Noise Detection Research Articles

A Survey of Deep Learning Methods for Noise Classification and Detection in Historical Image Analysis

Historical images often suffer from various types of noise due to aging, degradation, and other environmental factors, which can significantly impact their quality and usability. The advent of deep learning has revolutionized the field of image processing, offering robust methods for noise classification and detection. This survey provides a comprehensive overview of the current state-of-the-art deep learning techniques employed in historical image analysis for identifying and mitigating different types of noise. We review various deep learning architectures, including convolutional neural networks (CNNs), generative adversarial networks (GANs), and auto encoders, that have been applied to this problem. The paper discusses the strengths and limitations of each approach, highlights key challenges such as data scarcity and variability in noise types, and explores future directions in this field. The survey aims to serve as a resource for researchers and practitioners by summarizing the most effective methods, evaluating their performance on different datasets, and outlining potential avenues for further research

Magnetic noise detection and imaging from a superparamagnetic core-shell particle via an ensemble of nitrogen-vacancy centers in a thin diamond chip

Magnetic noise detection and imaging from a superparamagnetic core-shell particle via an ensemble of nitrogen-vacancy centers in a thin diamond chip

Performance evaluation of perceptible impulsive noise detection methods based on auditory models

Reference-free audio quality assessment is a valuable tool in many areas, such as audio recordings, vinyl production, and communication systems. Therefore, evaluating the reliability and performance of such tools is crucial. This paper builds on previous research by analyzing the performance of four additional algorithms in detecting perceptible impulsive noise (clicks) based on auditory models. We compared the results of eight algorithms, hypothesizing that computationally simpler algorithms could perform as well as more complex ones. We obtained a set of audio signals, with and without clicks, annotated by human subjects from a publicly available dataset. Audio signal sets are categorized based on the obtained annotation results to train the algorithms for different levels of the experiments. Experiments containing cross-validation are done for multiple parameters of algorithms. The algorithm training is based on maximizing a discriminability metric (A′). Evaluation criteria of the algorithms included the hit rate, false alarm rate, A′, and computational time. Our findings indicate that computationally simpler auditory models have performed as well as computationally more complex ones, while conventional models exhibit lower performance. Conclusively, the ERBlet transform based algorithm demonstrated superior performance in terms of A′ and robustness. This paper provides insights into the capabilities of auditory models in a practical use case of perceptible click detection. The results presented here can help research and develop such algorithms for vinyl production, audio archiving, podcasting, music production, and telecommunications.

Impact of slip velocity-dependent friction coefficient on surface traction, wear, RCF and curve squeal noise prediction in wheel-rail contact

Wheel-rail contact friction coefficient is often assumed to be constant through the entire contact patch for the calculation of surface traction. In reality, however, the friction value in a certain point decreases when transitioning from adhesion to slip regimes. Including this friction coefficient behaviour in the estimations of surface traction on the contact patch can potentially provide more accurate calculations of wear and rolling contact fatigue (RCF). In the present work, a slip velocity-dependent friction coefficient is implemented in the tangential contact solver using the concept of ‘Friction Memory’. The effect of this implementation on traction estimations and on the prediction of wear and RCF is analysed by comparing the results with a case with constant friction coefficient in the contact patch. Furthermore, the slip velocity-dependent friction coefficient provides a creep curve with a maximum creep forces value, and a decreasing creep force for higher creepages. This is commonly known as one of the possible mechanisms of curve squeal noise generation. The results provide insights into the likelihood of curve squeal generation, and an on-set curve squeal noise detection technique is proposed that also accounts for the influence of profile changes due to wear.

Real-Time Detection, Evaluation, and Mapping of Crowd Panic Emergencies Based on Geo-Biometrical Data and Machine Learning

