Since 2015, there has been an increase in articles on anomaly detection in robotic systems, reflecting its growing importance in improving the robustness and reliability of the increasingly utilized autonomous robots. This review paper investigates the literature on the detection of anomalies in Autonomous Robotic Missions (ARMs). It reveals different perspectives on anomaly and juxtaposition to fault detection. To reach a consensus, we infer a unified understanding of anomalies that encapsulate their various characteristics observed in ARMs and propose a classification of anomalies in terms of spatial, temporal, and spatiotemporal elements based on their fundamental features. Further, the paper discusses the implications of the proposed unified understanding and classification in ARMs and provides future directions. We envisage a study surrounding the specific use of the term anomaly, and methods for their detection could contribute to and accelerate the research and development of a universal anomaly detection system for ARMs.
Blood pressure (BP) monitoring is vital in daily healthcare, especially for cardiovascular diseases. However, BP values are mainly acquired through a contact-sensing method, which is inconvenient and unfriendly for BP monitoring. This paper proposes an efficient end-to-end network for estimating BP values from a facial video to achieve remote BP estimation in daily life. The network first derives a spatiotemporal map of a facial video. Then, it regresses the BP ranges with a designed blood pressure classifier and simultaneously calculates the specific value with a blood pressure calculator in each BP range based on the spatiotemporal map. In addition, an innovative oversampling training strategy was developed to handle the problem of unbalanced data distribution. Finally, we trained the proposed blood pressure estimation network on a private dataset, MPM-BP, and tested it on a popular public dataset, MMSE-HR. As a result, the proposed network achieved a mean absolute error (MAE) and root mean square error (RMSE) of 12.35 mmHg and 16.55 mmHg on systolic BP estimations, and those for diastolic BP were 9.54 mmHg and 12.22 mmHg, which were better than the values obtained in recent works. It can be concluded that the proposed method has excellent potential for camera-based BP monitoring in the indoor scenarios in the real world.
Statistical learning techniques and increased computational power have facilitated the development of self-driving car technology. However, a limiting factor has been the high expense of scaling and maintaining high-definition (HD) maps. These maps are a crucial backbone for many approaches to self-driving technology. In response to this challenge, we present an approach that fuses pre-built point cloud map data with images to automatically and accurately identify static landmarks such as roads, sidewalks, and crosswalks. Our pipeline utilizes semantic segmentation of 2D images, associates semantic labels with points in point cloud maps to pinpoint locations in the physical world, and employs a confusion matrix formulation to generate a probabilistic bird's-eye view semantic map from semantic point clouds. The approach has been tested in an urban area with different segmentation networks to generate a semantic map with road features. The resulting map provides a rich context of the environment that is valuable for downstream tasks such as trajectory generation and intent prediction. Moreover, it has the potential to be extended to the automatic generation of HD maps for semantic features. The entire software pipeline is implemented in the robot operating system (ROS), a widely used robotics framework, and made available.
The acousto-electric (AE) effect associated with the propagation of the Rayleigh wave in ZnO half-space was theoretically investigated by studying the changes in wave velocity and propagation loss induced by in-depth inhomogeneous changes in the ZnO electrical conductivity. An exponentially decaying profile for the electrical conductivity was attributed to the ZnO half-space, for some values of the exponential decay constant (from 100 to 500 nm), in order to simulate the photoconductivity effect induced by ultra-violet illumination. The calculated Rayleigh wave velocity and attenuation vs. ZnO conductivity curves have the form of a double-relaxation response as opposed to the single-relaxation response which characterizes the well-known AE effect due to surface conductivity changes onto piezoelectric media. As to the author’s knowledge, this is the first time the double-relaxation AE effect has been theoretically predicted.
This paper introduces a novel methodology that estimates the wind profile within the ABL by using a neural network along with predictions from a mesoscale model in conjunction with a single near-surface measurement. A major advantage of this solution compared to other solutions available in the literature is that it requires only near-surface measurements for prediction once the neural network has been trained. An additional advantage is the fact that it can be potentially used to explore the time evolution of the wind profile. Data collected by a LiDAR sensor located at the University of León (Spain) is used in the present research. The information obtained from the wind profile is valuable for multiple applications, such as preliminary calculations of the wind asset or CFD modeling.
