leads to cos(ωτ ) → 1 and sin(ωτ ) → 0. 17) thus becomes Q ij → KPij , (2.

DOA Estimation from Matrix Components

From Eqs.(3.2) - (3.4), we can estimate the GW source direction omegaˆ GW. p. 0

Where the paper says it

Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations? The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timings due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background.

For i = x and j = y, Eq.(2.21) reads Rxy = −Komegaˆ x GW, which is rearranged as omegaˆ y GW = −Rxy / Komegaˆ x GW, where omegaˆ x GW ≠ 0 is assumed. p. 6
Where the paper says it

21) reads Rxy = −KΩˆ x GW Ωˆ y GW , which is rearranged as Ωˆ y GW = − Rxy KΩˆ x GW , (3. 3) where Ωˆ x GW ̸= 0 is assumed.

Similarly, Eq. p. 6
Where the paper says it

3) where Ωˆ x GW ̸= 0 is assumed. Similarly, Eq.

(3.3) gives omegaˆ y GW = −Rxy / Komegaˆ x GW, where Rxy is related to the observed signals sa(t) and sb(t + τ). p. 6
Where the paper says it

For i = x and j = y, Eq. 21) reads Rxy = −KΩˆ x GW Ωˆ y GW , which is rearranged as Ωˆ y GW = − Rxy KΩˆ x GW , (3.

Therefore, the present method can be used in principle for any DOA. p. 6
Where the paper says it

9) Ωˆ y GW = − Ryz KΩˆz GW . 10) Therefore, the present method can be used in principle for any DOA.

Angular Resolution and Noise Contamination

The fluctuation in the pulsar direction is not dependent on time during the PTA observation. Eq.(4.3) becomes δpR ij ∼ 1/(Np) 2 Xmax K=1 X Np a=1 X Np b=1 δp zazb q ik a q kj b − 1/3 h(cos γab) squared − 1/3 i δ ij. p. 7

Where the paper says it

Namely, (1/Kmax) P K ∼ O(1). 3) becomes δpR ij ∼ 1 (Np) 2 X Np a=1 X Np b=1 δp zazb q ik a q kj b − 1 3 h (cos γab) 2 − 1 3 i δ ij , (4.

By using Eq. p. 1
Where the paper says it

4 Single dominant source over an isotropic stochastic background 13 5 Summary 13 A Derivation of Eq. 12) with CA 14 B Relationship between quadratic polarization tensors 16 1 Introduction The idea of a search for gravitational waves (GWs) by using radio pulse timings can be dated back to Reference [1–3].

(4.8) and (4.9) in Eq.(4.7), we obtain δpR ij ∼ (h+) 4 Np, where q ij a = O(1). p. 8
Where the paper says it

By using Eqs. 7), we obtain δpR ij2 ∼ (h tot) 4 Np , (4.

Therefore, the size of δpRij is obtained as δpR ij ∼ (h+) squared p Np! p. 8
Where the paper says it

10) where |q ij a | = O(1) is used. Therefore, the size of δpRij is obtained as δpR ij ∼ (h tot) 2 p Np .

24 Nobs! 1/2. This leads to the angular resolution limited by the number of pulsars, which is obtained as δpomegaˆ GW = sqrt(3 × r δpomegaˆ i GW) squared ∼ sqrt(3) p Np! p. 9
Where the paper says it

10) are used in the first line. This leads to the angular resolution limited by the number of pulsars, which is obtained as δpΩˆ GW = r δpΩˆ x GW 2 + δpΩˆ y GW 2 + δpΩˆz GW 2 ∼ √ 3 × r δpΩˆi GW 2 ∼ √ 3 p Np .

4.13.

Estimation Error from Random Noise

The ratio of δnRij to Rij gives us a typical size of the estimation error for a component of the GW directional vector. This leads to the typical size of an estimation error in the GW direction as δnomegaˆ GW = sqrt(3 × r δnomegaˆ i GW) squared ∼ sqrt(3) p Np! p. 9

Where the paper says it

2 radian (corresponding to ∼ 10 degrees) for Np ∼ 100 for instance. 3 Estimation error from a random noise In this subsection, next, we shall discuss a typical size of the estimation error due to random noises, where we suppose na(tK) and nb(tK′) are white random noises with the zero mean and the common standard deviation σ for its simplicity.

4.26. Roughly speaking, the estimation error is ∼ 0. p. 9
Where the paper says it

13) is the angular resolution limited by the number of pulsars in the present method. Roughly speaking, the angular resolution is ∼ 0.

1 radian (corresponding to ∼ 6 degrees) for e. p. 9
Where the paper says it

Roughly speaking, the angular resolution is ∼ 0. 2 radian (corresponding to ∼ 10 degrees) for Np ∼ 100 for instance.

g. Np ∼ 100, Nobs ∼ 24, σ/h+ ∼ σ/h× ∼ 10. p. 9
Where the paper says it

Roughly speaking, the angular resolution is ∼ 0. 2 radian (corresponding to ∼ 10 degrees) for Np ∼ 100 for instance.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI can estimate gravitational wave direction of arrival (DOA) by using correlations between quadrupole moments of pulsar timings, specifically by utilizing a three-dimensional traceless matrix with rank 2 that is related to a projection tensor for the gravitational wave.

  2. The improved system can perform angular resolution estimation limited by the number of pulsars, achieving an error bound of delta p omega GW squared approximately equal to sqrt(3) times r delta p omega i GW squared.

  3. The system can estimate the estimation error from random noise, yielding a typical size for the error in a component of the gravitational wave directional vector as 1 over Np root Nobs times sigma h tot squared.

Abstract

Can we estimate the direction of arrival (DOA) of a gravitational wave (GW) signal from pulsar timing array observations? The present paper addresses the inverse problem, for which we consider quadrupole moments of pulsar timing data due to GWs from a dominant isolated source such as a binary of supermassive black holes over an isotropic stochastic background. Correlations between the quadrupole moments are discussed as a possible alternative to the matched filtering technique, where the correlations between pulsar pairs over the full sky are taken into account. The correlations turn out to be in the form of a three-dimensional traceless matrix with rank 3 that can be closely related with a projection tensor for the GW. Thereby, we demonstrate that the rank-3 matrix allows to estimate the DOA of the GW without assuming a template bank nor a priori Gaussian noise. A mean over the celestial sphere is adopted instead of the pulsar averaging. A weighted average is used for a non-uniform distribution of pulsars on the sky, for which the Voronoi tessellation is employed to numerically calculate an area-weighted mean. In expectation of the forthcoming Square Kilometre Array, DOA estimation errors are also examined.

Related papers