With the exponential growth of the Internet of Things (IoT) landscape and the resulting spectrum congestion, innovative techniques for spectrum monitoring are crucial. This paper presents a groundbreaking approach to spectrum monitoring harnessing the power of spiking neural networks (SNNs) with a focus on image segmentation using the UNet architecture. Traditional methods, including energy detection, have been widely used but are not without challenges, especially in environments with varying signal-to-noise ratios. In contrast, the presented SNN approach in this paper leverages the leaky integrate-and-fire neuron model and provides superior energy efficiency, real-time inference capability, and higher detection performance. Through extensive simulations, the proposed SNN framework exhibited performance metrics that significantly surpass energy detection methods and closely align with conventional convolutional neural network techniques. Future explorations will delve into enhancing the framework using machine learning techniques for advanced feature extraction and multi-class segmentation.
An unidentified continuous functions is an function that looks infinitely differentiable from far away view. Here any unidentified continuous function is assumed to have either digital continuum (non-archimedian continuum) or archimedian continuum. Author mainly found an method to indicate an distinction for both continuum models. This method can be used on to unidentified continuous functions in order to obtain a discrete unit value relation of coarse structure. An example could be energy-frequency equation for photons.
Signal decomposition techniques aim to break down nonstationary signals into their oscillatory components, serving as a preliminary step in various practical signal processing applications. This has motivated researchers to explore different strategies, yielding several distinct approaches. A wellknown optimization-based method, the Variational Mode Decomposition (VMD), relies on the formulation of an optimization problem, utilizing constant bandwidth Wiener filters. However, this poses limitations in constant bandwidth and the need for constituent count. In this paper, a new method, namely Dynamic Bandwidth VMD (DB-VMD), is proposed to generalize VMD by addressing the Wiener filter limitations through enhancement of the optimization problem with an additional constraint. Experiments in synthetic signals highlight DB-VMDâ\euro™s noise robustness and adaptability in comparison to VMD, paving the way for many applications, especially when the analyzed signals are contaminated with noise.
The continuous emergence of new wireless communication systems increases the need for intelligent planning prior to their deployment. This planning phase includes determining the position and transmit power of wireless access points, to meet quality of service objectives along with electromagnetic compatibility standards. How to effectively place a set of transmitters in an environment is a long-standing problem that has been widely studied over the years. This process has mainly relied on expensive measurement campaigns, low-fidelity empirical models, or high-fidelity but time-consuming simulations. Recent advances in scientific machine learning create new possibilities for overcoming the standard dichotomy between speed and accuracy. In this paper, we train a deep neural network (U-Net) to predict received signal strength levels for a variety of different geometries and for positions of multiple transmitters. Then, we leverage the computational efficiency of the trained model to determine the position of access point transmitters in new geometries. This approach dramatically accelerates the process of selecting the position of access points, meeting multiple optimization objectivesin an efficient manner.Â
This review focuses on the integration of intelligent driving and intelligent cockpit systems. With the advancement of intelligent levels, autonomous vehicles are capable of driving in complex driving context. However, the lack of systems that consider both driver and driving scenarios information reduces the success probability of decision-making of autonomous vehicles in driving context where driver and driving scenarios information are tightly coupled during the human-vehicle cooperation stage. To solve this problem, we present the Cockpit-Driving Integration system (CDI) and review perception and decision-making algorithms for CDI systems. Additionally, to achieve human-centric autonomous vehicles, we propose that the CDI system should consider the personalized characteristics of drivers. Finally, we present a framework for CDI systemsÂ
Passive exoskeletons have been introduced to alleviate loading on the lumbar spine while increasing the wearerâ\euro™s productivity. However, few studies have examined the neurocognitive effects of short-term human-exoskeleton adaptation. The objective of the study was to develop a novel neural efficiency metric to assess short-term human exoskeleton adaptation during repetitive lifting. Twelve participants (gender-balanced) performed simulated asymmetric lifting tasks for a short duration (phase: early, middle, late) with and without a passive low back exoskeleton on two separate days. Phase, exoskeleton condition, and their interaction effects on biomechanical parameters, neural activation, and the novel neural efficiency metric were examined. Peak L5/S1 superior lateral shear forces were found to be significantly lower in the exoskeleton condition than the control condition, however other biomechanical and neural activation measures were comparable between conditions. The temporal change of neural efficiency metric was found to follow the motor adaptation process. Compared to the control condition, participants exhibited lower efficiency during the exoskeleton assisted lifting condition over time. The neural efficiency metric was capable of tracking the short-term task adaptation process during a highly ambulatory exoskeleton-assisted manual handling task. The exoskeleton-assisted task was less efficient and demanded longer adaptation period than the control condition, which may impact exoskeleton acceptance and/or intent to use.