To address the uncontrollable risks associated with the overreliance on ship operators’ driving in current ship safety braking methods, this study aims to reduce the impact of operator fatigue on navigation safety. Firstly, this study established a human–ship–environment monitoring system with functional and technical architecture, emphasizing the investigation of a ship braking model that integrates brain fatigue monitoring using electroencephalography (EEG) to reduce braking safety risks during navigation. Subsequently, the Stroop task experiment was employed to induce fatigue responses in drivers. By utilizing principal component analysis (PCA) to reduce dimensionality across multiple channels of the data acquisition device, this study extracted centroid frequency (CF) and power spectral entropy (PSE) features from channels 7 and 10. Additionally, a correlation analysis was conducted between these features and the Fatigue Severity Scale (FSS), a five-point scale for assessing fatigue severity in the subjects. This study established a model for scoring driver fatigue levels by selecting the three features with the highest correlation and utilizing ridge regression. The human–ship–environment monitoring system and fatigue prediction model proposed in this study, combined with the ship braking model, achieve a safer and more controllable ship braking process. By real-time monitoring and prediction of driver fatigue, appropriate measures can be taken in a timely manner to ensure navigation safety and driver health.
Two algorithms of data processing are proposed to shorten the unmeasurable dead-zone close to the zero-position of measurement, i.e., the minimum working distance of a dispersive interferometer using a femtosecond laser, which is a critical issue in millimeter-order short-range absolute distance measurement. After demonstrating the limitation of the conventional data processing algorithm, the principles of the proposed algorithms, namely the spectral fringe algorithm and the combined algorithm that combines the spectral fringe algorithm with the excess fraction method, are presented, together with simulation results for demonstrating the possibility of the proposed algorithms for shortening the dead-zone with high accuracy. An experimental setup of a dispersive interferometer is also constructed for implementing the proposed data processing algorithms over spectral interference signals. Experimental results demonstrate that the dead-zone using the proposed algorithms can be as small as half of that of the conventional algorithm while measurement accuracy can be further improved using the combined algorithm.
This paper presents a novel deep-learning (DL)-based approach for classifying digitally modulated signals, which involves the use of capsule networks (CAPs) together with the cyclic cumulant (CC) features of the signals. These were blindly estimated using cyclostationary signal processing (CSP) and were then input into the CAP for training and classification. The classification performance and the generalization abilities of the proposed approach were tested using two distinct datasets that contained the same types of digitally modulated signals, but had distinct generation parameters. The results showed that the classification of digitally modulated signals using CAPs and CCs proposed in the paper outperformed alternative approaches for classifying digitally modulated signals that included conventional classifiers that employed CSP-based techniques, as well as alternative DL-based classifiers that used convolutional neural networks (CNNs) or residual networks (RESNETs) with the in-phase/quadrature (I/Q) data used for training and classification.
This paper presents reported machine learning approaches in the field of Brillouin distributed fiber optic sensors (DFOSs). The increasing popularity of Brillouin DFOSs stems from their capability to continuously monitor temperature and strain along kilometer-long optical fibers, rendering them attractive for industrial applications, such as the structural health monitoring of large civil infrastructures and pipelines. In recent years, machine learning has been integrated into the Brillouin DFOS signal processing, resulting in fast and enhanced temperature, strain, and humidity measurements without increasing the system’s cost. Machine learning has also contributed to enhanced spatial resolution in Brillouin optical time domain analysis (BOTDA) systems and shorter measurement times in Brillouin optical frequency domain analysis (BOFDA) systems. This paper provides an overview of the applied machine learning methodologies in Brillouin DFOSs, as well as future perspectives in this area.