Crowd panic emergencies can pose serious risks to public safety, and effective detection and mapping of such events are crucial for rapid response and mitigation. In this paper, we propose a real-time system for detecting and mapping crowd panic emergencies based on machine learning and georeferenced biometric data from wearable devices and smartphones. The system uses a Gaussian SVM machine learning classifier to predict whether a person is stressed or not and then performs real-time spatial analysis to monitor the movement of stressed individuals. To further enhance emergency detection and response, we introduce the concept of CLOT (Classifier Confidence Level Over Time) as a parameter that influences the system’s noise filtering and detection speed. Concurrently, we introduce a newly developed metric called DEI (Domino Effect Index). The DEI is designed to assess the severity of panic-induced crowd behavior by considering factors such as the rate of panic transmission, density of panicked people, and alignment with the road network. This metric offers immeasurable benefits by assessing the magnitude of the cascading impact, enabling emergency responders to quickly determine the severity of the event and take necessary actions to prevent its escalation. Based on individuals’ trajectories and adjacency, the system produces dynamic areas that represent the development of the phenomenon’s spatial extent in real time. The results show that the proposed system is effective in detecting and mapping crowd panic emergencies in real time. The system generates three types of dynamic areas: a dynamic Crowd Panic Area based on the initial stressed locations of the persons, a dynamic Crowd Panic Area based on the current stressed locations of the persons, and the dynamic geometric difference between these two. These areas provide emergency responders with a real-time understanding of the extent and development of the crowd panic emergency, allowing for a more targeted and effective response. By incorporating the CLOT and the DEI, emergency responders can better understand crowd behavior and develop more effective response strategies to mitigate the risks associated with panic-induced crowd movements. In conclusion, our proposed system, enhanced by the incorporation of these two new metrics, proves to be a dependable and efficient tool for detecting, mapping, and assessing the severity of crowd panic emergencies, leading to a more efficient response and ultimately safeguarding public safety.

Molecular Ferroelectrics for Highly Sensitive Detection Toward Low-Frequency Sound Recognition.

Human hearing cannot sensitively detect sounds below 100Hz, which can affect the physical well-being and lead to dizziness, headaches, and nausea. Piezoelectric acoustic sensors still lack sensitivity to low-frequency sounds owing to the low piezoelectric coefficient or high elastic modulus of materials. The low elastic modulus and substantial piezoelectric coefficient of molecular ferroelectric materials make them excellent candidates for acoustic sensors. In this study, the molecular ferroelectric, [(CH3)3NCH2Cl]CdCl3, is used as a piezoelectric active layer in the construction of a piezoelectric acoustic sensor for low-frequency sound detection. The sensor exhibits high sensitivity (47.43mV Pa-1 cm-2) at 87Hz, with an excellent level of frequency resolution (up to 0.1Hz). This facilitates the accurate discrimination and detection of low-frequency sounds, which is suitable for noise detection applications. The sensor differentiates between various musical instruments and heartbeats, and recognizes audio signals. This study highlights the potential of molecular ferroelectric materials in piezoelectric acoustic device applications, including noise detection, health monitoring, and human-computer interactions.

High precision and low noise demodulation scheme using adaptive fast fourier transform for fiber-optic interferometric sensors.

This paper presents an adaptive fast Fourier transform (adaptive FFT) demodulation scheme, aimed at enhancing the precision and noise suppression capability of signal processing in fiber-optic interferometric sensors. By adaptively optimizing the length of the acquired spectrum and dynamically adjusting the frequency domain resolution, the proposed scheme can precisely calculate the eigenfrequency of the reflected spectrum. Therefore, the adaptive FFT demodulation scheme can effectively enhance the extraction ability of phase quadrature demodulation signal. In comparison to traditional FFT demodulation scheme, the proposed scheme demonstrates nearly a tenfold improvement in eigenfrequency calculation accuracy. The experimental results indicate a significant improvement in the stability of the adaptive FFT demodulation scheme. The fluctuations are reduced to approximately 0.03 dB. This value is roughly equivalent to 25% of the fluctuations typically observed in traditional FFT demodulation schemes. Additionally, the adaptive FFT scheme demonstrates superior adaptability to sensors with varying optical path differences (OPDs). It maintains noise levels below -100 dB@1 kHz across sensors with different OPDs, and substantially reduces noise fluctuations compared to traditional FFT scheme. Furthermore, the proposed scheme exhibits notable advantages in crosstalk suppression for multiplexed signal demodulation. The fundamental frequency crosstalk (FFC) is reduced to below -50 dB, which is approximately 10 dB lower than that achieved by traditional FFT methods. The adaptive FFT demodulation scheme shows promising potential for applications in low noise, high precision, and dynamic signal detection within fiber-optic interferometer sensors and multiplexing arrays.