Feature fusion is an effective solution for improving image retrieval performance. Although the more feature types, the better accuracy, complexity also increases. Applications in practice typically afford a limited number of feature types. Due to the strong complementarity, global and local features form an ideal combination for many fusion applications. However, the two kinds of features are intrinsically different in nature, thus cannot be fused in a straightforward way. In this work, we propose an integrated image retrieval and feature fusion framework for global and local features. It is based on inverted index fusion, a technique for efficient image retrieval. The core idea is to rank candidates by weighted voting during candidate selection, which is named pre-ranking. This procedure takes place before re-ranking, and is potentially superior to conventional late fusion. Extensive experiments on three public datasets show that the light-weight pre-ranking stage significantly contributes to accuracy, and brings substantial improvement when used together with re-ranking. Our method is robust and versatile, and can be applied to any scenario where inverted indexing is used. It is a promising technique for multimedia retrieval in the big data era.
Perovskite technology has been advancing at unprecedented levels in the last years, with efficiencies reaching up to 25.7%. State of art results are obtained on a very small area scale (<0.1cm2), by adopting high materials wasting processes not compatible with industry and with market exploitation. Silicon is a well-established technology and one of the advantages of Perovskite is its ability to pair with Silicon forming a tandem device that extracts charges reducing transmission and thermalization losses. In this work, we focus on finding a strategy to fabricate 15.2x15.2 cm2 Perovskite modules by avoiding any spin coating deposition and by adopting a green solvent cleaning step. Furthermore, we optimize the ITO top electrode deposition by adjusting sputtering process and buffer layer deposition; finally, we focused on light management by applying an antireflective coating. We obtained a semi-transparent and a tandem Silicon-Perovskite module in 4T configuration on 225cm2 (4T configuration) with 13.18% and 20.99% efficiency, respectively, passing ISOS-L1 (under continuous light soaking in air) test with a remarkable T80 of 1459hÂ
In this letter, we consider a secondary network (SN) in which an ambient backscatter device (BD) utilizes the secondary transmitter (ST) signal to communicate its own information to the secondary destination (SD). Optimizing performance of such networks is complicated by signal reflections by the BD.Â It is shown in this work how the secondary transmit power and the reflection coefficient of the BD can both be jointly optimized using a simple restricted one dimensional search to satisfy the quality of service (QoS) constraints of SN as well as the primary network (PN) while maximizing performance of the backscatter link, which is termed the tertiary network (TN).Â Only statistical channel knowledge is used for this purpose. It is seen that careful optimization can improve spectral efficiency. Simulations validate the derived analytical expressions.
As wireless networks continue to advance and the need for low-latency communication links increases, calibrating antennas installed in the field is essential. In this context, Unmanned Aerial Vehicles (UAVs) are very useful for applications such as UAV-based measurements. Given its light weight, wide bandwidth makes Printed Log-Periodic antenna (PLPDA) an ideal solution as UAV probe. Our study examines the performance of a PLPDA mounted on a UAV. Extensive simulations are performed to determine the optimal position for the PLPDA on a UAV. Simulations are carried out in CST Studio Suite 2022 using the time domain Finite Integration Technique (FIT) with appropriate mesh settings. At the optimized location PLPDA achieves a -10 dB bandwidth of 6.2 GHz.
