International Journal of Distributed Sensor Networks

Uneven water-air media distribution or irregular liquid flow can cause changes in light propagation, leading to blurring and distortion of the extracted image, which presents a challenge for object recognition accuracy. To address these issues, this paper proposes a repair network to correct object image distortion in water-air cross-media. Firstly, convolutional combination performs feature extraction on water-air cross-media images, which retains the same features at the same scale and marks feature points with large differences. Then, an attention correction module for geometric lines is proposed to correct geometric lines in water-air cross-media images by comparing and sensing the marked feature points with large differences and utilizing the line similarity in positive and negative samples. Finally, the blurring artifact elimination module eliminates artifacts caused by image blurring and geometric line correction by using multiscale fusion of individual U-Net information streams. This completes the image restoration of object distortion under water-air cross-media. The proposed method is feasible and effective for restoring aberrated objects in water-air cross-media environments, with numerous experiments conducted on water-air cross-media image datasets.

Implantable LED light sources have received a lot of attention in the field of optogenetic neuromodulation. This type of light modulation enables effective stimulation of neurons. However, as optogenetics moves towards clinical trials, combinatorial photostimulation based on arrays of LED light sources has emerged. This approach inevitably brings about a large increase in transient temperature, resulting in the inability to achieve precise stimulation of cells. We designed a wireless optogenetic hardware system to realize the control of the stimulation mode and temperature of the light source array. At the same time, a set of combined photothermal physical model was established to simulate the photothermal response of the whole experiment. The physical model can effectively guide the wireless optogenetic hardware circuits to perform effective stimulation within a controlled temperature range. Our model provides a new technical approach for photothermal studies in optogenetic clinical trials.

Wireless ad hoc/sensor networks (WASN) have seen increased application in three-dimensional (3D) environments, such as underwater and aerial scenarios. However, WASN may be fragmented or unable to connect continuously due to the harsh surrounding environment or high mobility. Therefore, restoring the network connectivity and transmit data in real time is very important. This paper focuses on the critical task of restoring network connectivity in 3D WASN, a complex issue given that existing connectivity restoration algorithms for two-dimensional environments are not directly applicable or become overly complicated in 3D contexts. We present a comprehensive analysis of the current research landscape, summarizing key findings related to various aspects of 3D WASN connectivity restoration. These aspects include the application environment (underwater, in the sky, and recovery disaster), opportunities for restoration (active, passive, and active/passive), implementation strategies (clustering and sleep scheduling), and resource constraints (node deployment and movement control). Our study also proposes a classification of connectivity restoration solutions for 3D WASN, identifying existing gaps and suggesting potential future research directions. By providing specific insights and a structured overview of the field, we aim to contribute to the ongoing development of robust and resilient 3D WASN.

During sonar detection, the fuzzy function of hyperbolic frequency modulation (HFM) in the waveform of active signal is in the shape of a “oblique blade,” which leads to the coupling characteristics of its fuzzy function in the time dimension and frequency dimension. Doppler causes a frequency shift in the frequency dimension and a time delay in the time dimension, which makes it impossible for a single HFM signal to measure speed accurately. However, the time delay caused by the same moving target to HFM signals with different frequency bands and different pulse widths is also different. A velocity measurement method combining HFM signals (VCH) is proposed, which employs the time delay correlation between the combined HFM signals to achieve the target distance and speed. In the VCH method, the echo signal is no longer matched and filtered as a whole but is divided into two channels and matched and filtered separately. And then, the combined HFM signals are used to obtain the distance and speed of the target. Extensive simulation results show that the proposed method can estimate the distance and speed of moving targets accurately, and it has reference value for engineering application.

The low-bit quantization can effectively reduce the deep neural network storage as well as the computation costs. Existing quantization methods have yielded unsatisfactory results when being applied to lightweight networks. Additionally, following network quantization, the differences in data types between the operators can cause issues when deploying networks on Field Programmable Gate Arrays (FPGAs). Moreover, some operators cannot be accelerated heterogeneously on FPGAs, resulting in frequent switching between the Advanced RISC Machine (ARM) and FPGA environments for computation tasks. To address these problems, this paper proposes a custom network quantization approach. Firstly, an improved PArameterized Clipping Activation (PACT) method is employed during the quantization aware training to restrict the value range of neural network parameters and reduce the loss of precision arising from quantization. Secondly, the Consecutive Execution Of Convolution Operators (CEOCO) strategy is utilized to mitigate the resource consumption caused by the frequent environment switching. The proposed approach is validated on Xilinx Zynq Ultrascale+MPSoC 3EG and Virtex UltraScale+XCVU13P platforms. The MobileNetv1, MobileNetv3, PPLCNet, and PPLCNetv2 networks were utilized as testbeds for the validation. Moreover, experimental results are on the miniImageNet, CIFAR-10, and OxFord 102 Flowers public datasets. In comparison to the original model, the proposed optimization methods result in an average decrease of 1.2% in accuracy. Compared to conventional quantization method, the accuracy remains almost unchanged, while the frames per second (FPS) on FPGAs improves by an average of 2.1 times.

For the characteristics of channel instability in wireless sensor networks, this paper proposes an intrusion detection algorithm based on FedAvg (federated averaging) and XGBoost (extreme gradient boosting) wireless sensor networks using fog computing architecture. First, the network edge is extended by introducing fog computing nodes to reduce the communication delay. It reduces the transmission bandwidth and privacy leakage risk while improving the accuracy of jointly learned global and local models. Then, the histogram-based approximation calculation method is improved to adapt to the unbalanced data characteristics of wireless sensor networks. Finally, by introducing TOP-K gradient selection, the number of model parameter uploads is minimized, and the efficiency of model parameter interaction is improved. The experimental results show that this algorithm has superior detection performance and low energy consumption. It is also compared with other algorithms to demonstrate the high detection rate and low computational complexity of this algorithm.