This study addresses the pressing issue of energy consumption and efficiency in the Kingdom of Saudi Arabia (KSA), a region experiencing growing demand for energy resources. Temperature control plays a vital role in achieving energy efficiency; however, traditional control systems may struggle to adapt to the non-linear and time-varying characteristics of the problem. To tackle this challenge, a fuzzy petri net (FPN) controller is proposed as a more suitable solution that combines fuzzy logic (FL) and petri nets (PN) to model and simulate complex systems. The main objective of this research is to develop an intelligent energy-saving framework that integrates advanced methodologies and air conditioning (AC) systems to optimize energy utilization and create a comfortable indoor environment. The proposed system incorporates user identification to authorize individuals who can set the temperature, and FL combined with PN is utilized to monitor and transmit their preferred temperature settings to a PID controller for adjustment. The experimental findings demonstrate the effectiveness of integrating the FPN controller with a convertible frequency AC compressor in significantly reducing energy consumption by 94% compared to using the PN controller alone. The utilization of the PN controller alone resulted in a 25% reduction in energy consumption. Conversely, employing a fixed-frequency compressor led to a 40% increase in energy consumption. These results emphasize the advantages of integrating FL into the PN model, as it effectively reduces energy consumption by half, highlighting the potential of the proposed approach for enhancing energy efficiency in AC systems.
In order to detect the refractive index (RI) of high refractive index materials such as trichlorobenzene and aniline in the near-infrared and mid-infrared spectra and expand the detection range of the refractive index, a surface plasmon resonance (SPR) photonic crystal fiber (PCF) sensor based on an elliptical sensing channel is proposed for high refractive index detection. The fiber core and the analyte channel are surrounded by two types of air holes with different sizes. When the surface plasmon resonance effect appears at the interface between the fiber core and the elliptical sensing layer, obvious resonance peaks appear in the near-infrared and mid-infrared bands. The full vector finite element method (FEM) is used to study the sensing characteristics of the sensor and the influence of structural parameters on the resonance peak. The results demonstrate that the sensor achieves detection in the refractive index range of 1.41–1.58, in the wavelength range of 1600–3200 nm. The average wavelength sensitivity is 9217.22 nm/RIU, and the refractive index resolution is 10.85 × 10−6 RIU. The proposed sensor realizes high refractive index detection in the near-infrared and mid-infrared bands, and obtains an ultra-wide detection range and higher sensitivity. The sensor has broad application prospects in chemical detection, biomedical sensing and other fields, and provides a theoretical reference for the design of a photonic crystal fiber surface plasmon resonance sensor.
This study aimed to evaluate the reliability of the RSQ Motion sensor and its validity against the Propriometer and electronic goniometer in measuring the active range of motion (ROM) of the shoulder. The study included 15 volunteers (mean age 24.73 ± 3.31) without any clinical symptoms with no history of trauma, disease, or surgery to the upper limb. Four movements were tested: flexion, abduction, external and internal rotation. Validation was assessed in the full range of active shoulder motion. Reliability was revised in full active ROM, a fixed angle of 90 degrees for flexion and abduction, and 45 degrees for internal and external rotation. Each participant was assessed three times: on the first day by both testers and on the second day only by one of the testers. Goniometer and RSQ Motion sensors showed moderate to excellent correlation for all tested movements (ICC 0.61–0.97, LOA < 23 degrees). Analysis of inter-rater reliability showed good to excellent agreement between both testers (ICC 0.74–0.97, LOA 13–35 degrees). Analysis of intra-rater reliability showed moderate to a good agreement (ICC 0.7–0.88, LOA 22–37 degrees). The shoulder internal and external rotation measurement with RSQ Motion sensors is valid and reliable. There is a high level of inter-rater and intra-rater reliability for the RSQ Motion sensors and Propriometer.
