Cell-free massive multi-input multi-output (MIMO) has recently gained much attention for its potentialin shaping the landscape of sixth-generation (6G) wireless systems. This paper proposes a hierarchical network architecture tailored for cell-free massive MIMO, seamlessly integrating co-located and distributed antennas. A central base station (CBS), equipped with an antenna array, positions itself near the center of the coverage area, complemented by distributed access points spanning the periphery. The proposed architecture remarkably outperforms conventional cell-free networks, demonstrating superior sum throughput while maintaining a comparable worst-case per-user spectral efficiency. Meanwhile, the implementation cost associated with the fronthaul network is substantially diminished.

This material is dedicated to the rather old, but always actual, topic in theoretical analysis of applied problems: how to simplify an analytical investigation by applying multidimensional Probability Density Functions (n-PDFs). A brief overview of the applied interpretations for n-PDF approximation is presented, which allows to make choice of the practically useful analytical presentations for n-PDFs. For analytical solutions, two approximations for the n-PDF were chosen: Poli-Gaussian approximation (mix of Gaussian PDF's) and Functional Approximation or Gaussian-like. The first one is found to be suitable (and is recommended) for the class of scenarios better described by the term "scenarios of random structure". The second seems to be more adequate for analysis of the numerical characteristics of the communication system's behavior Recommendations for validation of the models with given n-PDF approximations, based on goodness of fit criteria's can be found as well.

In electrical impedance tomography (EIT) systems, various current sources were designed to deliver constant electrical current towards the electrodes, irrespective of the conductivity of the conductive medium under test. This, however, is not possible to achieve in practice since the water conductivity ranges from 5 S/m for sea water down-to 5.5 x10-6 and 5 x 10-3 S/m for pure and drinking water respectively which are substantially low values [1]. Thus, even in case of high water-cut fluid, unless the water is very salty, the usage of such current sources may not be appropriate and their substitution with an electrical capacitance tomography system or a dielectric measurement sensor may be required. Indeed, even if the conductivity of the medium is known and high, the gross conductivity formed between a given pair of electrodes depend heavily on the phase’s distribution pattern between them and can be excessively low, causing high electric current fluctuations. One of the contributions of this paper is to experimentally assess the effect of the electrical current fluctuations on the accuracy of the EIT image reconstruction using three different cutting-edge and most widely current sources designs. To the authors’ best knowledge, this study is the first of its kind as all other prior works assume that the electrical current is constant in the formulation of the EIT forward and inverse problems. The paper also suggests a new cost-effective measurement circuit design that overcomes the fluctuations. It continuously measures the electrical current consumed during every single excitation cycle using a very high-speed analog-to-digital (ADC) converter, interfaced with Cyclone V field program gate array (FPGA), to compensate for the associate voltage readings accordingly during the image reconstruction procedure. The assessment of the system was conducted experimentally, and the results were compared with those provided by the three current sources. The associated results show the higher accuracy of the suggested design, when using Gauss Newton (GN) method, in terms of mean-squared error, which was decreased by 75%, and the image correlation coefficient as well.

Heart failure (HF) is considered a global pandemic because of increasing prevalence, high mortality rate, frequent hospitalization and associated economic burden. This study explores a noninvasive method that may help in managing HF patients by predicting HF readmission. Methods: Seismocardiogram (SCG) signal is the low frequency chest vibration produced by the mechanical activity of heart. SCG signal was acquired from 101 patients with HF including in those readmitted to the hospital during the study period. Features were extracted from SCG signals. Several conventional machine learning (ML) models were developed using selected SCG and heart rate variability features. Furthermore, SCG signals were transformed into images using a time-frequency distribution method. Images were used to train a deep learning model. The models were able to predict the readmission status of the HF patients. Results: ML algorithms achieved higher accuracy than the deep learning model in classifying the readmitted and non-readmitted HF patients. K-nearest neighbor (KNN) achieved the highest classification accuracy (89.4% accuracy, 87.8% sensitivity, and 90.1% specificity). The study results suggest that SCG signal may be useful for readmission prediction of HF patients. Significance: Use of SCG signal may help the management of HF patients and improve their quality of life.