Low-noise laser intensity noise evaluation system at Hz frequency band for ground-based gravitational wave detection

The direct detection of gravitational waves has opened a new window for understanding the universe and trailblazed multi-messenger astronomy. The frequency band of gravitational waves generated by various astronomical events can cover broadband range, and it is various for detection mechanism and scheme of gravitational waves in different frequency band. For example, The ground-based gravitational wave detection has the frequency band for 10 Hz to 10 kHz to mainly detect which based on Michelson interferometer; The space gravitational wave detection has the frequency band of 0.1 mHz to 1 Hz to mainly detect which based on space interferometer; The pulsar gravitational wave detection has the frequency band for 1×10<sup>-9</sup> Hz to 1×10<sup>-7</sup>Hz to mainly detect which based on pulsar timing array. The next generation ground-based gravitational wave project demands higher sensitivity to detect faint signals, necessitating an assessment system with minimal background noise to accurately measure the laser relative intensity noise. At present, the detection frequency band of ground-based gravitational wave detection devices in operation is mainly concentrated in the range of 10 Hz - 10 kHz. To satisfying the detection sensitivity requirements, the laser relative intensity noise should be accurately evaluated and suppressed to ≤2.0×10<sup>-9</sup>Hz<sup>-1/2</sup>@10 Hz and ≤4.0×10<sup>-7</sup>Hz<sup>-1/2</sup>@10 kHz by photoelectric feedback. This paper constructed an evaluation and characterization system for ground-based gravitational wave band laser intensity noise which based on low noise and high sensitivity photoelectric detection device and combined with LabVIEW and MATLAB algorithm programming for instrument control and data processing. This low noise evaluation system is used to test the background noise of FFT analyzer SR760, preamplifier SR560, photoelectric detector electronics noise and intensity noise of self-developed optical fiber amplifier, and then the data extraction and image processing are carried out by LabVIEW and MATLAB algorithms, and finally the evaluation of ground-based gravitational wave frequency band system is obtained. The experimental results show that the whole electronic noise for the preamplifier SR560 and FFT analyzer SR760 is lower than 3.8×10<sup>-9</sup> Hz<sup>-1/2</sup>@(10 Hz-10 kHz). The electronic noise for the photodetector is lower than 1.4×10<sup>-8</sup>V/√Hz@10 Hz&8.1×10<sup>-9</sup>V/√Hz@10 kHz and the accuracy of the system is calibrated and tested by the standard sinusoidal signal. Finally, the noise of commercial laser is evaluated and compared with the factory data to verify the accuracy of the evaluation system. Related research, device and system development provide hardware, software and theoretical basis for the preparation of high-power low-noise laser light source and gravitational wave detection and provide theoretical basis and evaluation criteria for ground-based gravitational wave detection in China.

Analysis of laser spot centroiding errors for laser acquisition system in gravitational wave detection missions.

In space-based gravitational wave detection, establishing ultra-long-distance and ultra-high-precision laser links between satellites is achieved through the laser acquisition and tracking system. The laser spot centroid positioning method, which offers low computational complexity and strong adaptability to beam shape, is currently the core measurement method during the laser acquisition phase. However, due to various interference factors encountered in practical tests, this algorithm often falls short of meeting the extremely high requirements. To address this challenge, this paper first defines the specific performance criteria for the centroid positioning method based on the needs of laser acquisition in gravitational wave detection. It then comprehensively analyzes how detector noise, window truncation effects, and beam wavefront aberrations impact the accuracy of angular measurements. Using derived analytical expressions, an improved centroiding algorithm is proposed to mitigate the effects of detector noise and wavefront aberrations simultaneously. Numerical simulations are conducted to design the specific parameters for the algorithm and the system, resulting in the ability to achieve an angular measurement accuracy of 60 nrad at the telescope front end.

Статистики вищих порядків в задачах виявлення сигналів на фоні корельованих негаусових завад

Signal detection in noise is critical in telecommunications, navigation and radar systems, image processing and biomedical research, where noise often deviates from a normal distribution, and sample values exhibit statistical dependence. Traditional methods for analyzing and designing such systems face significant limitations, including algorithmic and computational complexity, severely restricting their practical application. An effective approach to developing signal detection systems involves using moment and cumulant descriptions of random variables, simplifying the design of detection systems for signals with various probability density functions. The authors propose a novel approach based on one-dimensional (1D) and two-dimensional (2D) moment-cumulant models to describe correlated non-Gaussian processes. This approach enables the modification of the moment-based criterion for statistical hypothesis testing and synthesising polynomial stochastic decision rules for detecting signals in correlated non-Gaussian noise. The study demonstrates that the nonlinear processing of sample values and the application of higher-order statistics to describe the investigated processes account for the structure of non-Gaussian noise and its statistical dependencies. This reduces the error probabilities of the synthesized decision rules compared to traditional Gaussian models. The study aims to enhance the efficiency of signal detection systems under additive interaction with correlated non-Gaussian noise by developing new moment-cumulant models of the investigated processes, modifying the moment-based criterion for hypothesis testing, and designing polynomial decision rules. The study's practical significance lies in creating simple-to-implement algorithms with high accuracy that achieve lower error probabilities in decision rules compared to existing methods.