Intelligent behavior is an emergent phenomenon observed in biological organisms across all scales. It describes the cooperative behavior of low complexity entities to accomplish complex tasks, which exceed their individual capabilities. This property is particularly important for the Internet of Bio-Nano Things (IoBNT), which consists of Bio-Nano Things (BNTs) used in the human body, where they face many restrictions, such as bio-compatibility and size constraints. In this paper, we present a novel BNT-architecture, called Molecular Nano Neural Networks (M3N), which allows the implementation of intelligence on the micro-/nano-scale. The proposed structure consists of compartments (low complexity entities) that are connected to each other to form a network. Based on reaction and diffusion of molecules in and between connected compartments, this network mimics an artificial neural network, which is an important step towards Â artificial intelligence in the IoBNT. We provide design guidelines for the proposed M3N and successfully validate it by applying a regression and classification task.
This piece of research identifies how DTW and some common environmental parameters influence crop yield or vegetation. A total of 14 DTW and its variants and 10 environmental parameters that promote crop yield or vegetation were adopted.Â The work points out the potential gaps untapped within DTW application in crop yield prediction and the efficacy of this model and its variants in forecasting yield.
This paper introduces Continual Learning for Multilingual ASR (CL-MASR), a benchmark for continual learning applied to multilingual ASR. CL-MASR offers a curated selection of medium/low-resource languages, a modular and flexible platform for executing and evaluating various CL methods on top of existing large-scale pretrained multilingual ASR models such as Whisper and AWavLM, and a standardized set of evaluation metrics.
This project is to detect categories from the toy product description and names. The data set is an Amazon (toy) data set with manufacturer specific model, names and description of childrenâ\euro™s toys. The expected outcome includes labeling scheme and using CRF and bi-LSTM to measure the performance of our category extraction. As for the annotation, we performed the labeling and used product description decomposed into sequence of tokens labeled with BIO encoding, and the output of learning algorithm on a product description would be a sequence of labels. Three diffrent kinds of methods are used for our task. For the first method, We tried several traditional machine learning models like svm, logistic regression, and the linear svm for the second task. Linear svm gets the highest classification accuracy. Second model is CRF with hand-crafted features. And the last model is bi-directional LSTM. Note that there are too many categories in such a small dataset, which explains why even the best model results in relatively low accuracy.
Although there has been a large amount of research focusing on the noise performance of GaN HEMTs in L-bands and Ku-bands, few studies have focused specifically on the noise performance at mm-Wave frequencies. In this letter, the fabricated 120 nm MISHEMTs exhibited a low minimum noise figure (NFmin) of 1.3 dB with an associated gain (Ga) of 7.7 dB at 30 GHz, and NFmin of 1.7 dB and Ga of 6 dB at 40 GHz, showing the lowest values of NFmin compared with the reported GaN HEMTs with the same gate length.Â
In the past decade resistance-based memory devices, or the resistive random access memory Devices (RRAM) have emerged as a potential candidate for multi-state memory storage and non-conventional computing applications. Reports on conduction quantization (QC) have added an interesting layer to the utility of these RRAM devices for ultra-dense memory and neuromorphic computing applications due to the occurrence of integral and half-integral conduction (resistive) states. Since the first reports of QC phenomena in RRAM devices, there have been detailed studies on the nature of the conducting filaments, switching mechanisms, and tunability of the QC states, but there exists a scarcity of studies exploring controllability of QC phenomena in scalable device geometries. In this work, we report compliance current controlled tunable QC phenomena in crossbar RRAM cells based on electrochemical metallization switching mechanism. The devices exhibited robust bipolar resistive switching, with well separated high and low resistance states. Â The magnitude and number of the QC states were found to increase from ~2.5 to 3.5 and from 4 to 6, respectively as the IC increased from 50 to 200Î¼A. The Cu/Ta2O5/Pt device structure was chosen to strategically govern the metallic nature of the conduction filament (CF) formation, which helped postulating factors contributing to the tunability of the states via compliance current. We report the lateral dimension variability as the main factor governing the magnitude and number of quantized steps observed in RRAM devices, where we also discuss a numerical method to approximate the diameter of the CFs. The increase in number and magnitude of QC steps with IC was explained considering the fact that thicker CF obtained at higher ICC, when undergoes a gradual rupture during reset process, results in larger number of QC steps compared to a thinner CF.