The paper focuses on the multi-azimuth interpolation task of inverse synthetic aperture radar (ISAR) images for aircraft targets and complements incomplete ISAR image datasets. ISAR image automatic target recognition (ATR) has been widely applied in remote sensing and many fields. However, the imaging process is more challenging when compared to capturing optical and SAR image data, which reduces the accuracy and generalization performance of the ATR system. Therefore, in this paper, we leverage existing limited ISAR data to achieve autonomous data expansion. This approach helps mitigate the impact of low sample quantity and unbalanced distribution, ultimately improving the accuracy of the ATR system for target recognition. Most existing methods use generative networks for ISAR image expansion, but few focus on generating ISAR images with specific azimuths. The paper proposes a novel two-stage coarse-to-fine framework for ISAR object view interpolation (C2FIPNet) that combines flow estimation and GAN to interpolate ISAR images with intermediate azimuths using a set of ISAR image pairs. Flow estimation is employed for coarsegrained generation, determining the position and intensity of strong scattering points in the ISAR image. The GAN, on the other hand, is used for fine-grained completion to correct image distortion caused by flow estimation and enhance image details. Additionally, a suitable loss function is designed, incorporating both global and local features, allowing for priority generation in the region of strong scattering points. In conclusion, extensive simulation and comparative experiments have demonstrated that the interpolated ISAR images generated by the proposed C2FIPNet exhibit greater pixel-level authenticity.

In this study we systematically investigate a range of nonlinear functions in the fully-connected feed-forward portions of Vision Transformer (ViT) machine learning models. We limit our investigation to only GLU-type nonlinear functions, as the GLU-type function SwiGLU is the current standard in state-ofthe-art language and vision models. We identify 21 candidate layer functions with 1, 2, or 3 weight matrices, utilizing the activations: Sigmoid, Tanh, and Sin. ViT-Tiny models are implemented utilizing these functions and benchmarked on the image classification datasets CIFAR10, CIFAR100, and SVHN. Through these experiments we identify several previously uninvestigated functions which consistently outperform SwiGLU on all benchmarks. The most performant of these functions we call SinGLU. We further benchmark SwiGLU and SinGLU on the image classification dataset Imagenet64, and again SinGLU performs better. We note that periodic functions such as Sin are not common in neural networks. We perform a numerical investigation into sinusoidally activated neurons and suggest that their viability in Vision Transformers may be due to the losslandscape smoothing of Layer Normalization, Multi-Headed Self Attention (MHSA), and modern data augmentations such as label smoothing. However, our experiments on the CIFAR100 dataset indicate that layer nonlinearity remains a hyperparameter, with approximately piecewise-linear functions performing better than more complex functions when the problem space is not densely sampled.

Supercomputers poised to crack current encryption standards within a decade, traditional methods face an unprecedented threat. Training artificial intelligence (AI) models on plaintext data has sparked increasing concerns regarding privacy and security. The potential risks include the possibility of data leakage or theft. To address these critical challenges, we propose a groundbreaking architecture that seamlessly integrates homomorphic encryption (HE) and AI, enabling privacypreserving analysis of sensitive audio data and paving the way for a new era of secure AI applications in audio processing. Our approach introduces novel approximations for the sigmoid (Asigmoid) and rectified linear unit (A-ReLU) activation functions, designed to be compatible with the input data distribution and to minimize the percentage of squared error (PSE) between the approximation and the original activation function, optimizing processing efficiency on encrypted data while overcoming the limitations of existing methods. We underscore the crucial role of activation function selection and HE's parameter tuning in achieving a balance between computational efficiency and model accuracy within this framework. The evaluation results using the audioMNIST and musical instrument datasets demonstrate the system's robustness with a negligible 0.04% difference between plaintext and ciphertext conditions, highlighting its promising potential for secure audio data processing in various applications.

In the contemporary era of rapid technological advancement, the Industrial Internet of Things (IIoT) has become a pivotal element in revolutionizing industrial operations. This paper delves into the escalating cybersecurity challenges posed by the sprawling networks of IIoT, accentuating the inadequacy of traditional cybersecurity methods in the face of sophisticated cyber threats. We introduce machine learning (ML) as a transformative approach to fortify the cybersecurity landscape of IIoT systems. Our research primarily focuses on the application of machine learning algorithms to detect, analyze, and counteract diverse cyber threats in IIoT environments. These algorithms are trained to recognize and respond to a spectrum of cyber threats, thereby enhancing the resilience of IIoT networks. We present a novel Convolutional-GRU autoencoder model, which demonstrates superior performance over traditional machine learning models in terms of accuracy, precision, recall, and F1score. This model is adept at learning and adapting from complex data patterns, ensuring robust defense against cyber intrusions. We also address the challenges in applying ML to IIoT cybersecurity, considering the varied nature of IIoT devices and the dynamic landscape of cyber threats. This study is an important stride towards enhancing IIoT cybersecurity, highlighting the symbiotic relationship between ML and IIoT. It serves as a foundation for future research and a guide for current implementations, aiming to create more secure, reliable, and efficient IIoT environments. By exploring the potential of ML in cybersecurity, we pave the way for a new era in industrial digital protection, one that is adaptable, forward-thinking, and resilient against the ever-evolving digital threats.

