This paper presents a channel-wise average pooling and one dimension pixel-shuffle architecture for a denoising autoencoder (CPDAE) design that can be applied to efficiently remove electrode motion (EM) artifacts in an electrocardiogram (ECG) signal. The three advantages of the proposed design are as follows: (1) In the skip connection layer, less memory is needed to transfer the features extracted by the neural network; (2) Pixel shuffle and pixel unshuffle techniques with point-wise convolution are used to effectively reserve the key features generated from each layer in both the encoder and decoder; (3) Overall, fewer parameters are required to reconstruct the ECG signal. This paper describes three deep neural network models, namely CPDAE_Lite, CPDAE_Regular, and CPDAE_Full, which support various computational capacity and hardware arrangements. The three proposed structures involve an encoder and decoder with six, seven, and eight layers, respectively. Furthermore, the CPDAE_Lite, CPDAE_Regular, and CPDAE_Full structures require fewer multiply-accumulate operations—355.01, 56.96, and 14.69 million, respectively—and less parameter usage—2.69 million, 149.7 thousand, and 55.5 thousand, respectively. To evaluate the denoising performance, the MIT–BIH noise stress test database containing six signal-to-noise ratios (SNRs) of noisy ECGs was employed. The results demonstrated that the proposed models had a higher improvement of SNR and lower percentage root-mean-square difference than other state-of-the-art methods under various conditions of SNR.

Rodent infestations are a common problem that can result in several issues, including diseases, damage to property, and crop loss. Conventional methods of controlling rodent infestations often involve using mousetraps and applying rodenticides manually, leading to high manpower expenses and environmental pollution. To address this issue, we introduce a system for remotely monitoring rodent infestations using Internet of Things (IoT) nodes equipped with Long Range (LoRa) modules. The sensing nodes wirelessly transmit data related to rodent activity to a cloud server, enabling the server to provide real-time information. Additionally, this approach involves using images to auxiliary detect rodent activity in various buildings. By capturing images of rodents and analyzing their behavior, we can gain insight into their movement patterns and activity levels. By visualizing the recorded information from multiple nodes, rodent control personnel can analyze and address infestations more efficiently. Through the digital and quantitative sensing technology proposed at this stage, it can serve as a new objective indicator before and after the implementation of medication or other prevention and control methods. The hardware cost for the proposed system is approximately USD 43 for one sensor module and USD 17 for one data collection gateway (DCG). We also evaluated the power consumption of the sensor module and found that the 3.7 V 18,650 Li-ion batteries in series can provide a battery life of two weeks. The proposed system can be combined with rodent control strategies and applied in real-world scenarios such as restaurants and factories to evaluate its performance.

This paper presents a novel low-complexity impedance calculation based on a compact recursive discrete Fourier transform (RDFT) for electrical impedance spectroscopy (EIS) system design. Compared to a traditional fast Fourier transform, an arbitrary-length RDFT is not limited by the power-of-two transform length. By choosing a suitable integer ratio of the desired frequency/frequency resolution, the spectrum leakage can be significantly reduced by 4-Hz resolution per bin, which enables a lower magnitude relative error (MRE) and a phase absolute error (PAE) in impedance calculations. Here, the practical measurement results show that the maximum MRE and PAE are below 0.1% and 0.5°, respectively, making this approach superior to previously reported ones. Compared with the state-of-the-art Goertzel filter-based EIS system, the proposed method provides more frequency bins (+43.75%) within the same frequency band, and the operations in terms of square root, division, and arctangent are drastically reduced by 50%, 33%, and 50%, respectively. In addition, the proposed system only costs $474 USD and can serve as a suitable solution for various EIS applications in the future.