A method for detecting cracks on the trailing edge of wind turbine blades based on aeroacoustic noise analysis

This study uses aeroacoustic noise to identify cracking and debonding faults on the trailing edge (TE) of wind turbine blades, which is very meaningful because it can monitor early faults without affecting the normal operation of the wind turbine, thereby improving the safety and reliability of the wind turbines and extending the service life of the blades. First, the semi-empirical model called NAFNOISE is used to analyze the factors that influence the noise from the TE, such as chord length, angle of attack, and TE thickness. Computational aeroacoustics (CAA) as a more precise finite element method is further employed to calculate the areoacoustics noise (AAN) of the blade with TE cracking. Based on the vortex shedding diagram of the calculation results, a preliminary explanation is provided for the main reason high-frequency energy appears in the sound power spectrum when the blades have a TE cracking failure. Next, the semi-empirical model is compared with the noise waveforms obtained by CAA, the differences between different models in identifying TE cracking are analyzed, and the theoretical acoustic characteristics of AAN changes when blades fail are drawn. Finally, a noise detection experiment and corresponding signal processing technology are discussed. The results demonstrate that the method proposed in this article is not only capable of detecting debonding and cracking faults at the TE but also improves the theoretical research on active and noncontact damage monitoring of the blades.

High-Performance Perovskite Flat Panel X-Ray Imagers via Blade Coating.

Perovskite X-ray detectors are recognized as the most promising candidates for low-dose detectors due to their superior performance. However, it is still full of challenging in the fabrication of flat-panel X-ray imagers (FPXIs), primarily due to the absence of large area thick films that exhibit high uniformity and long-term performance stability. A general synthesis route is urgently needed to grow large-scale halide perovskite thick films directly on a pixeled thin-film transistor (TFT) backplane with high uniformity, closing the gap between the great potential of perovskite X-ray detectors and their entry into the market. In this work, an advanced precursor paste suitable for blade coating is developed to enable high-throughput manufacturing of FPXIs. Highly uniform perovskite films with a thickness of 300-micrometers are directly deposited onto pixeled TFT substrates by the blade-coating method using the above paste, which governs a stable dark current and minimal noise of the X-ray detectors. As a result, (BA)2(MA)9Pb10I31 perovskite X-ray detectors achieved a high sensitivity of 15200 µC Gyair -1 cm-2 and a limit of detection (LoD) of 26.8 nGyair s-1. Moreover, FPXIs with a spatial resolution of 0.95 lp mm-1 (0.475 lp pixel-1) are obtained, which exhibits negligible signal crosstalk and excellent X-ray imaging performance.

A physically modelled selection function for compact binary mergers in the LIGO-Virgo O3 run and beyond

Abstract Despite the observation of nearly 100 compact binary coalescence (CBC) events up to the end of the Advanced gravitational-wave (GW) detectors' third observing run (O3), there remain fundamental open questions regarding their astrophysical formation mechanisms and environments. Population analysis should yield insights into these questions, but requires careful control of uncertainties and biases. GW observations have a strong selection bias: this is due first to the dependence of the signal amplitude on the source's (intrinsic and extrinsic) parameters, and second to the complicated nature of detector noise and of current detection methods. In this work, we introduce a new physically-motivated model of the sensitivity of GW searches for CBC events, aimed at enhancing the accuracy and efficiency of population reconstructions. In contrast to current methods which rely on re-weighting simulated signals (injections) via importance sampling, we model the probability of detection of binary black hole (BBH) mergers as a smooth, analytic function of source masses, orbit-aligned spins, and distance, fitted to accurately match injection results. The estimate can thus be used for population models whose signal distribution over parameter space differs significantly from the injection distribution. Our method has already been used in population studies such as reconstructing the BBH merger rate dependence on redshift.