Functional near-infrared spectroscopy (fNIRS) is a non-invasive technique for monitoring brain activity. To better understand the brain, researchers often use deep learning to address the classification challenges of fNIRS data. Our study shows that while current networks in fNIRS are highly accurate for predictions within their training distribution, they falter at identifying and excluding abnormal data which is out-ofdistribution, affecting their reliability. We propose integrating metric learning and supervised methods into fNIRS research to improve networks capability in identifying and excluding out-of-distribution outliers. This method is simple yet effective. In our experiments, it significantly enhances the performance of various networks in fNIRS, particularly transformer-based one, which shows the great improvement in reliability. We will make our experiment data available on GitHub.

The rise in robotics technology has led to increased interest in three-wheeled mobile robots (TWMRs) due to their agility and adaptability across various applications. However, effectively controlling TWMRs presents a significant challenge owing to their inherent nonholonomic constraint, which restricts their independent movement in all directions. Additionally, factors like sensor noise, nonlinear system dynamics, and uncertain system parameters add to the complexity controlling of TWMRs. This research endeavors to enhance the precision of trajectory tracking in TWMRs. Specifically, it employs Backstepping Fuzzy Sliding Mode Control (BFSMC) with parameters optimized through Particle Swarm Optimization (PSO), coupled with the Extended Kalman Filter (EKF) for state estimation. The study conducts a comprehensive performance comparison between BFSMC and BSMC across various trajectory patterns, revealing substantial improvements in trajectory tracking accuracy with BFSMC. BFSMC demonstrates improved performance compared to BSMC across various trajectory types, quantified by calculating the percentage improvement in trajectory tracking using Integral Absolute Error (IAE). Specifically, it achieves a 51.97% improvement for circular trajectories, an 82.09% improvement for infinity trajectories, and an 84.073% improvement for spiral trajectories.. Moreover, BFSMC demonstrates superior robustness in the presence of disturbances, noise, parameter variations, and unmodeled dynamics compared to BSMC. The integration of the Extended Kalman Filter further improve accuracy, particularly in noisy conditions. Simulation results conducted using MATLAB/Simulink software validate the effectiveness of this approach in achieving superior trajectory tracking accuracy in TWMRs.

Identification of sources of seizures in the brain is of paramount importance, particularly for drug-resistant epilepsy patients who may require surgical operation. Interictal epileptiform discharges (IEDs), which may or may not be frequent, are known to originate from seizure networks. Delayed responses (DRs) to brain electrical stimulation have been recently discovered. If DRs and IEDs come from the same location and the DRs can be accurately localized, there will be a significant step in the identification of the source of seizures. The solution to this important question has been investigated in this paper. For this, we have exploited the morphology of these spike-type, events as well as the variability in their temporal location, to develop new constraints for an adaptive Bayesian beamformer that outperforms the conventional and recently proposed beamformers. This beamformer is applied to an array (a.k.a mat) of cortical EEG electrodes. As the significant outcome of applying this beamformer, it is very likely (if not certain) that the IEDs and DRs for an epileptic subject originate from the same location in the brain. This paves the way for a quick identification of the source(s) of seizure in the brain.