This study presents a novel hybrid algorithm merging Recursive Discrete Fourier Transform (RDFT) and Fast Fourier Transform (FFT) to meet the demands of Electrochemical Impedance Spectroscopy (EIS) applications. It particularly addresses scenarios where users need to calculate a reduced number or a variable quantity of frequency bins. By offering the flexibility for users to choose either RDFT or FFT based on their specific requirements, the algorithm overcomes the limitation encountered in traditional impedance calculation methods, where the FFT transform length is constrained to being a power of two. To ensure more precise observation of the Device Under Test (DUT) impedance spectrum, we set the sampling rate (Fs) and transform length (N) to 40960, achieving a frequency resolution of 1Hz. Unlike the traditional Radix-2 FFT method, the proposed hybrid algorithm eliminates the need for zero padding to supplement the signal to 65536 points. Consequently, the overall computation involves 1,163,264 multiplications and 2,359,296 additions, marking a notable reduction of 44.52% in multiplications and 43.94% in additions compared to the traditional radix-2 FFT method’s 2,097,152 multiplications and 4,194,304 additions. In order to validate measurement accuracy, experiments were performed using six pure resistors and two RC circuits. The RDFT-based impedance calculations, after averaging all measurements, resulted in magnitude relative error (MRE) and phase absolute error (PAE) values below 0.11% and 0.21°, respectively. Likewise, impedance calculations utilizing the hybrid algorithm showcased even greater precision, with MRE below 0.06% and PAE below 0.09°. These findings highlight the superior accuracy and streamlined complexity provided by the proposed system. For hardware implementation, the system costs only $352 USD, offering a cost-effective solution for various high-precision EIS measurement applications.

The Electrocardiogram (ECG) test detects and records cardiac-related electrical activity of the heart. The ECG test identifies and documents cardiac-related electrical activity in the heart. The use of ECG signals for cardiovascular disease nursing as a crucial component of preoperative evaluation is increasing. ECG signals need to denoise and display in a clear waveform due to the numerous noises. We have introduced Compact Shortcut Denoising Auto-encoder (CS-DAE) neural network, which reduces the noise from ECG signals. The Compact Shortcut approach compresses the features passed through the shortcut layers, which lowers the operation’s memory needs and improves the noise reduction impact. In addition, the encoder and decoder process the Pixel-Unshuffled and Pixel-Shuffled, which effectively mitigates the feature loss caused by down-sampling and up-sampling operations. As a result, the CS-DAE algorithm decreases the computation and required memory size while maintaining higher accuracy. We have used MITDB and NSTDB datasets for training and testing the proposed CS-DAE model, resulting in the average Percentage of Root Mean Square Difference (PRD) being 46.30% and the improvement of Signal-to-Noise Ratio (SNR_imp) being 10.50. In addition, we have designed VLSI architect ure for the proposed CS-DAE neural network to accelerate low hardware cost and less computation. The TUL PYNQTM-Z2 development platform runs the Verilog code, which is used for VLSI architecture and has the lowest power consumption of 1.65W.

Electrocardiograms (ECGs) are critical for cardiovascular disease diagnosis, but their accuracy is often compromised by electrode motion (EM) artifacts—large, nonstationary distortions caused by patient movement and electrode-skin interface shifts. These artifacts overlap in frequency with genuine cardiac signals, rendering traditional filtering methods ineffective and increasing the risk of false alarms and misdiagnosis, particularly in wearable and ambulatory ECG applications. To address this, we propose the Feature-Shuffle Multi-Head Attention Autoencoder (FMHA-AE), a novel architecture integrating multi-head self-attention (MHSA) and a feature-shuffle mechanism to enhance ECG denoising. MHSA captures long-range temporal and spatial dependencies, while feature shuffling improves representation robustness and generalization. Experimental results show that FMHA-AE achieves an average signal-to-noise ratio (SNR) improvement of 25.34 dB and a percentage root mean square difference (PRD) of 10.29%, outperforming conventional wavelet-based and deep learning baselines. These results confirm the model’s ability to retain critical ECG morphology while effectively removing noise. FMHA-AE demonstrates strong potential for real-time ECG monitoring in mobile and clinical environments. This work contributes an efficient deep learning approach for noise-robust ECG analysis, supporting accurate cardiovascular assessment under motion-prone conditions.