Abstract:Radio-frequency interference detection and flagging is one of the most difficult and urgent problems in 21 cm Epoch of Reionization research. In this work, we present $\chi^2$ from redundant calibration as a novel method for RFI detection and flagging, demonstrating it to be complementary to current state-of-the-art flagging algorithms. Beginning with a brief overview of redundant calibration and the meaning of the $\chi^2$ metric, we demonstrate a two-step RFI flagging algorithm which uses the values of this metric to detect faint RFI. We find that roughly 27.4\% of observations have RFI from digital television channel 7 detected by at least one algorithm of the three tested: 18.0\% of observations are flagged by the novel $\chi^2$ algorithm, 16.5\% are flagged by SSINS, and 6.8\% are flagged by AOFlagger (there is significant overlap in these percentages). Of the 27.4\% of observations with detected DTV channel 7 RFI, 37.1\% (10.2\% of the total observations) are detected by $\chi^2$ alone, and not by either SSINS or AOFlagger, demonstrating a significant population of as-yet undetected RFI. We find that $\chi^2$ is able to detect RFI events which remain undetectable to SSINS and AOFlagger, especially in the domain of long-duration, weak RFI from digital television. We also discuss the shortcomings of this approach, and discuss examples of RFI which seems undetectable using $\chi^2$ while being successfully flagged by SSINS and/or AOFlagger.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the difficult problem of radio - frequency interference (RFI) detection and flagging in the study of the 21 - centimeter epoch of reionization (EoR). Specifically, the author proposes using the χ² statistic in redundant calibration as a new RFI detection tool and demonstrates its effectiveness in detecting weak RFI. #### Background and problem description In the 21 - centimeter EoR study, radio - frequency interference (RFI) is one of the main sources of data contamination. Removing RFI is crucial for power - spectrum analysis because if RFI cannot be effectively removed, the measurement results will not accurately reflect the 21 - centimeter EoR signal. Existing RFI detection tools such as AOFlagger and SSINS are powerful but still cannot completely eliminate all RFI, especially the so - called "ultra - weak" RFI. #### Solution proposed in the paper The author proposes a new method based on the χ² statistic of redundant calibration to detect and flag RFI. Redundant calibration takes advantage of the redundancy of antenna baselines, that is, baselines with the same vector separation should measure the same sky visibility. By minimizing the χ² value, the antenna gain can be adjusted so that the actually measured visibility is as close as possible to the calculated "consensus" visibility. The author found that this method can detect RFI events that other existing algorithms failed to identify, especially in the case of long - duration, weak digital television (DTV) RFI. #### Experimental verification By conducting experiments on the data of the Murchison Widefield Array (MWA), the author compared the performance of the newly proposed χ² method with existing tools such as AOFlagger and SSINS. The results show that RFI from digital television channel 7 exists in about 27.4% of the observed data, of which 18.0% is flagged by the χ² algorithm, 16.5% by SSINS, and 6.8% by AOFlagger. It is worth noting that 37.1% of the DTV RFI is only detected by the χ² algorithm and not captured by the other two algorithms. #### Conclusion This study shows that the χ² statistic based on redundant calibration is an effective RFI detection tool, especially in detecting long - duration, weak RFI. Although this method still has some limitations, it provides new ideas and directions for future improvement of RFI detection algorithms. ### Key formulas 1. **Visibility model**: \[ v_{ij}(t, \nu, p) \approx g_i(t, \nu, p) g_j^*(t, \nu, p) y_{ij}(t, \nu, p)+n_{ij} \] where $v_{ij}$ is the measured visibility, $g_i$ and $g_j$ are antenna gains, $y_{ij}$ is the "true" sky visibility, and $n_{ij}$ is the noise term. 2. **χ² definition**: \[ \chi^2=\sum_{i < j}\frac{\vert v_{ij}-g_i g_j^* y_{ij}\vert^2}{\sigma_{ij}^2} \] where $y_{ij}$ is the "consensus" visibility of the baseline group, and $\sigma_{ij}^2$ is the estimated noise variance of this baseline type. 3. **Degree - of - freedom calculation**: \[ n_{\text{DoF}}=N_{\text{bl}}-N_{\text{ubl}}-N_{\text{ants}} + 2.5 \] where $N_{\text{bl}}$ is the total number of baselines, $N_{\text{ubl}}$ is the number of unique baseline types, and $N_{\text{ants}}$ is the total number of antennas. 4. **Modified z - score calculation**: \[ z_i = 0.6745\times\frac{x_i-\tilde{x}}{\text{MAD}} \] where $\tilde{x}$ is the median of the data set, M

$χ^2$ from Redundant Calibration as a Tool in the Detection of Faint Radio-frequency Interference

Calibration of Quasi-Redundant Interferometers

Radio Frequency Interference Detection and Mitigation Using Compressive Statistical Sensing

Spectral Redundancy for Calibrating Interferometers and Suppressing the Foreground Wedge in 21cm Cosmology

Spectral redundancy for calibrating interferometers and suppressing the foreground wedge in 21 cm cosmology

Spectral Kurtosis Based RFI Mitigation for CHIME

Multifrequency Array Calibration in Presence of Radio Frequency Interferences

Interference detection in radio astronomy applying Shapiro-Wilks normality test, spectral entropy, and spectral relative entropy

Reflection of Transmitting Antenna in Reverberation Chamber and Its Effect on Chi-square Validation

The Bias and Uncertainty of Redundant and Sky-based Calibration Under Realistic Sky and Telescope Conditions

A robust RFI identification for radio interferometry based on a convolutional neural network

Exploring the consequences of chromatic data excision in 21-cm Epoch of Reionization power spectrum observations

Multi-detector null-stream-based $χ^2$ statistic for compact binary coalescence searches

Quantifying Suppression of the Cosmological 21-cm Signal due to Direction Dependent Gain Calibration in Radio Interferometers

Quantifying Excess Power from Radio Frequency Interference in Epoch of Reionization Measurements

Trajectory Based RFI Subtraction and Calibration for Radio Interferometry

Mitigating calibration errors from mutual coupling with time-domain filtering of 21 cm cosmological radio observations

Redundancy Calibration of Phased Array Stations

A Digital Calibration Source for 21cm Cosmology Telescopes

Comparison of RFI Mitigation Strategies for Dispersed Pulse Detection

Simulating Spectral Kurtosis Mitigation Against Realistic RFI Signals