Intelligent Transportation Systems (ITS) operate within a highly intricate and dynamic environment characterized by complex spatial and temporal dynamics at various scales, further compounded by fluctuating conditions influenced by external factors such as social events, holidays, and weather conditions. Navigating the intricacies of modeling the intricate interaction among these elements, creating universal representations, and employing them to address transportation issues presents a significant endeavor. Yet, these intricacies comprise just one facet of the multifaceted trials confronting contemporary ITS. This paper offers an all-encompassing survey exploring Deep learning (DL) utilization in ITS, primarily focusing on practitioners' methodologies to address these multifaceted challenges. The emphasis lies on the architectural and problem-specific factors that guide the formulation of innovative solutions. In addition to shedding light on the state-of-the-art DL algorithms, we also explore potential applications of DL and large language models (LLMs) in ITS, including traffic flow prediction, vehicle detection and classification, road condition monitoring, traffic sign recognition, and autonomous vehicles. Besides, we identify several future challenges and research directions that can push the boundaries of ITS, including the critical aspects of explainability, transfer learning, hybrid models, privacy and security, and ultra-reliable low-latency communication. Our aim for this survey is to bridge the gap between the burgeoning DL and transportation communities. By doing so, we aim to facilitate a deeper comprehension of the challenges and possibilities within this field. We hope that this effort will inspire further exploration of fresh perspectives and issues, which, in turn, will play a pivotal role in shaping the future of transportation systems.

Current research uses TOPSIS to evaluate 14 Cricket World Cup 2023 teams. Data from the Espn Cricinfo website was used in this analysis. A comprehensive set of criteria (P1 to P11) was used to evaluate each squad, encompassing various game aspects. A numerical labeling system (A1 to A14) and parameter system (P1 to P11) were used to idenMfy team names and qualiMes more efficiently. The research calculates the normalized matrix and weighted matrix, then finds the best and worst values using TOPSIS. A normalized matrix creates a consistent and uniform framework for evaluaMng and comparing factors, ensuring imparMality and jusMficaMon. In contrast, the weighted matrix integrates each criterion's proporMonal importance into the evaluaMon process. For each criterion, the ideal best and ideal worst values indicate the best and worst performance. The TOPSIS analysis placed Australia first, Bangladesh second, and New Zealand third. In fourth and fiXh place were India and Sri Lanka. Afghanistan, West Indies, England, South Africa, and Pakistan rated sixth to tenth. Nepal was tenth, Ireland, the US, and Zimbabwe fourteenth.To understand team performance, the TOPSIS technique must be accepted. It is important to acknowledge that the Cricket World Cup 2023 results may vary owing to many factors. This study provides a systematic and comprehensive approach to team performance, making it a useful resource for cricket fans and experts interested in the event's competitive dynamics.

Time measurements are challenging in electronics 1 given their various applications. The main focus lies not in achieving greater precision, as conventional architectures have already reached picosecond levels. Instead, the challenge stems from the use of low resources and the substantial expansion in the number of channels. This study presents a novel architecture for the implementation of TDCs in applications where resources are constrained. The introduced FPGA-based TDC offers a resolution of 415.84 ps, a single-shot precision of 0.45 LSB (186 ps r.m.s), while maintaining a minimal resource occupancy. Built upon a multi-shift phase counter, the TDC is extended with a tap delay using the input delay available in the FPGA hardware input, doubling the resolution of the TDC. The resource utilization is minimized when compared to low-resources state-of-the-art TDCs. The number of LUTs has been reduced up to 102, and the number of registers to 213. Furthermore, the presented TDC exhibits favorable DNL (0.2 LSB) and INL 17 (0.15 LSB). The TDC has been successfully implemented on an Artix7-2 FPGA from Xilinx. This design provides a resource-effective solution for applications requiring high precision and  low resource consumption.

In this paper, we propose a generalized Moreau enhanced minimization induced (GME-MI) regularization model and its proximal splitting algorithm for further improvement of the MI penalty derived as the minimum of a convex function. We first design the GME-MI penalty function by applying the GME construction to the MI penalty, and derive an overall convexity condition for the GME-MI regularized least-squares model. Then, under the overall convexity condition, characterizing the solution set of the GME-MI model with a carefully designed averaged nonexpansive operator, we develop a proximal splitting algorithm which is guaranteed to converge to a globally optimal solution. Numerical examples demonstrate the effectiveness of the proposed approach.

This paper introduces a new spatial domain-based self-interference cancellation (SIC) precoding method named constrained minimum mean square error (C-MMSE) for an asymmetric massive multiple-input multiple-output (mMIMO) full-duplex (FD) system. The main idea is to translate the commonly used singular value decomposition (SVD)-based null-space projection approach, which is unfeasible in our considered system model, into an optimization problem under MMSE criterion, where additional constraints are implemented to perform SIC. Theoretical derivation of the C-MMSE precoder is presented, followed by performance comparison with conventional MMSE precoding, where no constraints are added for SIC. We theoretically show that the C-MMSE scheme outperforms the conventional one in terms of SIC, and allows the FD system to work under an almost interference-free environment. Additionally, we also assess the performance of the proposed method under imperfect channel state information (CSI), to further evaluate the robustness of our spatial precoder in more realistic conditions. We show that the C-MMSE precoder outperforms MMSE in terms of interference suppression ratio (ISR), even in CSI imperfection. Additionally, the C-MMSE achieves the same spectral efficiency (SE) as an hypothetical perfect SIC in a wide SNR range, whereas the MMSE is upper bounded in large SNR range.