Real-time anomaly detection in electric motor operation noise

Anomaly detection plays a very important role in many fields to identify abnormalities occurring in the system earlier. This study proposes a new abnormality detection solution for 3-phase electric motors based on their working noise. Normal and abnormal operating noise data sets for an electric motor were acquired in the laboratory. These datasets are converted into the corresponding two-dimensional gray spectrogram image sets. The normal set is used to train the autoencoder (AE) model to find the abnormality evaluation threshold. This threshold is validated again with anomalous data sets. The trained AE is then quantized to be installed on a system consisting of two duo-core microcontroller units (MCUs) for real-time testing. Free real-time operating system (FreeRTOS), a real-time operating system, is used to schedule tasks on the system. Experimental results show that the designed anomaly detector can accurately detect over 99% of abnormal events. The system can communicate with a supervisory control and data acquisition (SCADA) application running on the S7-1200 programmable logic controller (PLC) platform using the Modbus transmission control protocol (TCP) protocol. The SCADA application can continuously record evaluated results from the system and adjust abnormal thresholds for the system directly on the human-machine interface (HMI) screen.

An oversampling FCM-KSMOTE algorithm for imbalanced data classification

In recent years, imbalanced data classification has emerged as a challenging task. To address this issue, we propose a novel oversampling method named FCM-KSMOTE. The algorithm initially performs a density-based fuzzy clustering on the data, then iterates to partition regions and perform oversampling inside each cluster. Secondly, it merges the clusters and conducts noise detection to obtain a balanced dataset. Finally, we conducted the experiments on 19 public datasets and 3 synthetic datasets. Six evaluation metrics of Recall, Accuracy, G-mean, Specificity, AUC and F1-Score were used in the experiments. The experimental results demonstrate that our method can significantly improve the recognition rate of the minority class while maintaining high accuracy for the majority class. Particularly with the RF classifier, our method ranks first in all evaluation metrics, with a Recall difference of up to 0.2 compared to the least performing method, demonstrating its substantial performance advantage.

Enhancing Data Security through Digital Image Steganography: An Implementation of the Two Least Significant Bits (2LSB) Method

The rapid advancement of digital communication technologies has heightened the demand for robust data security methods, particularly for safeguarding sensitive information against cyber threats. Steganography, the art of concealing data within digital media, offers an effective means of securing communication by masking the very existence of hidden messages. This study evaluates the Two Least Significant Bits (2LSB) method as an enhancement to the conventional Least Significant Bit (LSB) technique, focusing on improving storage capacity and resistance to detection. The research employs a quantitative methodology, conducting experiments to analyze the impact of data embedding on image quality using metrics such as Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR). Data embedding was tested on images of varying sizes and resolutions, with results compared before and after embedding. The study also evaluates the system’s tolerance to compression and robustness against steganalysis attacks, such as histogram analysis and noise detection. The results demonstrate that the 2LSB method significantly enhances storage capacity while maintaining high image quality. For instance, PSNR values remained above 40 dB across all tested images, indicating minimal visual distortion. The method also showed resilience to lossless compression formats such as PNG, with minimal impact on data integrity. However, lossy compression formats like JPEG posed challenges in maintaining the embedded data. The MSE values before and after embedding were minor, ensuring that the hidden data remained undetectable under standard analysis. This research contributes to developing more efficient and secure steganography techniques, providing practical applications in protecting sensitive data during transmission. The findings underline the 2LSB method's potential as a viable solution for digital security, balancing enhanced storage capacity with robust resistance to detection. Future work should further explore integrating advanced encryption with 2LSB to strengthen data protection in diverse digital environments

VLSI implementation of a modified min-max median filter using an area and power competent tritonic sorter for image denoising

The prominence of image processing in today’s cutting-edge technology is undeniable. Integrating software with hardware leverages both strengths, resulting in a real-time processing system that is efficient and streamlined. Raw images are usually affected by noise, which hinders the acquisition of good-quality and detailed images; hence, denoising becomes necessary. This paper proposes a modified min-max median (MMM) filter to remove impulse noise and a Tritonic sorter to localize corrupted pixels. The proposed denoising method focuses on localizing noisy pixels, unlike traditional denoising approaches, which focus only on noise detection and filtering. A min-max sheet provides the location of the corrupted pixels, and filtering is performed on them. The Tritonic Sorter, consisting of a max locator and a min locator, compares three input values and finds the minimum, maximum and median values among them. Compared to other state-of-the-art methods, the proposed method minimizes the number of comparators needed to carry out the sorting process. The proposed method was synthesized in the ZedBoard Zynq kit using the Vivado tool. The results show that the area improved by 27%, and the power improved by 16.23% compared with those of the existing method.