Convolutional neural networks (CNNs) play a crucial role in many EdgeAI and TinyML applications, but their implementation usually requires external memory, which degrades the feasibility of such resource-hungry environments. To solve this problem, this paper proposes memory-reduction methods at the algorithm and architecture level, implementing a reasonable-performance CNN with the on-chip memory of a practical device. At the algorithm level, accelerator-aware pruning is adopted to reduce the weight memory amount. For activation memory reduction, a stream-based line-buffer architecture is proposed. In the proposed architecture, each layer is implemented by a dedicated block, and the layer blocks operate in a pipelined way. Each block has a line buffer to store a few rows of input data instead of a frame buffer to store the whole feature map, reducing intermediate data-storage size. The experimental results show that the object-detection CNNs of MobileNetV1/V2 and an SSDLite variant, widely used in TinyML applications, can be implemented even on a low-end FPGA without external memory.
Optical microresonators have proven to be especially useful for sensing applications. In most cases, the sensing mechanism is dispersive, where the resonance frequency of a mode shifts in response to a change in the ambient index of refraction. It is also possible to conduct dissipative sensing, in which absorption by an analyte causes measurable changes in the mode linewidth and in the throughput dip depth. If the mode is overcoupled, the dip depth response can be more sensitive than the linewidth response, but overcoupling is not always easy to achieve. We have recently shown theoretically that using multimode input to the microresonator can enhance the dip-depth sensitivity by a factor of several thousand relative to that of single-mode input and by a factor of nearly 100 compared to the linewidth sensitivity. Here, we experimentally confirm these enhancements using an absorbing dye dissolved in methanol inside a hollow bottle resonator. We review the theory, describe the setup and procedure, detail the fabrication and characterization of an asymmetrically tapered fiber to produce multimode input, and present sensing enhancement results that agree with all the predictions of the theory.
Sensor-based monitoring of natural gas pipelines is crucial for safety and detecting sensor faults is pivotal for reliable operations. This paper investigates the problem of fault diagnosis in a natural gas pipeline under transient flow, characterized by a system of hyperbolic partial differential equations (PDEs). A data-fusion approach based on an unscented Kalman filter (UKF) is employed to perform sensor-fault detection and isolation (SFDI). The architecture consists of a bank of local UKFs to provide the state estimates, which are further analyzed to identify sensor faults. The performance achieved by the proposed method is promising even in the case of a nonlinear flow model.
In this study, we employed temperature modulation to address the low selectivity issue of MOX sensors towards gases. We utilized low-frequency square wave and triangle wave signals as heating voltages to classify and predict the concentrations of ethanol, methanol, MEK, and ethyl ethanoate. The results demonstrated that the proposed approach achieved a classification accuracy of 100% under both heating voltages, while the root mean square error (RMSE) for concentration prediction was 0.6928% and 0.9198%, respectively.
Cell characterization is a topic of particular importance in terms of morphological and functional knowledge of a cell population. Currently, there are several techniques and devices used in both clinical and research settings. Recently, thermal dynamics of a cell has been proposed as viable tool for its characterization. Most of them are mainly based on the evaluation of temperature, heat capacity, thermal diffusivity, and conductivity of the cell. In this work, a device based on pyroelectric effect for cell characterization performed by thermal impedance measurements will be presented.
Local feature extractions have been verified to be effective for person re-identification (re-ID) in recent literature. However, existing methods usually rely on extracting local features from single part of a pedestrian while neglecting the relationship of local features among different pedestrian images. As a result, local features contain limited information from one pedestrian image, and cannot benefit from other pedestrian images. In this paper, we propose a novel approach named Local Relation-Aware Graph Convolutional Network (LRGCN) to learn the relationship of local features among different pedestrian images. In order to completely describe the relationship of local features among different pedestrian images, we propose overlap graph and similarity graph. The overlap graph formulates the edge weight as the overlap node number in the node's neighborhoods so as to learn robust local features, and the similarity graph defines the edge weight as the similarity between the nodes to learn discriminative local features. To propagate the information for different kinds of nodes effectively, we propose the Structural Graph Convolution (SGConv) operation. Different from traditional graph convolution operations where all nodes share the same parameter matrix, SGConv learns different parameter matrices for the node itself and its neighbor nodes to improve the expressive power. We conduct comprehensive experiments to verify our method on four large-scale person re-ID databases, and the overall results show LRGCN exceeds the state-of-the-art methods.