A document by Paola Arrubarrena. Click on the document to view its contents.

Functional near-infrared spectroscopy (fNIRS) is a valuable non-invasive tool for monitoring brain activity. The classification of fNIRS data in relation to conscious activity holds significance for advancing our understanding of the brain and facilitating the development of brain-computer interfaces (BCI). Many researchers have turned to deep learning to tackle the classification challenges inherent in fNIRS data due to its strong generalization and robustness. In the application of fNIRS, reliability is really important, and one mathematical formulation of the reliability of confidence is calibration. However, many researchers overlook the important issue of calibration. To address this gap, we propose integrating calibration into fNIRS field and assess the reliability of existing models. Surprisingly, our results indicate poor calibration performance in many proposed models. To advance calibration development in the fNIRS field, we summarize three practical tips. Through this letter, we hope to emphasize the critical role of calibration in fNIRS research and argue for enhancing the reliability of deep learning-based predictions in fNIRS classification tasks. All data from our experimental process are openly available on GitHub.

In this paper, we explore an indoor downlink cooperative hybrid visible light communication (VLC)/radio frequency (RF) scenario using a relay node to reduce system outage probability. In particular, information can be transmitted to the end user either directly through the VLC link or via the relay node. To re-transmit the decoded information to the end user through the RF link the relay utilizes harvested energy from the source light emitting diode (LED) at the ceiling. We derive the analytical expression for the outage probability of the relayaided hybrid VLC/RF system, considering the randomness of location and receiver orientation for both the relay and the end user. Furthermore, we investigate the effects of the direct current (DC) bias, data rate threshold, and different distributions for the location and orientation of the end user and relay on the outage probability of the system.

Speech overlap, which occurs when multiple people speak simultaneously, poses a significant challenge in audio and speech processing. The presence of overlapping speech segments significantly degrades the performance of technologies such as Automatic Speech Recognition (ASR), speaker identification, and diarization systems. This degradation in performance becomes more significant in diverse acoustic environments with background noise and reverberation. To effectively address this issue, we introduce BiConNet. This novel dual-branch architecture combines the strengths of Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) for robust detection of overlapping speech in diverse acoustic conditions. The CNN branch is used for frame-level spectral feature extraction, while the BiLSTM branch captures temporal dependencies from both forward and backward directions. Features from both branches are concatenated, resulting in a robust feature representation. We also examined the impact of Mel Frequency Cepstral Coefficients (MFCC), Gammatone Frequency Cepstral Coefficients (GTCC), and Power Normalized Cepstral Coefficients (PNCC) as spectral-based features on BiConNet's performance. To validate its effectiveness in various acoustic environments, we developed a constructed data set derived from the GRID corpus, including conversations with different gender combinations and recording conditions, such as clean, noisy, reverberant, and combined noise and reverberation conditions. Experimental results show that BiConNet outperforms various state-of-the-art methods in detecting overlapping speech segments under these conditions. Furthermore, our analysis of computational efficiency reveals that BiConNet provides competitive training and inference times, demonstrating its practicability for real-world applications.

This paper describes a method for automatically transforming the structure and characteristics of an image processing dataflow graph for the purpose of improving performance and/or lowering memory utilization as compared to the baseline tools. Embedded image processing applications are often executed on Digital Signal Processors, or their modern equivalent Visual Processor Units. The software usually performs a series of pixel-level operations for basic color conversion, channel extraction and combining, arithmetic, and filtering. These steps can often be efficiently described as a graph. For this reason, standard libraries such as OpenVX are used, which provide a graph-based programming model where the nodes are chosen from a repertoire of common pixel-level operations and the edges represent the flow of images as they progress though the processing stages. Generally speaking, each node is processed sequentially in the order implied by the data dependencies defined by the graph structure, with all intermediate values stored in external memory. In the proposed framework, we developed performance models for both the direct memory access subsystem and the L1 data cache to allow for selection of certain intermediate values to be stored in on-chip scratchpad memory as well as selecting the most appropriate tile size. In this way, we effectively decompose the graph in a way to fuse specific sets of nodes to associate their internal edges with on-chip buffers. Additionally, the tile size is optimized for each fused set of nodes. In this paper, we describe our performance models and approach for graph decomposition and tile size selection. The proposed performance models are accurate to within 2% on average, and the overall approach of graph optimization achieves an average speedup of 1.3 and allows for reduction of average DRAM utilization from 100% to as low as 15%.