Noise Detection and Suppression From ECG Through Improved CEEEMD and Adaptive Wavelet Soft Thresholding

Electrocardiogram (ECG) is a noninvasive, effective and economical biomedical signal that is vital in diagnosing cardiovascular diseases. However, the acquiring process contaminates the ECG signal with several types of noises like Motion Artifacts, Power Line Interference and Baseline Wander. Hence, this paper proposes a new approach to detect and suppress the noises from ECG signals. The complete methodology comprises two stages: noise detection and noise suppression. The former stage applies Improved Complete Ensemble Empirical Mode Decomposition (CEEMD) to decompose the noisy ECG into Intrinsic Mode Functions (IMFs). Next, Maximum Absolute Amplitude (MAA) and Auto-Correlation Maximum Amplitude (AMA) are extracted and used to classify the type of noises from ECG. Then, the noisy ECG segments are processed through the second stage and decomposed into sub-bands through Discrete Wavelet Transform (DWT). Then, the sub-bands are categorized into noise-dominant and signal-dominant frequency bins, and only noise-dominant frequency bins are subjected to noise suppression through a newly proposed adaptive soft thresholding mechanism. The effectiveness of the proposed method is assessed by contaminating the ECG signals acquired from the MIT-BIH arrhythmia database with different noises at different Signal-Noise Ratios (SNRs). Three performance metrics, namely Output SNR, Root Mean Square Error (RMSE) and Pearson Correlation Coefficient (PCC), are employed to explore the superiority of the proposed method over state-of-the-art methods, which considered EMD and CEEMD as decomposition filters. The proposed method improved by an average of 3.5 dB in output SNR and 0.0290 in RMSE.

Adaptive gravitational clustering algorithm integrated with noise detection

Clustering analysis is frequently used in data mining, image processing, artificial intelligence, and so on. Traditional approaches heavily rely on manually configured parameters, of which the initial selection exerts a profound influence on the clustering outcomes. In addition, they usually only consider the relationship between two individual samples when calculating distances, neglecting the overall structure of the dataset, which can negatively affect clustering performance. At the same time, many contemporary algorithms are tailored to specific datasets, posing challenges in achieving optimal clustering performance for intricate, noisy datasets. To address these limitations, we propose an Adaptive Gravitational Clustering Algorithm Integrated with Noise Detection called GCIND. Inspired by the law of gravitation, GCIND takes into account the natural neighborhood structure of the entire dataset, adaptively computing the gravitation between data points by leveraging shared neighbors and Euclidean distance relationships. Our algorithm initially identifies and eliminates outliers or edge points in the dataset. It subsequently uses gravitation to autonomously cluster the remaining core data. Finally, the removed data are reallocated to their respective clusters. GCIND has four notable advantages: (1) it uses gravitation to build the neighborhood graph, reflecting the overall dataset structure; (2) it demonstrates stronger robustness in handling noisy datasets; (3) it uses adaptive gravitational neighborhood graph clustering, removing manual parameter tuning; (4) it adapts to complex manifold-structured datasets, offering broad applicability. Experiments have shown that GCIND, without requiring any parameter settings, demonstrates slightly better performance than the algorithms compared in the study, especially when dealing with complex manifold datasets.

Suppression of clock-jitter noise and laser phase noise in arm locking

Abstract Arm-locking technique has been a focus of attention as one of the means to suppress the laser phase noise in space-based gravitational wave detector. The main idea of the arm-locking technique is to transfer the stability of the detector arm length to laser frequency by introducing a feedback control loop. Generally, laser phase noise will be suppressed by an amount similar to the magnitude of the controller gain. However, on the one hand, clock-jitter noise and optical bench motion noise, as the noise floor of the arm-locking technique, need to be suppressed. On the other hand, limited by the Doppler frequency pulling, the gain of the controller generally cannot be too large. It means that even if we do not consider clock-jitter noise and optical bench motion noise, it is difficult to suppress laser phase noise below the noise floor only by arm-locking technique. In this work, we combine self-referenced optical frequency combs and arm-locking technique to generate clock signals that are coherently referenced to the closed-loop laser beams, so that the clock-jitter noise is also suppressed by about the level of controller gain. We conduct a simulation on the above configuration, and the results show that the performance of the arm-locking is no longer limited by clock-jitter noise in the low-frequency band. To address the issue of insufficient laser phase noise suppression by the arm-locking technique, we further investigate time-delay interferometry (TDI) combinations under outputs of arbitrary arm-locking configurations. We obtain the equations for eliminating laser phase noise. To ensure that the TDI combinations can directly operate in the time domain, we derive a restricted solution space by assuming a specific form for the solutions.