Cryptography has become an essential tool in information security, preserving data confidentiality, integrity, and availability. However, despite rigorous analysis, cryptographic algorithms may still be susceptible to attack when used on real-world devices. Side-channel attacks (SCAs) are physical attacks that target cryptographic equipment through quantifiable phenomena such as power consumption, operational times, and EM radiation. These attacks are considered to be a significant threat to cryptography since they compromise the integrity of the algorithm by obtaining the internal cryptographic key of a device by seeing its physical implementation. The literature on SCAs has focused on real-world devices, yet with the growing popularity of sophisticated devices like smartphones, fresh approaches to SCAs are necessary. One such approach is electromagnetic side-channel analysis (EM-SCA), which gathers information by listening to electromagnetic (EM) radiation. EM-SCA has been demonstrated to recover sensitive data like encryption keys and has the potential to identify malicious software, retrieve data, and identify program activity. This study aims to evaluate how well EM-SCA compromises encryption under various application scenarios, as well as examine the role of EM-SCA in digital forensics and law enforcement. Regarding this, addressing the susceptibility of encryption algorithms to EM-SCA approaches can provide digital forensic investigators with the tools they desire to overcome the challenges posed by strong encryption, allowing them to continue playing a crucial role in law enforcement and the justice system. Furthermore, this paper seeks to define the current state of EM-SCA in terms of attacking encryption, the encryption algorithms and encrypted devices that are most vulnerable and resistant to EM-SCA, and the most promising EM-SCA on encryption approaches. This study will provide a comprehensive analysis of EM-SCA in the context of law enforcement and digital forensics and point towards potential directions for further research.

The application of millimeter-Wave (mmWave) Radar sensors for people monitoring raised a lot of interest in the context of Active Assisted Living (AAL), especially since the processing of Radar signals can provide interesting information about the observed subjects. Correct recognition of the ongoing behavior, however, cannot disregard from detecting where the subject is positioned. Detection approaches, based on Constant False Alarm Rate (CFAR) algorithms, sometimes fail to correctly identify the presence of targets within the observed scenario, especially in complex environments such as indoors. This paper proposes the use of a mmWave Multiple Input Multiple Output (MIMO) Radar in combination with a You Only Look Once (YOLO) neural network-based algorithm for the detection of moving people in indoor environments by processing all the data cube information at the same time. Results are validated through experimental tests which involve subjects walking in linear or random mode, different Radar configurations, and different indoor environments. By exploiting at the same time information such as the angle, Doppler, and range distance of the target, the proposed approach proves to be very effective in the examined scenarios. Experimental results will be discussed in this work to demonstrate the effectiveness of the proposed method.

Motor unit (MU) decomposition, generally, requires a time-consuming and labor-intensive manual inspection/editing process from human operators to ensure high accuracy. In this study, we propose and validate a rule-based auto-editing method that could potentially substitute the manual process. Methods: The proposed auto-editing framework (autoeditor) consists of four main rules for adding and removing spikes based on the height of the innervation pulse train (IPT) and the regularity of the firing rate of the identified motor unit. The rules were optimized and validated based on an open-source database including raw MU spike trains estimated from the convolution kernel compensation method and the manually edited MU spike trains from eight human operators. Results: Across 110 motor units, the average rate of agreement between the auto-editor and human operators reached 99.2% after the auto-editor corrected more than 10 edits for each motor unit on average from the raw spike trains. More importantly, the characteristics of the motor unit behaviors, including the MU firing rate and recruitment threshold, were consistent across human operators and the proposed auto-editor. Conclusion: With a simple but effective rulebased auto-editing framework, comparable performance in MU refinement was achieved as human operators. Significance: The proposed auto-editing framework has the potential to standardize the MU editing practice, lower the requirements for expert knowledge and specialized training for MU decomposition, and provide an expandable framework allowing contributions from the community.